diff --git a/CHANGELOG.md b/CHANGELOG.md index c2fb8cff2765..d6dd6859c861 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -2,6 +2,64 @@ Release notes for OpenCode V2, newest first. +## v2.1.2 — 2026-10-21 + +### Highlights + +- **Clear Free-Limit Recovery**: When your free messages run out, Desktop and the TUI now show the reset time, restore your submitted prompt, and offer direct upgrade or model-switching actions instead of a generic error. ([#57466](https://github.com/anomalyco/opencode/pull/57466), [#57467](https://github.com/anomalyco/opencode/pull/57467), [#57471](https://github.com/anomalyco/opencode/pull/57471)) +- **Hide Free Models**: Set `OPENCODE_DISABLE_FREE_MODELS=1` or `true` to remove zero-cost models from OpenCode and OpenCode Go catalogs, defaults, and fallbacks. ([#57478](https://github.com/anomalyco/opencode/pull/57478)) +- **Configurable Location Unloading**: Set `experimental.location_idle_timeout` to a duration such as `"4 hours"`, or to `false`, to control when inactive location services unload. ([#57476](https://github.com/anomalyco/opencode/pull/57476)) +- **Reliable Fresh Installs**: The installer now creates `~/.zshrc` or Fish’s `config.fish` when missing, adds OpenCode to `PATH`, and tells you to run `exec $SHELL`. ([#57348](https://github.com/anomalyco/opencode/pull/57348)) + +### Core + +#### Improvements + +- Added `OPENCODE_DISABLE_FREE_MODELS` for hiding free OpenCode and OpenCode Go models. ([#57478](https://github.com/anomalyco/opencode/pull/57478)) +- Added `experimental.location_idle_timeout`, accepting a duration string or `false`. ([#57476](https://github.com/anomalyco/opencode/pull/57476)) + +#### Bug fixes + +- Fixed fresh zsh and Fish installations not adding OpenCode to `PATH` when their configuration file did not exist. ([#57348](https://github.com/anomalyco/opencode/pull/57348)) +- Kept locations loaded while sessions are waiting for input, running long tools, or owning background jobs. ([#57416](https://github.com/anomalyco/opencode/pull/57416)) +- Fixed legacy `console.opencode.ai` credentials not being forwarded to `opencode.ai/console`. ([#57588](https://github.com/anomalyco/opencode/pull/57588)) +- Reused recent Console configuration checks when inference is rejected, avoiding unnecessary repeated checks. ([#57465](https://github.com/anomalyco/opencode/pull/57465)) + +### TUI + +#### Improvements + +- Added an inline free-limit notice with the reset time and `/connect` and `/models` actions, while restoring the submitted prompt when the composer is empty. ([#57467](https://github.com/anomalyco/opencode/pull/57467)) +- Restored the animated “Free limit reached” OpenCode Go dialog with **Subscribe**, **Dismiss**, and **Don’t show again** actions. ([#57471](https://github.com/anomalyco/opencode/pull/57471)) + +#### Bug fixes + +- Stopped showing redundant failure notifications when the free-message limit is reached. ([#57467](https://github.com/anomalyco/opencode/pull/57467)) +- Only shows “location unavailable” when the server explicitly reports the location as missing. ([#57594](https://github.com/anomalyco/opencode/pull/57594)) + +### Desktop + +#### Bug fixes + +- Restored the free-limit dialog, reset-time timeline notice, and submitted prompt after free usage is exhausted. Subscribing to OpenCode Go now updates the current session’s model instead of opening another draft. ([#57466](https://github.com/anomalyco/opencode/pull/57466)) + +### SDK + +#### Improvements + +- Added `retryAfterMs` to structured session errors so clients can display provider-supplied retry timing. ([#57146](https://github.com/anomalyco/opencode/pull/57146)) + +#### Bug fixes + +- Fixed retry timing being dropped from quota, rate-limit, and provider-internal session errors. ([#57589](https://github.com/anomalyco/opencode/pull/57589)) +- Fixed OpenAI realtime client-secret requests by supplying the required `expires_after.anchor`. + +### Extensions + +#### Bug fixes + +- Fixed Code Mode async jobs interleaving in the middle of statements and losing writes, while preserving concurrent tool calls. Thanks [@RhysSullivan](https://github.com/RhysSullivan)! ([#54276](https://github.com/anomalyco/opencode/pull/54276)) + ## v2.1.0 — 2026-10-20 ### Highlights diff --git a/bun.lock b/bun.lock index 10923cf993de..70364f66c561 100644 --- a/bun.lock +++ b/bun.lock @@ -34,7 +34,7 @@ }, "packages/ai": { "name": "@opencode/ai", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@aws-sdk/credential-providers": "3.1057.0", "@opencode/schema": "workspace:*", @@ -56,7 +56,7 @@ }, "packages/cli": { "name": "@opencode/cli", - "version": "2.1.0", + "version": "2.1.2", "bin": { "opencode": "./bin/opencode.cjs", "opencode2": "./bin/opencode2.cjs", @@ -75,7 +75,7 @@ "@opencode/util": "workspace:*", "@opentui/core": "catalog:", "@opentui/solid": "catalog:", - "@opentunnel/client": "0.4.0", + "@opentunnel/client": "0.4.1", "@parcel/watcher": "2.5.1", "@silvia-odwyer/photon-node": "0.3.4", "diff": "catalog:", @@ -124,7 +124,7 @@ }, "packages/client": { "name": "@opencode/client", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@opencode/protocol": "workspace:*", "@opencode/schema": "workspace:*", @@ -149,7 +149,7 @@ }, "packages/codemode": { "name": "@opencode/codemode", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "acorn": "8.15.0", "effect": "catalog:", @@ -162,7 +162,7 @@ }, "packages/core": { "name": "@opencode/core", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@ai-sdk/cohere": "3.0.27", "@ai-sdk/perplexity": "3.0.26", @@ -228,7 +228,7 @@ }, "packages/http-recorder": { "name": "@opencode/http-recorder", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@effect/platform-node-shared": "catalog:", }, @@ -247,7 +247,7 @@ }, "packages/httpapi-codegen": { "name": "@opencode/httpapi-codegen", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "effect": "catalog:", "prettier": "3.6.2", @@ -319,7 +319,7 @@ }, "packages/latex": { "name": "@opencode/latex", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@opencode/plugin": "workspace:*", "@opentui/core": "catalog:", @@ -333,7 +333,7 @@ }, "packages/merman": { "name": "@opencode/merman", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@opencode/plugin": "workspace:*", "@opentui/core": "catalog:", @@ -348,7 +348,7 @@ }, "packages/plugin": { "name": "@opencode/plugin", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@ai-sdk/provider": "3.0.8", "@opencode/ai": "workspace:*", @@ -385,7 +385,7 @@ }, "packages/plugin-browser": { "name": "@opencode/plugin-browser", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@opencode/plugin": "workspace:*", "@opencode/schema": "workspace:*", @@ -400,13 +400,13 @@ }, "packages/plugins/figma": { "name": "@opencode/plugin-figma", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "gray-matter": "4.0.3", }, "devDependencies": { "@tsconfig/bun": "1.0.9", - "@types/bun": "1.4.0", + "@types/bun": "1.4.3", "@typescript/native": "npm:typescript@7.0.2", }, "peerDependencies": { @@ -415,7 +415,7 @@ }, "packages/protocol": { "name": "@opencode/protocol", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@opencode/schema": "workspace:*", "effect": "catalog:", @@ -430,7 +430,7 @@ }, "packages/schema": { "name": "@opencode/schema", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@standard-schema/spec": "catalog:", "effect": "catalog:", @@ -454,7 +454,7 @@ }, "packages/sdk": { "name": "@opencode/sdk", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@opencode/client": "workspace:*", "@opencode/core": "workspace:*", @@ -475,7 +475,7 @@ }, "packages/server": { "name": "@opencode/server", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@effect/platform-node": "catalog:", "@effect/platform-node-shared": "catalog:", @@ -498,7 +498,7 @@ }, "packages/simulation": { "name": "@opencode/simulation", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@opencode/ai": "workspace:*", "@opencode/core": "workspace:*", @@ -518,7 +518,7 @@ }, "packages/theme": { "name": "@opencode/theme", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@opentui/core": "catalog:", "effect": "catalog:", @@ -532,7 +532,7 @@ }, "packages/tui": { "name": "@opencode/tui", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@opencode/client": "workspace:*", "@opencode/core": "workspace:*", @@ -565,20 +565,19 @@ }, "packages/ui": { "name": "@opencode/ui", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@kitlangton/solid-motion": "0.2.2", "@kobalte/core": "catalog:", "@pierre/diffs": "catalog:", + "@pierre/highlights": "catalog:", "@solid-primitives/event-listener": "catalog:", "@solid-primitives/resize-observer": "catalog:", "fuzzysort": "catalog:", "katex": "0.16.47", "marked": "catalog:", - "marked-shiki": "catalog:", "motion": "12.34.5", "remeda": "catalog:", - "shiki": "catalog:", "solid-list": "catalog:", "solid-sonner": "catalog:", }, @@ -601,7 +600,7 @@ }, "packages/util": { "name": "@opencode/util", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "@effect/opentelemetry": "catalog:", "@effect/platform-node": "catalog:", @@ -661,7 +660,7 @@ }, "services/update": { "name": "@opencode/update", - "version": "2.1.0", + "version": "2.1.2", "dependencies": { "jose": "6.0.11", "semver": "catalog:", @@ -713,7 +712,7 @@ "@opentui/solid": "catalog:", "@types/bun": "catalog:", "@types/node": "catalog:", - "bun-types": "1.4.2", + "bun-types": "1.4.4", "effect": "catalog:", "open": "11.0.4", "solid-js": "catalog:", @@ -734,11 +733,11 @@ "@opentui/core": "0.6.6", "@opentui/keymap": "0.6.6", "@opentui/solid": "0.6.6", - "@pierre/diffs": "1.5.1", + "@pierre/diffs": "1.6.1", + "@pierre/highlights": "0.2.0", "@playwright/test": "1.60.0", "@sentry/solid": "10.71.0", "@sentry/vite-plugin": "5.4.0", - "@shikijs/stream": "4.4.3", "@solid-primitives/event-listener": "2.4.6", "@solid-primitives/media": "2.3.6", "@solid-primitives/resize-observer": "2.2.0", @@ -751,7 +750,7 @@ "@tanstack/solid-virtual": "3.13.37", "@tsconfig/bun": "1.0.9", "@tsconfig/node22": "22.0.2", - "@types/bun": "1.4.0", + "@types/bun": "1.4.3", "@types/cross-spawn": "6.0.6", "@types/luxon": "3.7.1", "@types/node": "24.12.2", @@ -1509,9 +1508,9 @@ "@opentui/solid": ["@opentui/solid@0.6.6", "", { "dependencies": { "@babel/core": "7.29.7", "@babel/preset-typescript": "7.27.1", "@opentui/core": "0.6.6", "babel-plugin-module-resolver": "5.0.2", "babel-preset-solid": "1.9.12", "entities": "7.0.1", "s-js": "^0.4.9" }, "peerDependencies": { "solid-js": "1.9.12" } }, "sha512-YqwrLIiqjUCLUNQuN4kA1pDEVcoYaXaSbg6kt8OlKh+uuHRwq43qmCAIz19md6Ld0FrPtKKbkxPMLKyiSyVLUg=="], - "@opentunnel/client": ["@opentunnel/client@0.4.0", "", { "dependencies": { "@opentunnel/protocol": "0.4.0" }, "peerDependencies": { "effect": "^4.0.0" } }, "sha512-Y6UiO/ouJ/vtdIy4QmCTRY2DPTRpHy2+jR0F6+AKXoD57ozd8W9ZTTP2J1KDqTb7ZfJmRTuqtx1ZR2TnejlZGQ=="], + "@opentunnel/client": ["@opentunnel/client@0.4.1", "", { "dependencies": { "@opentunnel/protocol": "0.4.1" }, "peerDependencies": { "effect": "^4.0.0" } }, "sha512-TLHCko/cSoF6LY/IBvex9s/u9vh4CB3K/S06q7Zwu0HUISikMC+j/4437Y39slF7hmx1oSpTo7AySELu7j5DQg=="], - "@opentunnel/protocol": ["@opentunnel/protocol@0.4.0", "", { "peerDependencies": { "effect": "^4.0.0" } }, "sha512-rxmfgJ5O1EYxgzawkOti9DyiM44l0Bkeue5KuHpCFaPAP2n9i4xDYCU8XQp9WAEFt7ifZZqI9giOyzJE61faaA=="], + "@opentunnel/protocol": ["@opentunnel/protocol@0.4.1", "", { "peerDependencies": { "effect": "^4.0.0" } }, "sha512-ebAO6/H5VdeAtweqWhMBqTthTMqQfsmZRsbHJULpDLK3HGJa02L1VO4e4LnigLbIPhP/PULqPjFe5nmcMXUvNQ=="], "@oslojs/encoding": ["@oslojs/encoding@1.1.0", "", {}, "sha512-70wQhgYmndg4GCPxPPxPGevRKqTIJ2Nh4OkiMWmDAVYsTQ+Ta7Sq+rPevXyXGdzr30/qZBnyOalCszoMxlyldQ=="], @@ -1611,11 +1610,13 @@ "@parcel/watcher-win32-x64": ["@parcel/watcher-win32-x64@2.5.1", "", { "os": "win32", "cpu": "x64" }, "sha512-9lHBdJITeNR++EvSQVUcaZoWupyHfXe1jZvGZ06O/5MflPcuPLtEphScIBL+AiCWBO46tDSHzWyD0uDmmZqsgA=="], - "@pierre/diffs": ["@pierre/diffs@1.5.1", "", { "dependencies": { "@pierre/theme": "2.0.0", "@pierre/theming": "1.0.1", "@shikijs/transformers": "^3.0.0 || ^4.0.0", "diff": "9.0.0", "hast-util-to-html": "9.0.5", "lru_map": "0.4.1", "shiki": "^3.0.0 || ^4.0.0" }, "peerDependencies": { "react": "^18.3.1 || ^19.0.0", "react-dom": "^18.3.1 || ^19.0.0" }, "optionalPeers": ["react", "react-dom"] }, "sha512-+EXNfz4ZXI6FHr7P6ToWEORwyYiPVfAmzhA70xF+0y/oNvj7gbNTLJJ6jNzkBjuMbRlu564TAmgYZLc5OyHURA=="], + "@pierre/diffs": ["@pierre/diffs@1.6.1", "", { "dependencies": { "@pierre/highlights": "0.2.0", "@pierre/theme": "2.0.0", "@pierre/theming": "1.0.2", "diff": "9.0.0", "hast-util-to-html": "9.0.5", "lru_map": "0.4.1", "shiki": "^3.0.0 || ^4.0.0" }, "peerDependencies": { "react": "^18.3.1 || ^19.0.0", "react-dom": "^18.3.1 || ^19.0.0" }, "optionalPeers": ["react", "react-dom"] }, "sha512-XTfLW4bFETyosZfFNUoyRh1zPE4tmAijOoxdHx4X805tlaq/7/fuQXIiJSm21+jLkDo8vn5elQ+2O4YtMTKM0Q=="], + + "@pierre/highlights": ["@pierre/highlights@0.2.0", "", {}, "sha512-hTpcOU1EzDAP4bYtEhxmMu2jwMvtkUydRUBUpWMSHTs3dw0IaC3Ld5ansMy18L0GE3b6Vfl2OGkLK+pWWpnBmw=="], "@pierre/theme": ["@pierre/theme@2.0.0", "", {}, "sha512-yNDd9GYLQl1mEUJR8AneJ5e4ohLIHQd/wZLWr4fagt78vS2RwwZNW530vVgHqXFAyFVcFlRmGUD5ramXH46OXw=="], - "@pierre/theming": ["@pierre/theming@1.0.1", "", { "peerDependencies": { "@pierre/theme": "^1.1.0 || ^2.0.0", "@shikijs/themes": "^3.0.0 || ^4.0.0", "react": "^18.3.1 || ^19.0.0", "react-dom": "^18.3.1 || ^19.0.0", "shiki": "^3.0.0 || ^4.0.0" }, "optionalPeers": ["@pierre/theme", "@shikijs/themes", "react", "react-dom", "shiki"] }, "sha512-WCI5Qd7iprDpISL9fBYOLe8RV53+b7mFNA3bPzl60/2CKCSrsKN8zEcep6Y3BAzvARlmca50zGjDodqPGiTUKA=="], + "@pierre/theming": ["@pierre/theming@1.0.2", "", { "peerDependencies": { "@pierre/theme": "^1.1.0 || ^2.0.0", "@shikijs/themes": "^3.0.0 || ^4.0.0", "react": "^18.3.1 || ^19.0.0", "react-dom": "^18.3.1 || ^19.0.0", "shiki": "^3.0.0 || ^4.0.0" }, "optionalPeers": ["@pierre/theme", "@shikijs/themes", "react", "react-dom", "shiki"] }, "sha512-RmXLLqtx31UGcR8kZyO5hGAQ1EXz6loT1Cu0BFAGsbnBboIxp2BlomkLfWyvdT4WZXnBGyu25onL1gKpFQuSUw=="], "@pkgjs/parseargs": ["@pkgjs/parseargs@0.11.0", "", {}, "sha512-+1VkjdD0QBLPodGrJUeqarH8VAIvQODIbwh9XpP5Syisf7YoQgsJKPNFoqqLQlu+VQ/tVSshMR6loPMn8U+dPg=="], @@ -1739,8 +1740,6 @@ "@shikijs/themes": ["@shikijs/themes@4.4.3", "", { "dependencies": { "@shikijs/types": "4.4.3" } }, "sha512-w8UHjeUnIR965KMWJHUPXOc2mNJUnK3vpVLYLvw5IYU2mnTTJ89E24OrJDBNiJDQ0qzb0tc4l7mrIXx5cFeIyw=="], - "@shikijs/transformers": ["@shikijs/transformers@3.20.0", "", { "dependencies": { "@shikijs/core": "3.20.0", "@shikijs/types": "3.20.0" } }, "sha512-PrHHMRr3Q5W1qB/42kJW6laqFyWdhrPF2hNR9qjOm1xcSiAO3hAHo7HaVyHE6pMyevmy3i51O8kuGGXC78uK3g=="], - "@shikijs/types": ["@shikijs/types@4.4.3", "", { "dependencies": { "@shikijs/vscode-textmate": "^10.0.2", "@types/hast": "^3.0.5" } }, "sha512-UEJxmRR++MAGR6hugn0vgVS2W/6lWAts84FFSrnlH9sP0LNol7E5+NQ792pH8liWUhyMyjhTgSUH3k7iD7tc5g=="], "@shikijs/vscode-textmate": ["@shikijs/vscode-textmate@10.0.2", "", {}, "sha512-83yeghZ2xxin3Nj8z1NMd/NCuca+gsYXswywDy5bHvwlWL8tpTQmzGeUuHd9FC3E/SBEMvzJRwWEOz5gGes9Qg=="], @@ -1909,7 +1908,7 @@ "@types/babel__traverse": ["@types/babel__traverse@7.28.0", "", { "dependencies": { "@babel/types": "^7.28.2" } }, "sha512-8PvcXf70gTDZBgt9ptxJ8elBeBjcLOAcOtoO/mPJjtji1+CdGbHgm77om1GrsPxsiE+uXIpNSK64UYaIwQXd4Q=="], - "@types/bun": ["@types/bun@1.4.0", "", { "dependencies": { "bun-types": "1.4.0" } }, "sha512-K+lZULY23vRgK/CfTjFIV+tyifaNdSMlPh9j+6mQ/cLfpOznLyAuzgV/JQysyECpkBQLVMSyvjlr2fBUSA9wFQ=="], + "@types/bun": ["@types/bun@1.4.3", "", { "dependencies": { "bun-types": "1.4.3" } }, "sha512-N67Pqy8JXbAY/j0tNraS3V5mubGvc6rrrNDmDCtZLlFzccbUJPDb0tvgUBC5q4m8ybkP93Jcn1QB+RYQ6PvBhw=="], "@types/cacache": ["@types/cacache@20.0.1", "", { "dependencies": { "@types/node": "*", "minipass": "*" } }, "sha512-QlKW3AFoFr/hvPHwFHMIVUH/ZCYeetBNou3PCmxu5LaNDvrtBlPJtIA6uhmU9JRt9oxj7IYoqoLcpxtzpPiTcw=="], @@ -2159,7 +2158,7 @@ "bun-pty": ["bun-pty@0.4.11", "", {}, "sha512-DpBfXCoWbcDziqL4gT3sFw/O6vfe2Gm1LmsnjfGjpqNhmsPjBcVorTrRv7OVATsy88PiXfn486vp0w4x3/Q84g=="], - "bun-types": ["bun-types@1.4.2", "", { "dependencies": { "@types/node": "*" } }, "sha512-bxV1FgK7yBIzjRe5zBozIM4Bem11ZJcCXSrjWRG3YWLt8yFDePu4cLjpebO8OvPeIE9trbyPF4fuj3Cia4Fj3w=="], + "bun-types": ["bun-types@1.4.4", "", { "dependencies": { "@types/node": "*", "undici-types": "*" } }, "sha512-yjrJh5oGX7mMPKGd19YXyIjmEuj4G4zNC30bs7PM9Az7NQ3ndF3j6hbeRdvJTRQEhJ/FYQwW0a2moweEi2hMxQ=="], "bundle-name": ["bundle-name@4.1.0", "", { "dependencies": { "run-applescript": "^7.0.0" } }, "sha512-tjwM5exMg6BGRI+kNmTntNsvdZS1X8BFYS6tnJ2hdH0kVxM6/eVZ2xy+FqStSWvYmtfFMDLIxurorHwDKfDz5Q=="], @@ -2667,8 +2666,6 @@ "marked": ["marked@18.0.11", "", { "bin": { "marked": "bin/marked.js" } }, "sha512-HnslJfsZkRPBDJRHvVtAaWlZHEpSu7u8LgQuJCELjRKuWR+hpq4A7sLq3p8HaI9ypVoXDXxV34CsQJEe1+J5Aw=="], - "marked-shiki": ["marked-shiki@1.2.1", "", { "peerDependencies": { "marked": ">=7.0.0", "shiki": ">=1.0.0" } }, "sha512-yHxYQhPY5oYaIRnROn98foKhuClark7M373/VpLxiy5TrDu9Jd/LsMwo8w+U91Up4oDb9IXFrP0N1MFRz8W/DQ=="], - "math-intrinsics": ["math-intrinsics@1.1.0", "", {}, "sha512-/IXtbwEk5HTPyEwyKX6hGkYXxM9nbj64B+ilVJnC/R6B0pH5G4V3b0pVbL7DBj4tkhBAppbQUlf6F6Xl9LHu1g=="], "mdast-util-definitions": ["mdast-util-definitions@6.0.0", "", { "dependencies": { "@types/mdast": "^4.0.0", "@types/unist": "^3.0.0", "unist-util-visit": "^5.0.0" } }, "sha512-scTllyX6pnYNZH/AIp/0ePz6s4cZtARxImwoPJ7kS42n+MnVsI4XbnG6d4ibehRIldYMWM2LD7ImQblVhUejVQ=="], @@ -3555,10 +3552,6 @@ "@rollup/pluginutils/estree-walker": ["estree-walker@2.0.2", "", {}, "sha512-Rfkk/Mp/DL7JVje3u18FxFujQlTNR2q6QfMSMB7AvCBx91NGj/ba3kCfza0f6dVDbw7YlRf/nDrn7pQrCCyQ/w=="], - "@shikijs/transformers/@shikijs/core": ["@shikijs/core@3.20.0", "", { "dependencies": { "@shikijs/types": "3.20.0", "@shikijs/vscode-textmate": "^10.0.2", "@types/hast": "^3.0.4", "hast-util-to-html": "^9.0.5" } }, "sha512-f2ED7HYV4JEk827mtMDwe/yQ25pRiXZmtHjWF8uzZKuKiEsJR7Ce1nuQ+HhV9FzDcbIo4ObBCD9GPTzNuy9S1g=="], - - "@shikijs/transformers/@shikijs/types": ["@shikijs/types@3.20.0", "", { "dependencies": { "@shikijs/vscode-textmate": "^10.0.2", "@types/hast": "^3.0.4" } }, "sha512-lhYAATn10nkZcBQ0BlzSbJA3wcmL5MXUUF8d2Zzon6saZDlToKaiRX60n2+ZaHJCmXEcZRWNzn+k9vplr8Jhsw=="], - "@vscode/emmet-helper/jsonc-parser": ["jsonc-parser@2.3.1", "", {}, "sha512-H8jvkz1O50L3dMZCsLqiuB2tA7muqbSg1AtGEkN0leAqGjsUzDJir3Zwr02BhqdcITPg3ei3mZ+HjMocAknhhg=="], "anymatch/picomatch": ["picomatch@2.3.2", "", {}, "sha512-V7+vQEJ06Z+c5tSye8S+nHUfI51xoXIXjHQ99cQtKUkQqqO1kO/KCJUfZXuB47h/YBlDhah2H3hdUGXn8ie0oA=="], diff --git a/bunfig.toml b/bunfig.toml index edbde6d0fd0c..6cf566729074 100644 --- a/bunfig.toml +++ b/bunfig.toml @@ -2,7 +2,7 @@ exact = true # Only install newly resolved package versions published at least 3 days ago. minimumReleaseAge = 259200 -minimumReleaseAgeExcludes = ["@ai-sdk/amazon-bedrock", "@ai-sdk/anthropic", "@brendonovich/vite-plugin-opencode", "@kitlangton/solid-motion", "@opencode/sdk", "@opencode-ai/pty", "@opencode-ai/pty-darwin-arm64", "@opencode-ai/pty-darwin-x64", "@opencode-ai/pty-linux-arm64-gnu", "@opencode-ai/pty-linux-arm64-musl", "@opencode-ai/pty-linux-x64-gnu", "@opencode-ai/pty-linux-x64-musl", "@opencode-ai/pty-win32-arm64", "@opencode-ai/pty-win32-x64", "@opentui/core", "@opentui/core-darwin-arm64", "@opentui/core-darwin-x64", "@opentui/core-linux-arm64", "@opentui/core-linux-arm64-musl", "@opentui/core-linux-x64", "@opentui/core-linux-x64-musl", "@opentui/core-win32-arm64", "@opentui/core-win32-x64", "@opentui/keymap", "@opentui/solid", "@opentunnel/client", "@opentunnel/protocol", "gitlab-ai-provider", "opencode-gitlab-auth", "@ff-labs/fff-node", "@ff-labs/fff-bun", "@ff-labs/fff-bin-darwin-arm64", "@ff-labs/fff-bin-darwin-x64", "@ff-labs/fff-bin-linux-arm64-gnu", "@ff-labs/fff-bin-linux-arm64-musl", "@ff-labs/fff-bin-linux-x64-gnu", "@ff-labs/fff-bin-linux-x64-musl", "@ff-labs/fff-bin-win32-arm64", "@ff-labs/fff-bin-win32-x64", "@pierre/diffs", "@pierre/theming", "app-builder-lib", "dmg-builder", "electron", "electron-builder", "electron-publish", "blume", "mermaid"] +minimumReleaseAgeExcludes = ["@ai-sdk/amazon-bedrock", "@ai-sdk/anthropic", "@brendonovich/vite-plugin-opencode", "@kitlangton/solid-motion", "@opencode/sdk", "@opencode-ai/pty", "@opencode-ai/pty-darwin-arm64", "@opencode-ai/pty-darwin-x64", "@opencode-ai/pty-linux-arm64-gnu", "@opencode-ai/pty-linux-arm64-musl", "@opencode-ai/pty-linux-x64-gnu", "@opencode-ai/pty-linux-x64-musl", "@opencode-ai/pty-win32-arm64", "@opencode-ai/pty-win32-x64", "@opentui/core", "@opentui/core-darwin-arm64", "@opentui/core-darwin-x64", "@opentui/core-linux-arm64", "@opentui/core-linux-arm64-musl", "@opentui/core-linux-x64", "@opentui/core-linux-x64-musl", "@opentui/core-win32-arm64", "@opentui/core-win32-x64", "@opentui/keymap", "@opentui/solid", "@opentunnel/client", "@opentunnel/protocol", "gitlab-ai-provider", "opencode-gitlab-auth", "@ff-labs/fff-node", "@ff-labs/fff-bun", "@ff-labs/fff-bin-darwin-arm64", "@ff-labs/fff-bin-darwin-x64", "@ff-labs/fff-bin-linux-arm64-gnu", "@ff-labs/fff-bin-linux-arm64-musl", "@ff-labs/fff-bin-linux-x64-gnu", "@ff-labs/fff-bin-linux-x64-musl", "@ff-labs/fff-bin-win32-arm64", "@ff-labs/fff-bin-win32-x64", "@pierre/diffs", "@pierre/highlights", "@pierre/theming", "app-builder-lib", "dmg-builder", "electron", "electron-builder", "electron-publish", "blume", "mermaid", "bun-types"] [test] root = "./do-not-run-tests-from-root" diff --git a/github/README.md b/github/README.md index 17b24ffb1d6e..e08ae25899ca 100644 --- a/github/README.md +++ b/github/README.md @@ -119,7 +119,7 @@ To test locally: MODEL=anthropic/claude-sonnet-4-20250514 \ ANTHROPIC_API_KEY=sk-ant-api03-1234567890 \ GITHUB_RUN_ID=dummy \ - MOCK_TOKEN=github_pat_1234567890 \ + TOKEN=github_pat_1234567890 \ MOCK_EVENT='{"eventName":"issue_comment",...}' \ bun /path/to/opencode/github/index.ts ``` @@ -127,7 +127,7 @@ To test locally: - `MODEL`: The model used by opencode. Same as the `MODEL` defined in the GitHub workflow. - `ANTHROPIC_API_KEY`: Your model provider API key. Same as the keys defined in the GitHub workflow. - `GITHUB_RUN_ID`: Dummy value to emulate GitHub action environment. - - `MOCK_TOKEN`: A GitHub personal access token. This token is used to verify you have `admin` or `write` access to the test repo. Generate a token [here](https://github.com/settings/personal-access-tokens). + - `TOKEN`: A GitHub personal access token with write access to the test repo, used directly for GitHub API calls. Generate a token [here](https://github.com/settings/personal-access-tokens). - `MOCK_EVENT`: Mock GitHub event payload (see templates below). - `/path/to/opencode`: Path to your cloned opencode repo. `bun /path/to/opencode/github/index.ts` runs your local version of `opencode`. diff --git a/github/index.ts b/github/index.ts index e8acd9a9cef3..48cd05c1a0df 100644 --- a/github/index.ts +++ b/github/index.ts @@ -334,7 +334,6 @@ function useEnvShare() { function useEnvMock() { return { mockEvent: process.env["MOCK_EVENT"], - mockToken: process.env["MOCK_TOKEN"], } } @@ -343,8 +342,7 @@ function useEnvGithubToken() { } function isMock() { - const { mockEvent, mockToken } = useEnvMock() - return Boolean(mockEvent || mockToken) + return Boolean(useEnvMock().mockEvent) } function isPullRequest() { @@ -367,29 +365,17 @@ function useShareUrl() { } async function getAccessToken() { - const { repo } = useContext() - const envToken = useEnvGithubToken() if (envToken) return envToken + if (isMock()) throw new Error("TOKEN is required when running with MOCK_EVENT") - let response - if (isMock()) { - response = await fetch("https://api.opencode.ai/exchange_github_app_token_with_pat", { - method: "POST", - headers: { - Authorization: `Bearer ${useEnvMock().mockToken}`, - }, - body: JSON.stringify({ owner: repo.owner, repo: repo.repo }), - }) - } else { - const oidcToken = await core.getIDToken("opencode-github-action") - response = await fetch("https://api.opencode.ai/exchange_github_app_token", { - method: "POST", - headers: { - Authorization: `Bearer ${oidcToken}`, - }, - }) - } + const oidcToken = await core.getIDToken("opencode-github-action") + const response = await fetch("https://api.opencode.ai/exchange_github_app_token", { + method: "POST", + headers: { + Authorization: `Bearer ${oidcToken}`, + }, + }) if (!response.ok) { const responseJson = (await response.json()) as { error?: string } diff --git a/install b/install index 8d4fa6639fcf..5e8a3cd4cfbc 100755 --- a/install +++ b/install @@ -366,6 +366,7 @@ fi install_legacy_shim +path_added=false add_to_path() { local config_file=$1 local command=$2 @@ -375,6 +376,7 @@ add_to_path() { elif [[ -w $config_file ]]; then echo -e "\n# opencode" >> "$config_file" echo "$command" >> "$config_file" + path_added=true print_message info "${MUTED}Successfully added ${NC}opencode ${MUTED}to \$PATH in ${NC}$config_file" else print_message warning "Manually add the directory to $config_file (or similar):" @@ -416,6 +418,18 @@ if [[ "$no_modify_path" != "true" ]]; then fi done + # A new Mac has no ~/.zshrc, so create the shell's main config file rather than leaving opencode off PATH. + if [[ -z $config_file ]]; then + case $current_shell in + zsh) new_config_file="${ZDOTDIR:-$HOME}/.zshrc" ;; + fish) new_config_file="$HOME/.config/fish/config.fish" ;; + *) new_config_file="" ;; + esac + if [[ -n $new_config_file ]] && mkdir -p "$(dirname "$new_config_file")" 2>/dev/null && touch "$new_config_file" 2>/dev/null; then + config_file=$new_config_file + fi + fi + if [[ -z $config_file ]]; then print_message warning "No config file found for $current_shell. You may need to manually add to PATH:" print_message info " export PATH=$INSTALL_DIR:\$PATH" @@ -459,6 +473,9 @@ echo -e "" echo -e "" echo -e "${MUTED}OpenCode includes free models, to start:${NC}" echo -e "" +if [[ $path_added == "true" ]]; then + echo -e "exec \$SHELL ${MUTED}# Reload PATH (or open a new terminal)${NC}" +fi echo -e "cd ${MUTED}# Open directory${NC}" echo -e "opencode ${MUTED}# Run command${NC}" echo -e "" diff --git a/migration-tracking/technical-notes/baseline/pinned-v2-baseline.md b/migration-tracking/technical-notes/baseline/pinned-v2-baseline.md index 4c2e3df4eff1..1621ff7d7ed1 100644 --- a/migration-tracking/technical-notes/baseline/pinned-v2-baseline.md +++ b/migration-tracking/technical-notes/baseline/pinned-v2-baseline.md @@ -1,10 +1,43 @@ # Pinned upstream v2 baseline -Status: living. Refreshed 2026-10-23 on the `v2.1.1` sync branch `sync-v2-1-1`, synced to upstream `v2.1.1`. The base moves with every upstream sync (plan A4), so re-measure after each sync. The 2026-10-16 capture at `v2.0.30` and the 2026-10-11 capture at the initial base are kept below as previous captures. The 2026-08-13 capture at `76dbaf20` (old branch `johnnyeric/v2-baseline`) survives in git history only. +Status: living. Refreshed 2026-10-23 on the `v2.1.3` sync branch `sync-v2-1-3`, synced to upstream `v2.1.3` (the sync includes `v2.1.2`). The base moves with every upstream sync (plan A4), so re-measure after each sync. The `v2.1.1` capture (2026-10-23), the 2026-10-16 capture at `v2.0.30` and the 2026-10-11 capture at the initial base are kept below as previous captures. The 2026-08-13 capture at `76dbaf20` (old branch `johnnyeric/v2-baseline`) survives in git history only. This file records the upstream release `origin/v2` is synced to, the validation commands that this tree defines, and the result of running them. Compare any failing check against this file, and against `origin/v2` before your change, before you attribute it to a change (`migration-context.md`, Conventions). -## Current capture: `v2.1.1` (2026-10-23) +## Current capture: `v2.1.3` (2026-10-23) + +### Refs + +| Ref | SHA | Date | Note | +|---|---|---|---| +| Sync branch (measured) | `021d6b31848` | 2026-10-23 | `sync-v2-1-3`: the `v2.1.3` sync merge `0d8ae9be328` plus `origin/v2` `2af82eef6a8` (#15405). After the later merge of `origin/v2` `aa7d29226e4` (#15345, #15407) at `d75e69380b1`, `check`, the generated-file checks, the Kilo checks and the `tui` (1,604), `plugin` (9), `kilo-plugin` (188) and `kilo-cli` (62) tests were run again and pass | +| Current upstream base | `37d973754ba6cc31732f93421c0c18b675841573` | 2026-10-22 | Release commit of `v2.1.3` (parent of the tag) | +| `upstream/dev` at capture | `2a3e7c0f349` | 2026-10-23 | 72 commits past the current base. Lag: 1 release (`v2.1.4`) | + +Measured with `bun migration-tracking/technical-notes/script/v2-baseline.ts --checks` in a clean detached worktree outside the repository. Toolchain: bun 1.4.4 (the new `packageManager`), node v25.2.1, macOS 26.6 arm64, macOS bash 3.2, `rg` on PATH (Homebrew), login shell fish. `origin/v2` `2af82eef6a8` was measured with the same script (`--only test`) and bun 1.4.2. + +### Results + +| Check | Result | Detail | +|---|---|---| +| `install`, `check` (`bun run check`), effect simplification lint and rule tests | pass | `bun.lock` unchanged. `check`: 28/28 turbo tasks, oxlint 209 warnings, 0 errors | +| `test` (`bun turbo test --continue --force`, isolated) | **fail** | 25/26 tasks pass, 4 min 8 s. `@opencode/core`: 6,732 pass, 13 fail (`origin/v2`: 6,693 pass, the same 13 fail). All other packages pass, for example `@opencode/cli` 339, `@opencode/tui` 1,601, `@opencode/server` 87, `@kilocode/gateway-core` 104, `@kilocode/kilo-plugin` 188, `@kilocode/kilo-cli` 62 | +| `client`, `www`, `protocol` `check:generated` | pass | No drift | +| `sdk-verify-package`, `codemode-publish-dry-run` | pass | | +| CLI build, binary version, service smoke | pass | | +| Kilo build, host smoke, CLI startup ratchet | pass | | +| Workflow allowlist, annotations, shared patches, progress map, Kilo import rule | pass | Shared patches: 46 files, 334 marked and 39 unmarked lines against `37d973754ba` (at `d75e69380b1`) | +| Node build and smoke | not run | CI pins Node 26.4.0; this host has v25.2.1 | +| `lint:effect-patterns` | **fail** | 27 errors, unchanged. Not run by CI | + +### Current reds + +The 13 `@opencode/core` cases of the `v2.1.1` capture, unchanged. The failure set is identical on `origin/v2` `2af82eef6a8` (before the sync) on this host. None is attributable to the sync or to a Kilo change. + +1. **macOS bash 3.2, 7 cases:** `ShellTool conditional process substitution` (4), `ShellTool compound syntax approval compatibility` (2), `real-shell compound syntax > bash` (1). Green on Linux CI. +2. **Login shell, 6 cases:** `Session.shell > …` (`packages/core/test/session-shell.test.ts`) times out when `SHELL` is fish. With `SHELL=/bin/zsh` the file passes (9 pass). + +## Previous capture: `v2.1.1` (2026-10-23) ### Refs diff --git a/nix/hashes.json b/nix/hashes.json index aad281678517..ea1a80418090 100644 --- a/nix/hashes.json +++ b/nix/hashes.json @@ -1,7 +1,7 @@ { "nodeModules": { - "x86_64-linux": "sha256-XWwlhjvmCdt68mIHTUQsmDO/lBzrYjO34/3tob3d1tA=", - "aarch64-linux": "sha256-/310R94wlUd6rSDE9PJFf1FVXSyS6hAan8zi2NqEj4k=", - "aarch64-darwin": "sha256-0NZH0SXQ7e64wiCy2C9sKlygrr41fqe7JRoBLD8rRIg=" + "x86_64-linux": "sha256-y9Zr4/sIxFh5i/6AYl/1N9PzdFBfPnyXEyjqW/idPes=", + "aarch64-linux": "sha256-4JBTGcwS4GsfSSvnKvcBR/6aCBghTwUSRz+E8lsQpO4=", + "aarch64-darwin": "sha256-hn62idpf73/uW+nCLLO5hhn/HTwwdV2NmmkqNTqo+2g=" } } diff --git a/package.json b/package.json index 581b0ccf1049..a9a817945f11 100644 --- a/package.json +++ b/package.json @@ -2,10 +2,10 @@ "$schema": "https://json.schemastore.org/package.json", "name": "opencode", "description": "AI-powered development tool", - "version": "2.1.0", + "version": "2.1.2", "private": true, "type": "module", - "packageManager": "bun@1.4.2", + "packageManager": "bun@1.4.4", "scripts": { "dev": "bun script/dev-isolated.ts", "dev:live": "sh -c 'OPENCODE_TUI_CHANNEL=dev OPENCODE_PASSWORD=\"$(opencode service get password)\" exec bun run dev \"$@\" --server \"$(opencode service status)\"' --", @@ -50,7 +50,7 @@ "@effect/platform-node-shared": "4.0.1", "@effect/sql-sqlite-bun": "4.0.1", "@npmcli/arborist": "9.4.0", - "@types/bun": "1.4.0", + "@types/bun": "1.4.3", "@types/cross-spawn": "6.0.6", "@octokit/rest": "22.0.0", "@hono/standard-validator": "0.2.0", @@ -59,7 +59,6 @@ "@opentui/keymap": "0.6.6", "@opentui/solid": "0.6.6", "@tanstack/solid-virtual": "3.13.37", - "@shikijs/stream": "4.4.3", "@standard-schema/spec": "1.1.0", "ulid": "3.0.1", "@kobalte/core": "0.13.13", @@ -70,7 +69,8 @@ "@tsconfig/node22": "22.0.2", "@tsconfig/bun": "1.0.9", "@cloudflare/workers-types": "4.20251008.0", - "@pierre/diffs": "1.5.1", + "@pierre/diffs": "1.6.1", + "@pierre/highlights": "0.2.0", "@solid-primitives/event-listener": "2.4.6", "@solid-primitives/media": "2.3.6", "@solid-primitives/resize-observer": "2.2.0", @@ -164,7 +164,7 @@ "@effect/platform-node-shared": "catalog:", "@types/bun": "catalog:", "@types/node": "catalog:", - "bun-types": "1.4.2", + "bun-types": "1.4.4", "effect": "catalog:", "open": "11.0.4", "solid-js": "catalog:" diff --git a/packages/ai/package.json b/packages/ai/package.json index 69cff6648b47..686897ceb0b2 100644 --- a/packages/ai/package.json +++ b/packages/ai/package.json @@ -1,6 +1,6 @@ { "$schema": "https://json.schemastore.org/package.json", - "version": "2.1.0", + "version": "2.1.2", "name": "@opencode/ai", "type": "module", "license": "MIT", diff --git a/packages/ai/src/cache-policy.ts b/packages/ai/src/cache-policy.ts index 81ae0f4ed5c6..b00d8cc24807 100644 --- a/packages/ai/src/cache-policy.ts +++ b/packages/ai/src/cache-policy.ts @@ -107,8 +107,13 @@ const markSystemBoundaries = (system: LLMRequest["system"], hint: CacheHint, bud const lastIndexOfRole = (messages: ReadonlyArray, role: Message["role"]): number => messages.findLastIndex((m) => m.role === role) +type CacheablePart = Exclude + +const isCacheablePart = (part: ContentPart): part is CacheablePart => + part.type !== "compaction" && part.type !== "effort" + // Mark the last text part of `messages[index]`. If no text part exists, mark -// the last content part regardless of type — that's the breakpoint position +// the last cacheable content part — that's the breakpoint position // in tool-result-only messages too. const markMessageAt = ( messages: ReadonlyArray, @@ -120,11 +125,11 @@ const markMessageAt = ( const target = messages[index]! if (target.content.length === 0) return messages const lastTextIndex = target.content.findLastIndex((part) => part.type === "text") - const markAt = lastTextIndex >= 0 ? lastTextIndex : target.content.length - 1 - const existing = target.content[markAt]! - if (("cache" in existing && existing.cache) || budget.remaining === 0) return messages + const markAt = lastTextIndex >= 0 ? lastTextIndex : target.content.findLastIndex(isCacheablePart) + const existing = target.content[markAt] + if (!existing || !isCacheablePart(existing) || existing.cache || budget.remaining === 0) return messages budget.remaining -= 1 - const nextContent = target.content.map((part, i) => (i === markAt ? ({ ...part, cache: hint } as ContentPart) : part)) + const nextContent = target.content.map((part, i) => (i === markAt ? { ...existing, cache: hint } : part)) const next = new Message({ ...target, content: nextContent }) // Single pass over `messages`, substituting the one updated entry. Long // conversations call this on every request, so avoid `.map()` here — its @@ -164,7 +169,7 @@ const countHints = (request: LLMRequest) => (count, message) => count + message.content.reduce( - (contentCount, part) => contentCount + ("cache" in part && part.cache !== undefined ? 1 : 0), + (contentCount, part) => contentCount + (isCacheablePart(part) && part.cache !== undefined ? 1 : 0), 0, ), 0, diff --git a/packages/ai/src/experimental.ts b/packages/ai/src/experimental.ts index 4128d367e0d4..6e67322a546b 100644 --- a/packages/ai/src/experimental.ts +++ b/packages/ai/src/experimental.ts @@ -33,4 +33,5 @@ export type { RealtimeSession, RealtimeTransport, RealtimeTransportInput, + RealtimeTransportType, } from "./experimental/realtime.js" diff --git a/packages/ai/src/experimental/evaluation.ts b/packages/ai/src/experimental/evaluation.ts index 7c56d1ceac91..06d656c67b6c 100644 --- a/packages/ai/src/experimental/evaluation.ts +++ b/packages/ai/src/experimental/evaluation.ts @@ -108,14 +108,11 @@ export type EvaluationOptions = Record export interface EvaluationRoute { readonly id: string - readonly evaluate: ( - request: EvaluationRequestFor, - execute: Execute, - ) => Effect.Effect + evaluate(request: EvaluationRequestFor, execute: Execute): Effect.Effect } export class EvaluationModel { - declare protected readonly _Options: (options: Options) => Options + declare protected readonly _Options: Options readonly id: ModelID readonly provider: ProviderID readonly route: EvaluationRoute @@ -182,14 +179,14 @@ export type EvaluationRequestFor< } export type EvaluationRequestInput< - Model extends object = EvaluationModel, + Model extends EvaluationModel = EvaluationModel, Questions extends EvaluationQuestions = EvaluationQuestions, > = Omit[0], "model" | "questions" | "options" | "http"> & { readonly model: Model readonly questions: Questions readonly options?: NoInfer> readonly http?: HttpOptions.Input -} & (Model extends EvaluationModel> ? unknown : never) +} export class EvaluationRounding extends Schema.Class("Evaluation.Rounding")({ probabilityDecimals: Schema.optional(Schema.Int), @@ -208,26 +205,26 @@ export type EvaluationResponseFor = Omit< readonly answers: AnswersFor } -export function request( +export function request( input: EvaluationRequestInput, ): EvaluationRequestFor, Questions> export function request(input: EvaluationRequest): EvaluationRequest +export function request(input: EvaluationRequest | EvaluationRequestInput): EvaluationRequest export function request(input: EvaluationRequest | EvaluationRequestInput) { if (input instanceof EvaluationRequest) return input return new EvaluationRequest({ ...input, - model: input.model as unknown as EvaluationModel, http: HttpOptions.make(input.http), }) } -export function run( +export function run( input: EvaluationRequestInput, ): Effect.Effect, AIError, Service> export function run(input: EvaluationRequest): Effect.Effect export function run(input: EvaluationRequest | EvaluationRequestInput) { return Effect.try({ - try: () => (input instanceof EvaluationRequest ? input : request(input)), + try: () => request(input), catch: (cause) => new AIError({ reason: new InvalidRequestError({ @@ -235,11 +232,7 @@ export function run(input: EvaluationRequest | EvaluationRequestInput) { cause, }), }), - }).pipe( - Effect.flatMap((request) => - EvaluationClient.evaluate(request as EvaluationRequestFor), - ), - ) + }).pipe(Effect.flatMap(EvaluationClient.evaluate)) } export const Evaluation = { diff --git a/packages/ai/src/experimental/realtime.ts b/packages/ai/src/experimental/realtime.ts index 4af26e34701a..c5f5026a005f 100644 --- a/packages/ai/src/experimental/realtime.ts +++ b/packages/ai/src/experimental/realtime.ts @@ -1,7 +1,7 @@ import { Clock, Effect, Redacted, Schema } from "effect" import { HttpClientRequest } from "effect/http" import { RequestExecutor } from "../route/executor.js" -import { AIError, HttpContext, InvalidProviderOutputError, InvalidRequestError } from "../schema/index.js" +import { AIError, InvalidProviderOutputError, InvalidRequestError } from "../schema/index.js" import { encodeJson } from "../utils/json.js" const OPENAI_BASE_URL = "https://api.openai.com/v1" @@ -12,6 +12,8 @@ export type RealtimeTransportInput = | { readonly type: "webrtc"; readonly sdp: string; readonly dcid?: number } | { readonly type: "websocket" } +export type RealtimeTransportType = RealtimeTransportInput["type"] + export type RealtimeTransport = | { readonly type: "webrtc"; readonly sdp: string } | { readonly type: "websocket"; readonly url: string; readonly token: string; readonly expiresAt: number } @@ -25,11 +27,23 @@ export interface RealtimeConnectInput { readonly transport: RealtimeTransportInput } -const engines = [ - { family: "gpt-live", id: "openai-live", provider: "openai" }, - { family: "gpt-realtime", id: "openai-realtime", provider: "openai" }, - { family: "grok-voice", id: "xai-realtime", provider: "xai" }, - { family: "gemini-3.8-live", id: "gemini-live", provider: "google" }, +export const engines = [ + { family: "gpt-live", model: "gpt-live-1", id: "openai-live", provider: "openai", transports: ["webrtc"] }, + { + family: "gpt-realtime", + model: "gpt-realtime-2.1", + id: "openai-realtime", + provider: "openai", + transports: ["webrtc", "websocket"], + }, + { family: "grok-voice", model: "grok-voice-3.0", id: "xai-realtime", provider: "xai", transports: ["websocket"] }, + { + family: "gemini-3.8-live", + model: "gemini-3.8-live", + id: "gemini-live", + provider: "google", + transports: ["websocket"], + }, ] as const // Matches the bare family name (`gpt-realtime`) and its versions (`gpt-realtime-2.1`, `gpt-realtime-mini`). @@ -39,12 +53,12 @@ export const engine = (model: string) => export const connect = Effect.fn("Realtime.connect")(function* (input: RealtimeConnectInput) { const resolved = engine(input.model) if (!resolved) return yield* invalidRequest(`Unsupported voice model: ${input.model}`) + if (!resolved.transports.some((type) => type === input.transport.type)) + return yield* invalidRequest(`${input.model} only supports ${resolved.transports.join(", ")} transport`) const session = { ...input.session, model: input.model } switch (resolved.id) { case "openai-live": { - if (input.transport.type !== "webrtc") - return yield* invalidRequest("OpenAI GPT-Live only supports WebRTC transport") const data = yield* post( resolved.id, `${OPENAI_BASE_URL}/live/sessions`, @@ -56,7 +70,10 @@ export const connect = Effect.fn("Realtime.connect")(function* (input: RealtimeC } case "openai-realtime": { if (input.transport.type === "websocket") - return yield* clientSecret(resolved.id, OPENAI_BASE_URL, input.model, input.apiKey, session) + return yield* clientSecret(resolved.id, OPENAI_BASE_URL, input.model, input.apiKey, { + expires_after: { anchor: "created_at", seconds: 60 }, + session, + }) const form = new FormData() form.append("sdp", input.transport.sdp) form.append("session", encodeJson(session)) @@ -70,14 +87,11 @@ export const connect = Effect.fn("Realtime.connect")(function* (input: RealtimeC ) return { type: "webrtc" as const, sdp } } - case "xai-realtime": { - if (input.transport.type !== "websocket") - return yield* invalidRequest("xAI Grok Voice only supports WebSocket transport") - return yield* clientSecret(resolved.id, XAI_BASE_URL, input.model, input.apiKey) - } + case "xai-realtime": + return yield* clientSecret(resolved.id, XAI_BASE_URL, input.model, input.apiKey, { + expires_after: { seconds: 60 }, + }) case "gemini-live": { - if (input.transport.type !== "websocket") - return yield* invalidRequest("Google Gemini Live only supports WebSocket transport") const now = yield* Clock.currentTimeMillis const expiresAt = now + 60_000 const data = yield* post( @@ -104,6 +118,7 @@ export const connect = Effect.fn("Realtime.connect")(function* (input: RealtimeC }) export const Realtime = { + engines, engine, connect, } as const @@ -115,7 +130,7 @@ const LiveSessionResponse = Schema.fromJsonString( ) // xAI's ephemeral-token docs (https://docs.x.ai/developers/model-capabilities/audio/ephemeral-tokens) -// document the request body (`{ expires_after: { seconds } }`) but not the response payload; we assume +// document the request body (`{ expires_after: { seconds } }`, no `anchor` or `session`) but not the response payload; we assume // it mirrors OpenAI's `{ value, expires_at }` response shape. const ClientSecretResponse = Schema.fromJsonString( Schema.Struct({ @@ -152,7 +167,7 @@ const post = ( : HttpClientRequest.bodyText(encodeJson(body), "application/json"), ), ) - const http = new HttpContext({ url: res.request.url, status: res.status, headers: res.headers }) + const http = RequestExecutor.responseHttp(res) const fail = (message: string, cause?: unknown, raw?: string) => new AIError({ reason: new InvalidProviderOutputError({ route, message, body: raw, http, cause }) }) const text = yield* res.text.pipe(Effect.mapError((cause) => fail(`Failed to read ${route} response`, cause))) @@ -166,15 +181,9 @@ const clientSecret = ( baseURL: string, model: string, apiKey: Redacted.Redacted, - session?: Record, + body: Record, ) => - post( - route, - `${baseURL}/realtime/client_secrets`, - apiKey, - { expires_after: { seconds: 60 }, session }, - ClientSecretResponse, - ).pipe( + post(route, `${baseURL}/realtime/client_secrets`, apiKey, body, ClientSecretResponse).pipe( Effect.map( (data): RealtimeTransport => ({ type: "websocket", diff --git a/packages/ai/src/experimental/system-one.ts b/packages/ai/src/experimental/system-one.ts index 3e924f2810e4..e3be4e109cb0 100644 --- a/packages/ai/src/experimental/system-one.ts +++ b/packages/ai/src/experimental/system-one.ts @@ -11,13 +11,15 @@ import { type EvaluationOptions, } from "./evaluation.js" import { Auth, type Definition as AuthDefinition } from "../route/auth.js" +import { RequestExecutor } from "../route/executor.js" +import { JsonObject, ProviderShared } from "../protocols/shared.js" import { AIError, - HttpContext, HttpOptions, InvalidProviderOutputError, InvalidRequestError, ModelID, + ProviderMetadata, Usage, mergeJsonRecords, } from "../schema/index.js" @@ -51,7 +53,7 @@ const Request = Schema.StructWithRest( state: EvaluationInput, questions: Schema.Record(Schema.String, Question), }), - [Schema.Record(Schema.String, Schema.Any)], + [JsonObject], ) const Probability = Schema.Number.check(Schema.isBetween({ minimum: 0, maximum: 1 })) @@ -75,7 +77,7 @@ const NativeUsage = Schema.StructWithRest( input_tokens: Schema.optional(Schema.Number), output_tokens: Schema.optional(Schema.Number), }), - [Schema.Record(Schema.String, Schema.Unknown)], + [JsonObject], ) const Response = Schema.Struct({ model: Schema.String, @@ -83,7 +85,7 @@ const Response = Schema.Struct({ usage: Schema.optional(NativeUsage), id: Schema.optional(Schema.String), provider: Schema.optional(Schema.String), - provider_metadata: Schema.optional(Schema.Record(Schema.String, Schema.Record(Schema.String, Schema.Unknown))), + provider_metadata: Schema.optional(ProviderMetadata), }) export interface ModelInput { @@ -132,7 +134,7 @@ export const model = (cfg HttpClientRequest.bodyText(body, "application/json"), ), ) - const http = new HttpContext({ url: res.request.url, status: res.status, headers: res.headers }) + const http = RequestExecutor.responseHttp(res) const fail = (message: string, cause: unknown, body?: string) => new AIError({ reason: new InvalidProviderOutputError({ route: "system-one", message, body, http, cause }) }) const text = yield* res.text.pipe( @@ -182,10 +184,11 @@ export const model = (cfg ? new Usage({ inputTokens: data.usage.input_tokens, outputTokens: data.usage.output_tokens, - totalTokens: - data.usage.input_tokens === undefined && data.usage.output_tokens === undefined - ? undefined - : (data.usage.input_tokens ?? 0) + (data.usage.output_tokens ?? 0), + totalTokens: ProviderShared.totalTokens( + data.usage.input_tokens, + data.usage.output_tokens, + undefined, + ), providerMetadata: { [cfg.providerMetadataKey]: data.usage }, }) : undefined, diff --git a/packages/ai/src/generation.ts b/packages/ai/src/generation.ts index 6c4d88b0dcdc..6cb471ffde55 100644 --- a/packages/ai/src/generation.ts +++ b/packages/ai/src/generation.ts @@ -202,6 +202,6 @@ const timeoutError = (id: string, timeout: Duration.Duration) => new AIError({ reason: new TimeoutError({ message: `Generation ${id} did not finish within ${Duration.format(timeout)}`, - timeoutMs: Duration.toMillis(timeout), + timeoutMs: Math.min(Number.MAX_SAFE_INTEGER, Math.max(0, Math.ceil(Duration.toMillis(timeout)))), }), }) diff --git a/packages/ai/src/image.ts b/packages/ai/src/image.ts index b25efee811fc..be96446bf8fb 100644 --- a/packages/ai/src/image.ts +++ b/packages/ai/src/image.ts @@ -3,7 +3,14 @@ import { Generation, ProgressEvent, QueuedEvent, type AwaitOptions } from "./gen import { Media } from "./media.js" import { MediaModel, composeRoute, tryRequest } from "./media-model.js" import { MediaRoute } from "./route/media.js" -import { AIError, HttpOptions, MediaUsage, ProviderMetadata, type OpenString } from "./schema/index.js" +import { + AIError, + HttpOptions, + MediaUsage, + ProviderMetadata, + ProviderOptions, + type OpenString, +} from "./schema/index.js" import { ImageClient, Service } from "./image-client.js" // --------------------------------------------------------------------------- @@ -74,7 +81,7 @@ export class ImageRequest extends Schema.Class("Image.Request")({ aspectRatio: Schema.optional(ImageAspectRatio), seed: Schema.optional(Schema.Number), format: Schema.optional(Schema.String), - providerOptions: Schema.optional(Schema.Record(Schema.String, Schema.Unknown)), + providerOptions: Schema.optional(ProviderOptions), http: Schema.optional(HttpOptions), }) { declare protected readonly _ImageRequest: void diff --git a/packages/ai/src/llm.ts b/packages/ai/src/llm.ts index f28804927b1a..7a7059e1f787 100644 --- a/packages/ai/src/llm.ts +++ b/packages/ai/src/llm.ts @@ -37,19 +37,17 @@ export type RequestInput( - input: RequestInput, + input: RequestInput | LLMRequest, options?: StreamOptions, ): Effect.Effect -export function generate(input: LLMRequest, options?: StreamOptions): Effect.Effect export function generate(input: RequestInput | LLMRequest, options?: StreamOptions) { return requestEffect(input).pipe(Effect.flatMap((request) => LLMClient.generate(request, options))) } export function stream( - input: RequestInput, + input: RequestInput | LLMRequest, options?: StreamOptions, ): Stream.Stream -export function stream(input: LLMRequest, options?: StreamOptions): Stream.Stream export function stream(input: RequestInput | LLMRequest, options?: StreamOptions) { return Stream.unwrap(requestEffect(input).pipe(Effect.map((request) => LLMClient.stream(request, options)))) } @@ -130,10 +128,8 @@ const runGenerateObject = Effect.fn("LLM.generateObject")(function* ( toolChoice: ToolChoice.named(GENERATE_OBJECT_TOOL_NAME), }) const response = yield* LLMClient.generate(generateRequest) - const call = response.toolCalls.find( - (event) => LLMEvent.is.toolCall(event) && event.name === GENERATE_OBJECT_TOOL_NAME, - ) - if (!call || !LLMEvent.is.toolCall(call)) + const call = response.toolCalls.find((event) => event.name === GENERATE_OBJECT_TOOL_NAME) + if (!call) return yield* new AIError({ reason: new InvalidProviderOutputError({ message: `generateObject: model did not call the forced \`${GENERATE_OBJECT_TOOL_NAME}\` tool`, diff --git a/packages/ai/src/media-model.ts b/packages/ai/src/media-model.ts index 6e453ab7ad33..8d611564d740 100644 --- a/packages/ai/src/media-model.ts +++ b/packages/ai/src/media-model.ts @@ -2,7 +2,14 @@ import { Effect } from "effect" import { Endpoint } from "./route/endpoint.js" import { MediaRoute } from "./route/media.js" import type { MediaProtocol } from "./route/media-protocol.js" -import { AIError, HttpOptions, InvalidRequestError, ModelID, ProviderID } from "./schema/index.js" +import { + AIError, + HttpOptions, + InvalidRequestError, + ModelID, + ProviderID, + type ProviderOptions, +} from "./schema/index.js" /** * What every media model carries: ids, the configured route, and deployment `http` overlays. Modality classes @@ -27,7 +34,7 @@ export class MediaModel { } export namespace MediaModel { - export type Options = Record + export type Options = ProviderOptions export interface Input { readonly id: string | ModelID diff --git a/packages/ai/src/media.ts b/packages/ai/src/media.ts index 91c0d832c9eb..0de96d0d21e9 100644 --- a/packages/ai/src/media.ts +++ b/packages/ai/src/media.ts @@ -4,9 +4,9 @@ import { Effect, FileSystem, Schema, SchemaGetter } from "effect" import { Base64 } from "effect/encoding" import { HttpClientRequest } from "effect/http" import { ProviderID } from "./schema/ids.js" -import { AIError, HttpContext, InvalidProviderOutputError, InvalidRequestError } from "./schema/errors.js" +import { AIError, InvalidProviderOutputError, InvalidRequestError } from "./schema/errors.js" import { ProviderMetadata } from "./schema/options.js" -import { Service } from "./route/executor-service.js" +import { responseHttp, Service } from "./route/executor-service.js" import { detectMediaType, fileMediaType } from "./utils/media-type.js" export { detectMediaType } from "./utils/media-type.js" @@ -254,7 +254,7 @@ const download = Effect.fn("Media.download")(function* ( new AIError({ reason: new InvalidProviderOutputError({ message: `Failed to read media from ${source.url}`, - http: new HttpContext({ url: response.request.url, status: response.status, headers: response.headers }), + http: responseHttp(response), cause, }), }), diff --git a/packages/ai/src/promise.ts b/packages/ai/src/promise.ts index d4dd8dddc606..fe13b006cc06 100644 --- a/packages/ai/src/promise.ts +++ b/packages/ai/src/promise.ts @@ -4,9 +4,8 @@ import type { AwaitOptions, Event, Generation, Snapshot } from "./generation.js" import { Image, type ImageModel, type ImageRequest, type ImageRequestInput } from "./image.js" import { LLM } from "./index.js" import { Media } from "./media.js" -import { tryRequest } from "./media-model.js" import { RequestExecutor } from "./route/executor.js" -import { AIError, InvalidRequestError, LanguageModel, LLMRequest } from "./schema/index.js" +import { AIError, InvalidRequestError, LanguageModel, type LLMRequest } from "./schema/index.js" import type { RequestInput } from "./llm.js" import { Speech, type SpeechModel, type SpeechRequest, type SpeechRequestInput } from "./speech.js" import { @@ -93,9 +92,6 @@ export const make = (options: Options = {}) => { cancel: (options) => run(generation.cancel(), options), }) - const llmRequest = (input: RequestInput | LLMRequest) => - input instanceof LLMRequest ? Effect.succeed(input) : tryRequest(() => LLM.request(input)) - return { run, /** Decoded asset bytes, downloading `url` sources through the executor. */ @@ -137,9 +133,9 @@ export const make = (options: Options = {}) => { llm: { request: LLM.request, generate: (input: RequestInput | LLMRequest, options?: RunOptions) => - run(Effect.flatMap(llmRequest(input), LLM.generate), options), + run(LLM.generate(input), options), stream: (input: RequestInput | LLMRequest, options?: RunOptions) => - iterate(Stream.unwrap(Effect.map(llmRequest(input), LLM.stream)), options), + iterate(LLM.stream(input), options), }, image: { request: Image.request, diff --git a/packages/ai/src/protocols/alibaba-chat.ts b/packages/ai/src/protocols/alibaba-chat.ts index 7ad7ea35cc4f..910117da969a 100644 --- a/packages/ai/src/protocols/alibaba-chat.ts +++ b/packages/ai/src/protocols/alibaba-chat.ts @@ -16,7 +16,7 @@ const Options = Schema.Struct({ clearThinking: Schema.optional(Schema.Boolean), thinking: Schema.optional( Schema.Struct({ - type: Schema.declare<"adaptive" | "disabled" | (string & {})>(Schema.is(Schema.String)), + type: ProviderShared.knownString<"adaptive" | "disabled">(), }), ), toolStream: Schema.optional(Schema.Boolean), @@ -24,7 +24,7 @@ const Options = Schema.Struct({ repetitionPenalty: Schema.optional(Schema.Number), responseFormat: Schema.optional( Schema.Struct({ - type: Schema.declare<"text" | "json_object" | "json_schema" | (string & {})>(Schema.is(Schema.String)), + type: ProviderShared.knownString<"text" | "json_object" | "json_schema">(), json_schema: Schema.optional(JsonObject), }), ), @@ -32,9 +32,7 @@ const Options = Schema.Struct({ searchOptions: Schema.optional( Schema.Struct({ forced_search: Schema.optional(Schema.Boolean), - search_strategy: Schema.optional( - Schema.declare<"turbo" | "max" | "agent" | "agent_max" | (string & {})>(Schema.is(Schema.String)), - ), + search_strategy: Schema.optional(ProviderShared.knownString<"turbo" | "max" | "agent" | "agent_max">()), enable_search_extension: Schema.optional(Schema.Boolean), }), ), diff --git a/packages/ai/src/protocols/alibaba-messages.ts b/packages/ai/src/protocols/alibaba-messages.ts index 2dd99c020c15..1dec225f9287 100644 --- a/packages/ai/src/protocols/alibaba-messages.ts +++ b/packages/ai/src/protocols/alibaba-messages.ts @@ -9,7 +9,7 @@ const Options = Schema.Struct({ effort: Schema.optional(OpenResponsesOptions.ReasoningEffort), thinking: Schema.optional( Schema.Struct({ - type: Schema.declare<"enabled" | "disabled" | (string & {})>(Schema.is(Schema.String)), + type: ProviderShared.knownString<"enabled" | "disabled">(), budgetTokens: Schema.optional(Schema.Int), budget_tokens: Schema.optional(Schema.Int), }), diff --git a/packages/ai/src/protocols/anthropic-messages.ts b/packages/ai/src/protocols/anthropic-messages.ts index 60282befbc13..1e8aeb84f030 100644 --- a/packages/ai/src/protocols/anthropic-messages.ts +++ b/packages/ai/src/protocols/anthropic-messages.ts @@ -535,13 +535,13 @@ const providerMetadata = (key: string, metadata: Record): Provi const signatureFromMetadata = (metadata: ProviderMetadata | undefined, key: string): string | undefined => { const provider = metadata?.[key] - if (!ProviderShared.isRecord(provider)) return undefined + if (!provider) return undefined return typeof provider.signature === "string" ? provider.signature : undefined } const redactedDataFromMetadata = (metadata: ProviderMetadata | undefined, key: string): string | undefined => { const provider = metadata?.[key] - if (!ProviderShared.isRecord(provider)) return undefined + if (!provider) return undefined return typeof provider.redactedData === "string" ? provider.redactedData : undefined } @@ -617,7 +617,7 @@ const lowerServerToolResult = Effect.fnUntraced(function* (part: ToolResultPart, }) const fileIdFromMetadata = (metadata: MediaPart["metadata"]): string | undefined => { - if (!ProviderShared.isRecord(metadata)) return undefined + if (!metadata) return undefined const anthropic = metadata.anthropic if (ProviderShared.isRecord(anthropic)) { if (typeof anthropic.file_id === "string") return anthropic.file_id @@ -631,7 +631,7 @@ const fileIdFromMetadata = (metadata: MediaPart["metadata"]): string | undefined const transformationsFromMetadata = ( metadata: MediaPart["metadata"], ): AnthropicImageBlock["transformations"] | undefined => { - if (!ProviderShared.isRecord(metadata)) return undefined + if (!metadata) return undefined const anthropic = ProviderShared.isRecord(metadata.anthropic) ? metadata.anthropic : undefined const raw = anthropic?.transformations ?? metadata.transformations if (ProviderShared.isRecord(raw)) { @@ -644,17 +644,17 @@ const transformationsFromMetadata = ( } const documentTitleFromPart = (part: MediaPart): string | undefined => { - if (ProviderShared.isRecord(part.metadata)) { + if (part.metadata !== undefined) { const anthropic = part.metadata.anthropic if (ProviderShared.isRecord(anthropic) && typeof anthropic.title === "string") return anthropic.title if (typeof part.metadata.title === "string") return part.metadata.title } - if (typeof part.filename === "string" && part.filename.length > 0) return part.filename + if (part.filename && part.filename.length > 0) return part.filename return undefined } const documentContextFromMetadata = (metadata: MediaPart["metadata"]): string | undefined => { - if (!ProviderShared.isRecord(metadata)) return undefined + if (!metadata) return undefined const anthropic = ProviderShared.isRecord(metadata.anthropic) ? metadata.anthropic : undefined if (anthropic && typeof anthropic.context === "string") return anthropic.context if (typeof metadata.context === "string") return metadata.context @@ -662,7 +662,7 @@ const documentContextFromMetadata = (metadata: MediaPart["metadata"]): string | } const citationsFromMetadata = (metadata: MediaPart["metadata"]): AnthropicDocumentBlock["citations"] | undefined => { - if (!ProviderShared.isRecord(metadata)) return undefined + if (!metadata) return undefined const raw = ProviderShared.isRecord(metadata.anthropic) ? (metadata.anthropic.citations ?? metadata.citations) : metadata.citations @@ -940,7 +940,7 @@ const lowerMessages = Effect.fnUntraced(function* (request: LLMRequest, breakpoi content.push({ type: "redacted_thinking", data: redactedData }) continue } - if (typeof signature !== "string" || signature.trim().length === 0) { + if (!signature || signature.trim().length === 0) { if (part.text.trim().length === 0) continue if (!requireThinkingSignature(request)) { content.push({ type: "thinking", thinking: part.text, signature: "" }) @@ -1199,9 +1199,7 @@ const isServerToolResultType = (type: string): type is AnthropicServerToolResult const serverToolResultEvent = (block: AnthropicStreamBlock, providerMetadataKey: string): LLMEvent | undefined => { if (!block.type || !isServerToolResultType(block.type)) return undefined const errorPayload = - typeof block.content === "object" && block.content !== null && "type" in block.content - ? String((block.content as Record).type) - : "" + ProviderShared.isRecord(block.content) && "type" in block.content ? String(block.content.type) : "" const isError = errorPayload.endsWith("_tool_result_error") return LLMEvent.toolResult({ id: block.tool_use_id ?? "", diff --git a/packages/ai/src/protocols/bedrock-converse.ts b/packages/ai/src/protocols/bedrock-converse.ts index 2954615a00a7..69030465f60b 100644 --- a/packages/ai/src/protocols/bedrock-converse.ts +++ b/packages/ai/src/protocols/bedrock-converse.ts @@ -261,12 +261,12 @@ const providerMetadata = (key: string, metadata: Record): Provi const reasoningSignature = (part: ReasoningPart, providerMetadataKey: string) => { const metadata = part.providerMetadata?.[providerMetadataKey] if (part.encrypted !== undefined) return part.encrypted - if (ProviderShared.isRecord(metadata) && typeof metadata.signature === "string") return metadata.signature + if (typeof metadata?.signature === "string") return metadata.signature } const reasoningRedactedData = (part: ReasoningPart, providerMetadataKey: string) => { const metadata = part.providerMetadata?.[providerMetadataKey] - if (ProviderShared.isRecord(metadata) && typeof metadata.redactedData === "string") return metadata.redactedData + if (typeof metadata?.redactedData === "string") return metadata.redactedData } const removeEmptyToolInputKeys = (input: unknown): unknown => { @@ -806,6 +806,7 @@ const step = (state: ParserState, event: BedrockEvent) => return yield* new AIError({ reason: classifyProviderFailure({ message, + status: event.exception.details.originalStatusCode, rawBody: body, }), }) @@ -859,7 +860,7 @@ export const protocol = Protocol.make({ from: fromRequest, }, stream: { - event: BedrockEvent, + event: BedrockEventStream.Frame.pipe(Schema.decodeTo(BedrockEvent)), initial: (request) => ({ providerMetadataKey: request.model.route.providerMetadataKey ?? String(request.model.provider), tools: ToolStream.empty(), diff --git a/packages/ai/src/protocols/bedrock-event-stream.ts b/packages/ai/src/protocols/bedrock-event-stream.ts index af81da859c2b..6a385b442be8 100644 --- a/packages/ai/src/protocols/bedrock-event-stream.ts +++ b/packages/ai/src/protocols/bedrock-event-stream.ts @@ -1,6 +1,6 @@ import { EventStreamCodec } from "@smithy/eventstream-codec" import { fromUtf8, toUtf8 } from "@smithy/util-utf8" -import { Effect, Stream } from "effect" +import { Effect, Schema, Stream } from "effect" import { Base64 } from "effect/encoding" import { AIError, AIErrorReason, InvalidProviderOutputError } from "../schema/index.js" import { Framing } from "../route/framing.js" @@ -23,6 +23,9 @@ interface FrameBufferState { const initialFrameBuffer: FrameBufferState = { buffer: new Uint8Array(0), offset: 0 } +export const Frame = Schema.StructWithRest(Schema.Struct({ rawBody: Schema.String }), [ProviderShared.JsonObject]) +export type Frame = typeof Frame.Type + type FrameInput = { readonly _tag: "Chunk"; readonly bytes: Uint8Array } | { readonly _tag: "End" } const endOfStream: FrameInput = { _tag: "End" } @@ -55,7 +58,7 @@ const consumeFrames = (route: string) => (state: FrameBufferState, input: FrameI } let cursor = appendChunk(state, input.bytes) - const out: object[] = [] + const out: Frame[] = [] while (cursor.buffer.length - cursor.offset >= 4) { const view = cursor.buffer.subarray(cursor.offset) const totalLength = new DataView(view.buffer, view.byteOffset, view.byteLength).getUint32(0, false) @@ -100,7 +103,7 @@ const consumeFrames = (route: string) => (state: FrameBufferState, input: FrameI // before handing the object to the chunk schema. JSON decode goes // through the shared Schema-driven codec to satisfy the package rule // against ad-hoc `JSON.parse` calls. - const parsed = (yield* ProviderShared.parseJson( + const parsed = yield* ProviderShared.parseJson( route, payload, "Failed to parse Bedrock Converse event-stream payload", @@ -111,12 +114,12 @@ const consumeFrames = (route: string) => (state: FrameBufferState, input: FrameI reason: AIErrorReason.make({ ...error.reason, message: error.message, cause: error.reason.cause, body }), }), ), - )) as Record - delete parsed.p + ) + if (!ProviderShared.isRecord(parsed)) + return yield* ProviderShared.eventError(route, "Failed to parse Bedrock Converse event-stream payload", body) + const { p: _, ...event } = parsed out.push({ - ...(messageType === "exception" - ? { exception: { type: eventType, details: parsed } } - : { [eventType]: parsed }), + ...(messageType === "exception" ? { exception: { type: eventType, details: event } } : { [eventType]: event }), rawBody: body, }) } @@ -129,9 +132,9 @@ const consumeFrames = (route: string) => (state: FrameBufferState, input: FrameI * under its `:event-type` header so the chunk schema can match the JSON * payload directly. */ -export const framing = (route: string): Framing.Definition => ({ +export const framing = (route: string): Framing.Definition => ({ id: "aws-event-stream", - body: (frame) => ("rawBody" in frame && typeof frame.rawBody === "string" ? frame.rawBody : undefined), + body: (frame) => frame.rawBody, frame: (bytes) => bytes.pipe( Stream.map((bytes): FrameInput => ({ _tag: "Chunk", bytes })), diff --git a/packages/ai/src/protocols/cohere-chat.ts b/packages/ai/src/protocols/cohere-chat.ts index 3b1756823c9b..876d5e297660 100644 --- a/packages/ai/src/protocols/cohere-chat.ts +++ b/packages/ai/src/protocols/cohere-chat.ts @@ -4,7 +4,7 @@ import { Endpoint } from "../route/endpoint.js" import { Framing } from "../route/framing.js" import { Protocol } from "../route/protocol.js" import { LLMEvent, Usage, type FinishReasonDetails, type LLMRequest } from "../schema/index.js" -import { ProviderShared } from "./shared.js" +import { JsonObject, ProviderShared } from "./shared.js" import { Lifecycle } from "./utils/lifecycle.js" import { ToolStream } from "./utils/tool-stream.js" @@ -49,7 +49,7 @@ const Body = Schema.Struct({ function: Schema.Struct({ name: Schema.String, description: Schema.optional(Schema.String), - parameters: Schema.Unknown, + parameters: JsonObject, }), }), ), diff --git a/packages/ai/src/protocols/gemini.ts b/packages/ai/src/protocols/gemini.ts index 20d4fbfc2d41..5bcf833a3962 100644 --- a/packages/ai/src/protocols/gemini.ts +++ b/packages/ai/src/protocols/gemini.ts @@ -248,8 +248,17 @@ const GeminiPromptFeedback = Schema.StructWithRest( ) type GeminiPromptFeedback = Schema.Schema.Type +const GeminiError = Schema.StructWithRest( + Schema.Struct({ + code: lenient(Schema.Number), + message: lenient(Schema.String), + status: lenient(Schema.String), + }), + [Schema.Record(Schema.String, Schema.Unknown)], +) + const GeminiEvent = Schema.Struct({ - error: Schema.optional(Schema.Unknown), + error: Schema.optional(GeminiError), candidates: optionalNull(Schema.Array(GeminiCandidate)), promptFeedback: optionalNull(GeminiPromptFeedback), usageMetadata: optionalNull(GeminiUsage), @@ -302,9 +311,7 @@ const providerMetadata = (key: string, metadata: Record): Provi const thoughtSignature = (metadata: ProviderMetadata | undefined, key: string) => { const value = metadata?.[key] - return ProviderShared.isRecord(value) && typeof value.thoughtSignature === "string" - ? value.thoughtSignature - : undefined + return typeof value?.thoughtSignature === "string" ? value.thoughtSignature : undefined } const lowerToolCall = (part: ToolCallPart, omitIds: boolean, metadataKey: string) => ({ @@ -637,27 +644,18 @@ const step = (state: ParserState, event: GeminiEvent | NextRequest) => { }, [], ] as const) - if (ProviderShared.isRecord(event.error)) { + if (event.error) { const body = ProviderShared.encodeJson(event) return Effect.fail( new AIError({ reason: classifyProviderFailure({ - message: - typeof event.error.message === "string" && event.error.message.length > 0 - ? event.error.message - : typeof event.error.status === "string" && event.error.status.length > 0 - ? event.error.status - : "Gemini provider error", - status: typeof event.error.code === "number" ? event.error.code : undefined, + message: event.error.message || event.error.status || "Gemini provider error", + status: event.error.code, rawBody: body, }), }), ) } - if ("error" in event) - return Effect.fail( - ProviderShared.eventError(state.route, `Invalid ${state.route} stream event`, ProviderShared.encodeJson(event)), - ) const candidate = event.candidates?.[0] const nextState = { ...state, diff --git a/packages/ai/src/protocols/google-interactions.ts b/packages/ai/src/protocols/google-interactions.ts index 836177fdeb21..dd74422c06dc 100644 --- a/packages/ai/src/protocols/google-interactions.ts +++ b/packages/ai/src/protocols/google-interactions.ts @@ -132,13 +132,22 @@ const OutputStep = Schema.Struct({ content: Schema.optional(Schema.Array(Schema.Struct({ type: Schema.String, text: Schema.optional(Schema.String) }))), }) type OutputStep = typeof OutputStep.Type +const KNOWN_DELTA_TYPES = new Set(["text", "arguments_delta", "thought_signature", "thought_summary"]) +const UnknownDelta = Schema.Struct({ + type: Schema.String.check(Schema.makeFilter((type) => !KNOWN_DELTA_TYPES.has(type))), +}).pipe( + Schema.decodeTo(Schema.Struct({ type: Schema.Literal("unknown"), rawType: Schema.String }), { + decode: SchemaGetter.transform((delta) => ({ type: "unknown" as const, rawType: delta.type })), + encode: SchemaGetter.forbidden(() => "Unknown Google Interactions deltas are decode-only"), + }), +) const Delta = Schema.Union([ Schema.Struct({ type: Schema.Literal("text"), text: Schema.String }), Schema.Struct({ type: Schema.Literal("arguments_delta"), arguments: Schema.String }), Schema.Struct({ type: Schema.Literal("thought_signature"), signature: Schema.String }), Schema.Struct({ type: Schema.Literal("thought_summary"), content: Text }), // Unknown output modalities must fail explicitly rather than disappearing from a successful response. - Schema.Struct({ type: Schema.String }), + UnknownDelta, ]) const Interaction = Schema.StructWithRest( Schema.Struct({ @@ -222,9 +231,7 @@ const mediaContent = Effect.fnUntraced(function* (asset: Media.Asset) { const signature = (metadata: ProviderMetadata | undefined, key: string) => { const value = metadata?.[key] - return ProviderShared.isRecord(value) && typeof value.interactionSignature === "string" - ? value.interactionSignature - : undefined + return typeof value?.interactionSignature === "string" ? value.interactionSignature : undefined } const lowerMessages = Effect.fnUntraced(function* (request: LLMRequest) { @@ -430,17 +437,17 @@ const onDelta = Effect.fnUntraced(function* ( if (!step) return yield* ProviderShared.eventError(ADAPTER, "Interactions delta without step.start", encodeJson(delta)) const events: LLMEvent[] = [] - if (delta.type === "text" && "text" in delta && step.type === "model_output") + if (delta.type === "text" && step.type === "model_output") return [ { ...state, lifecycle: Lifecycle.textDelta(state.lifecycle, events, String(index), delta.text) }, events, ] satisfies StepResult - if (delta.type === "thought_summary" && "content" in delta && step.type === "thought") + if (delta.type === "thought_summary" && step.type === "thought") return [ { ...state, lifecycle: Lifecycle.reasoningDelta(state.lifecycle, events, String(index), delta.content.text) }, events, ] satisfies StepResult - if (delta.type === "thought_signature" && "signature" in delta) { + if (delta.type === "thought_signature") { const next = { ...step, signature: delta.signature } const tool = state.tools[index] return [ @@ -452,7 +459,7 @@ const onDelta = Effect.fnUntraced(function* ( events, ] satisfies StepResult } - if (delta.type === "arguments_delta" && "arguments" in delta && step.type === "function_call") { + if (delta.type === "arguments_delta" && step.type === "function_call") { const result = ToolStream.appendExisting( ADAPTER, state.tools, @@ -463,7 +470,11 @@ const onDelta = Effect.fnUntraced(function* ( if (ToolStream.isError(result)) return yield* result return [{ ...state, tools: result.tools }, result.events] satisfies StepResult } - return yield* ProviderShared.eventError(ADAPTER, `Unsupported Interactions delta: ${delta.type}`, encodeJson(delta)) + return yield* ProviderShared.eventError( + ADAPTER, + `Unsupported Interactions delta: ${delta.type === "unknown" ? delta.rawType : delta.type}`, + encodeJson(delta), + ) }) const onStop = Effect.fnUntraced(function* (state: ParserState, index: number) { diff --git a/packages/ai/src/protocols/mistral-chat.ts b/packages/ai/src/protocols/mistral-chat.ts index 12cdfc91f8f2..fe573bb7e5ee 100644 --- a/packages/ai/src/protocols/mistral-chat.ts +++ b/packages/ai/src/protocols/mistral-chat.ts @@ -388,9 +388,6 @@ export const fromRequest = Effect.fn("MistralChat.fromRequest")(function* (reque const options = yield* ProviderShared.validateWith(Schema.decodeUnknownEffect(MistralOptions))( request.providerOptions ?? {}, ) - const selected = request.toolChoice?.type === "tool" ? request.toolChoice.name : undefined - if (request.toolChoice?.type === "tool" && !selected) - return yield* ProviderShared.invalidRequest("Mistral Chat tool choice requires a tool name") if (options.reasoningEffort !== undefined && options.promptMode !== undefined) return yield* ProviderShared.invalidRequest( "Mistral Chat reasoningEffort and promptMode provider options are mutually exclusive", @@ -505,7 +502,7 @@ const thinkingUnits = (value: unknown): ReadonlyArray => { } const thinkingText = (thinking: ReadonlyArray) => - thinking.flatMap((unit) => (typeof unit.text === "string" ? [unit.text] : [])).join("") + thinking.flatMap((unit) => (unit.text !== undefined ? [unit.text] : [])).join("") const thinkingMetadata = (thinking: MistralThinkingContent) => ({ mistral: { thinking } }) diff --git a/packages/ai/src/protocols/open-responses.ts b/packages/ai/src/protocols/open-responses.ts index 5074d9c3541f..ca5be36690a2 100644 --- a/packages/ai/src/protocols/open-responses.ts +++ b/packages/ai/src/protocols/open-responses.ts @@ -476,7 +476,7 @@ export const lowerToolChoice = (protocolName: string, toolChoice: NonNullable { const metadata = providerMetadata?.[providerMetadataKey] - if (!ProviderShared.isRecord(metadata) || typeof metadata.itemId !== "string") return undefined + if (!metadata || typeof metadata.itemId !== "string") return undefined const separator = metadata.itemId.indexOf("_") return separator > 0 && separator < metadata.itemId.length - 1 ? metadata.itemId : undefined } @@ -495,7 +495,7 @@ const lowerToolCall = (part: ToolCallPart, providerMetadataKey: string): OpenRes const lowerReasoning = (part: ReasoningPart, providerMetadataKey: string): OpenResponsesReasoningInput | undefined => { const metadata = part.providerMetadata?.[providerMetadataKey] - if (!ProviderShared.isRecord(metadata)) return undefined + if (!metadata) return undefined const id = itemID(part.providerMetadata, providerMetadataKey) const encryptedContent = typeof metadata.reasoningEncryptedContent === "string" || metadata.reasoningEncryptedContent === null @@ -1021,7 +1021,7 @@ export const onReasoningDelta = (state: ParserState, event: Event, itemID: strin // as a single delta unless that summary index already streamed one. export const onReasoningDone = (state: ParserState, event: Event, itemID: string): StepResult => { const item = state.reasoningItems[itemID] - if (!item || typeof event.text !== "string") return [state, NO_EVENTS] + if (!item || event.text === undefined) return [state, NO_EVENTS] const index = event.summary_index ?? 0 if (item.deltaIndexes.has(index)) return [state, NO_EVENTS] return onReasoningDelta(state, { ...event, delta: event.text }, itemID) diff --git a/packages/ai/src/protocols/openai-chat.ts b/packages/ai/src/protocols/openai-chat.ts index 6440a78ed016..1b80e93739a1 100644 --- a/packages/ai/src/protocols/openai-chat.ts +++ b/packages/ai/src/protocols/openai-chat.ts @@ -25,7 +25,7 @@ import { import { classifyProviderFailure } from "../provider-error.js" import { isRecord, JsonObject, optionalArray, optionalNull, ProviderShared } from "./shared.js" import { cacheControl } from "./utils/cache.js" -import { OpenAIOptions } from "./utils/openai-options.js" +import { OpenResponsesOptions } from "./utils/open-responses-options.js" import { Lifecycle } from "./utils/lifecycle.js" import { ToolStream } from "./utils/tool-stream.js" @@ -181,7 +181,7 @@ export const bodyFields = { stream_options: Schema.optional(Schema.Struct({ include_usage: Schema.Boolean })), store: Schema.optional(Schema.Boolean), prompt_cache_key: Schema.optional(Schema.String), - reasoning_effort: Schema.optional(OpenAIOptions.OpenAIReasoningEffort), + reasoning_effort: Schema.optional(OpenResponsesOptions.ReasoningEffort), tool_stream: Schema.optional(Schema.Boolean), max_completion_tokens: Schema.optional(Schema.Number), max_tokens: Schema.optional(Schema.Number), @@ -757,7 +757,7 @@ const detectZaiToolStream = (provider: string, baseURL: string | undefined, mode } const lowerOptions = (request: LLMRequest, supportsStore: boolean) => { - const options = OpenAIOptions.resolve(request) + const options = OpenResponsesOptions.resolve(request) // Default off: strict providers 400 on unknown body fields, so only send // the key where compatibility explicitly allows it. Header-based affinity // (x-session-affinity, x-grok-conv-id, ...) is unaffected. @@ -1028,10 +1028,9 @@ const step = (state: ParserState, event: OpenAIChatEvent) => const choice = event.choices?.[0] // Moonshot (and a few other OpenAI-compatible providers) attach usage to // `choice.usage` instead of the top-level `usage` field. - const choiceUsage = (choice as unknown as { usage?: OpenAIChatEvent["usage"] })?.usage const usage = mapUsage(event.usage, state.providerMetadataKey) ?? - (choiceUsage ? mapUsage(choiceUsage, state.providerMetadataKey) : undefined) ?? + mapUsage(choice?.usage, state.providerMetadataKey) ?? state.usage const rawFinishReason = choice?.finish_reason const finishReason = rawFinishReason diff --git a/packages/ai/src/protocols/utils/open-responses-options.ts b/packages/ai/src/protocols/utils/open-responses-options.ts index fbd0f0dc5cfd..8973e44531ac 100644 --- a/packages/ai/src/protocols/utils/open-responses-options.ts +++ b/packages/ai/src/protocols/utils/open-responses-options.ts @@ -1,15 +1,12 @@ import { Schema } from "effect" import { ReasoningEffort, ReasoningEfforts, type LLMRequest } from "../../schema/index.js" -import { lenient } from "../shared.js" +import { lenient, ProviderShared } from "../shared.js" export { ReasoningEffort, ReasoningEfforts } export const TextVerbosities = ["low", "medium", "high"] as const export type TextVerbosity = (typeof TextVerbosities)[number] | (string & {}) -export const TextVerbosity = Schema.declare( - (value): value is TextVerbosity => typeof value === "string", - { title: "TextVerbosity" }, -) +export const TextVerbosity = ProviderShared.knownString<(typeof TextVerbosities)[number]>() export const ResponseIncludables = [ "file_search_call.results", @@ -25,18 +22,13 @@ export type ResponseIncludable = (typeof ResponseIncludables)[number] | (string export const ServiceTiers = ["auto", "default", "flex", "priority"] as const export type ServiceTier = (typeof ServiceTiers)[number] | (string & {}) -export const ServiceTier = Schema.declare((value): value is ServiceTier => typeof value === "string", { - title: "ServiceTier", -}) +export const ServiceTier = ProviderShared.knownString<(typeof ServiceTiers)[number]>() export const Truncations = ["auto", "disabled"] as const export type Truncation = (typeof Truncations)[number] export const TextVerbositySchema = TextVerbosity -export const ResponseIncludableSchema = Schema.declare( - (value): value is ResponseIncludable => typeof value === "string", - { title: "ResponseIncludable" }, -) +export const ResponseIncludableSchema = ProviderShared.knownString<(typeof ResponseIncludables)[number]>() export const ServiceTierSchema = ServiceTier export const TruncationSchema = Schema.Literals(Truncations) diff --git a/packages/ai/src/protocols/utils/openai-options.ts b/packages/ai/src/protocols/utils/openai-options.ts deleted file mode 100644 index 00f6592006ed..000000000000 --- a/packages/ai/src/protocols/utils/openai-options.ts +++ /dev/null @@ -1,25 +0,0 @@ -import { OpenResponsesOptions } from "./open-responses-options.js" - -export const OpenAIReasoningEfforts = OpenResponsesOptions.ReasoningEfforts -export type OpenAIReasoningEffort = OpenResponsesOptions.ReasoningEffort -export const OpenAITextVerbosities = OpenResponsesOptions.TextVerbosities -export type OpenAITextVerbosity = OpenResponsesOptions.TextVerbosity - -// Mirrors OpenAI's `ResponseIncludable` union from the official SDK. Keep this -// in lockstep with `openai-node/src/resources/responses/responses.ts`. -export const OpenAIResponseIncludables = OpenResponsesOptions.ResponseIncludables -export type OpenAIResponseIncludable = OpenResponsesOptions.ResponseIncludable -// Mirrors OpenAI's `ServiceTier` union from the official SDK. -export const OpenAIServiceTiers = [...OpenResponsesOptions.ServiceTiers, "scale", "fast", "ultrafast"] as const -export type OpenAIServiceTier = (typeof OpenAIServiceTiers)[number] | (string & {}) - -export const OpenAIReasoningEffort = OpenResponsesOptions.ReasoningEffort -export const OpenAITextVerbosity = OpenResponsesOptions.TextVerbosity -export const OpenAIResponseIncludable = OpenResponsesOptions.ResponseIncludableSchema -export const OpenAIServiceTier = OpenResponsesOptions.ServiceTierSchema - -export const isReasoningEffort = (effort: unknown): effort is OpenAIReasoningEffort => typeof effort === "string" - -export const resolve = OpenResponsesOptions.resolve - -export * as OpenAIOptions from "./openai-options.js" diff --git a/packages/ai/src/protocols/utils/partial-json.ts b/packages/ai/src/protocols/utils/partial-json.ts index 345a19e4dcaa..d2567ba7dd27 100644 --- a/packages/ai/src/protocols/utils/partial-json.ts +++ b/packages/ai/src/protocols/utils/partial-json.ts @@ -26,13 +26,14 @@ */ import { Schema } from "effect" +import { decodeJson } from "../../utils/json.js" import { Allow } from "./partial-json-options.js" export * from "./partial-json-options.js" export class PartialJSON extends Error {} export class MalformedJSON extends Error {} -const decodeJson = Schema.decodeUnknownSync(Schema.fromJsonString(Schema.Unknown)) +const decodeJsonString = Schema.decodeUnknownSync(Schema.fromJsonString(Schema.String)) /** Parse complete or incomplete JSON, restricted by the supplied partial-value flags. */ export function parseJSON(jsonString: string, allowPartial = Allow.ALL): unknown { @@ -175,16 +176,16 @@ const _parseJSON = (jsonString: string, allow: number) => { } if (jsonString.charAt(index) === '"') { try { - return decodeJson(jsonString.substring(start, ++index - Number(escape))) as string + return decodeJsonString(jsonString.substring(start, ++index - Number(escape))) } catch (error) { throwMalformedError(String(error)) } } if (Allow.STR & allow) { try { - return decodeJson(`${jsonString.substring(start, index - Number(escape))}"`) as string + return decodeJsonString(`${jsonString.substring(start, index - Number(escape))}"`) } catch { - return decodeJson(`${jsonString.substring(start, jsonString.lastIndexOf("\\"))}"`) as string + return decodeJsonString(`${jsonString.substring(start, jsonString.lastIndexOf("\\"))}"`) } } return markPartialJSON("Unterminated string literal") diff --git a/packages/ai/src/protocols/utils/tool-schema.ts b/packages/ai/src/protocols/utils/tool-schema.ts index 69b6a6c78693..b2525eb93d5f 100644 --- a/packages/ai/src/protocols/utils/tool-schema.ts +++ b/packages/ai/src/protocols/utils/tool-schema.ts @@ -43,9 +43,6 @@ const moonshot = (schema: JsonSchema): JsonSchema => { return isRecord(projected) ? projected : {} } -const openAI = (schema: JsonSchema): JsonSchema => schema -const responses = openAI - const gemini = GeminiJsonSchema.normalize const MODEL_NAMES = [ @@ -106,7 +103,5 @@ const tools = (entries: LLMRequest["tools"], model: LanguageModel): LLMRequest[" export const ToolSchemaProjection = { gemini, moonshot, - openAI, - responses, tools, } as const diff --git a/packages/ai/src/protocols/venice-chat.ts b/packages/ai/src/protocols/venice-chat.ts index 01bf830ae1c0..cd2c1b36ecfe 100644 --- a/packages/ai/src/protocols/venice-chat.ts +++ b/packages/ai/src/protocols/venice-chat.ts @@ -87,7 +87,7 @@ const Assistant = Schema.StructWithRest( tool_calls: optionalArray(ToolCall), thought_signature: Schema.optional(Schema.String), }), - [Schema.Record(Schema.String, Schema.Unknown)], + [JsonObject], ) const Messages = Schema.Array( Schema.Union([ @@ -130,11 +130,11 @@ const Delta = Schema.StructWithRest( tool_calls: optionalNull(Schema.Array(ToolDelta)), thought_signature: optionalNull(Schema.String), }), - [Schema.Record(Schema.String, Schema.Unknown)], + [JsonObject], ) const Choice = Schema.StructWithRest( Schema.Struct({ ...OpenAIChat.OpenAIChatChoice.schema.fields, delta: optionalNull(Delta) }), - [Schema.Record(Schema.String, Schema.Unknown)], + [JsonObject], ) const Usage = Schema.StructWithRest( Schema.Struct({ @@ -146,11 +146,11 @@ const Usage = Schema.StructWithRest( cache_write_tokens: optionalNull(Schema.Number), cache_creation_input_tokens: optionalNull(Schema.Number), }), - [Schema.Record(Schema.String, Schema.Unknown)], + [JsonObject], ), ), }), - [Schema.Record(Schema.String, Schema.Unknown)], + [JsonObject], ) const Event = Schema.StructWithRest( Schema.Struct({ @@ -165,7 +165,7 @@ const Event = Schema.StructWithRest( }), ), }), - [Schema.Record(Schema.String, Schema.Unknown)], + [JsonObject], ) interface State { diff --git a/packages/ai/src/protocols/xai-images.ts b/packages/ai/src/protocols/xai-images.ts index 438b93a701ef..8011c0ba323e 100644 --- a/packages/ai/src/protocols/xai-images.ts +++ b/packages/ai/src/protocols/xai-images.ts @@ -5,7 +5,7 @@ import { Media } from "../media.js" import { MediaProtocol } from "../route/media-protocol.js" import { MediaRoute } from "../route/media.js" import { mergeJsonRecords, type OpenString } from "../schema/index.js" -import { ProviderShared, optionalNull } from "./shared.js" +import { JsonObject, ProviderShared, lenient, optionalNull } from "./shared.js" import { MediaInput } from "./utils/media-input.js" const route = MediaProtocol.identity({ id: "xai-images", name: "xAI Images", provider: "xai" }) @@ -39,7 +39,7 @@ const XAIImageResponse = Schema.Struct({ mime_type: optionalNull(Schema.String), }), ), - usage: Schema.optional(Schema.Unknown), + usage: lenient(JsonObject), }) // --------------------------------------------------------------------------- @@ -100,12 +100,11 @@ const decodeResponse = Effect.fn("XAIImages.decodeResponse")(function* ( }), ) if (images.length === 0) return yield* output.invalid(`${route.name} returned no images`) - const usage = ProviderShared.isRecord(decoded.usage) ? decoded.usage : undefined // xAI reports a USD cost (`cost_in_usd_ticks`) rather than tokens, seconds, or credits; the raw record stays in // provider metadata. return new ImageResponse({ images, - providerMetadata: usage === undefined ? undefined : { xai: { usage } }, + providerMetadata: decoded.usage === undefined ? undefined : { xai: { usage: decoded.usage } }, }) }) diff --git a/packages/ai/src/provider-error.ts b/packages/ai/src/provider-error.ts index 722728c3a031..286a5ae41f28 100644 --- a/packages/ai/src/provider-error.ts +++ b/packages/ai/src/provider-error.ts @@ -12,6 +12,8 @@ import { type HttpContext, type HttpRateLimitDetails, } from "./schema/index.js" +import { Json } from "./utils/json.js" +import { isRecord } from "./utils/record.js" const patterns = [ /prompt is too long/i, @@ -115,7 +117,7 @@ export const isRetryable = (error: AIError) => { } } -const decodeJson = Schema.decodeUnknownOption(Schema.fromJsonString(Schema.Unknown)) +const decodeJson = Schema.decodeUnknownOption(Json) // OpenCode Zen reports account caps as typed 429/402 errors that are not throttles. const QUOTA_CODES = new Set([ "insufficient_quota", @@ -263,7 +265,7 @@ export function classifyProviderFailure(input: ProviderFailure): AIError["reason codes.some((code) => QUOTA_CODES.has(code)) || (input.status === 429 && QUOTA_TEXT.test(text)) ) - return new QuotaExceededError(details) + return new QuotaExceededError({ ...details, retryAfterMs: input.retryAfterMs }) if ( input.status === 401 || input.status === 403 || @@ -333,7 +335,3 @@ function providerCodes(value: unknown) { exception?.type, ].filter((value): value is string => typeof value === "string") } - -function isRecord(value: unknown): value is Record { - return typeof value === "object" && value !== null -} diff --git a/packages/ai/src/providers/cloudflare-ai-gateway.ts b/packages/ai/src/providers/cloudflare-ai-gateway.ts index d5beba1bc12e..9ed255f992a3 100644 --- a/packages/ai/src/providers/cloudflare-ai-gateway.ts +++ b/packages/ai/src/providers/cloudflare-ai-gateway.ts @@ -1,10 +1,11 @@ -import { Schema, type Config, type Redacted } from "effect" +import type { Config, Redacted } from "effect" import type { ProviderPackage } from "../provider-package.js" import { AnthropicMessages, OpenAIChat, OpenAIResponses } from "../protocols/index.js" import { Auth } from "../route/auth.js" import type { AtLeastOne, ProviderAuthOption } from "../route/auth-options.js" import type { RouteDefaultsInput } from "../route/client.js" import { ProviderConfigurationError, ProviderID, type ModelID } from "../schema/index.js" +import { encodeJson } from "../utils/json.js" import type { OpenAIProviderOptionsInput } from "./openai-options.js" export const id = ProviderID.make("cloudflare-ai-gateway") @@ -86,9 +87,7 @@ const credential = (input: LanguageModelOptions) => .orElse(Auth.config(authEnvVars[1])) const headers = (input: LanguageModelOptions) => ({ - ...(input.metadata === undefined - ? {} - : { "cf-aig-metadata": Schema.encodeSync(Schema.fromJsonString(Schema.Unknown))(input.metadata) }), + ...(input.metadata === undefined ? {} : { "cf-aig-metadata": encodeJson(input.metadata) }), ...(input.cacheTtl === undefined ? {} : { "cf-aig-cache-ttl": String(input.cacheTtl) }), ...(input.cacheKey === undefined ? {} : { "cf-aig-cache-key": input.cacheKey }), ...(input.skipCache === undefined ? {} : { "cf-aig-skip-cache": String(input.skipCache) }), diff --git a/packages/ai/src/providers/openai-options.ts b/packages/ai/src/providers/openai-options.ts index 0e1bfdf51e94..8b248afe1c21 100644 --- a/packages/ai/src/providers/openai-options.ts +++ b/packages/ai/src/providers/openai-options.ts @@ -1,9 +1,12 @@ import { mergeProviderOptions, type ProviderOptions } from "../schema/index.js" -import type { OpenAIServiceTier } from "../protocols/utils/openai-options.js" -import type { Options } from "../protocols/utils/open-responses-options.js" +import type { Options, ResponseIncludable, ServiceTier } from "../protocols/utils/open-responses-options.js" import type { ContextManagement } from "../protocols/openai-responses.js" -export type { OpenAIResponseIncludable, OpenAIServiceTier } from "../protocols/utils/openai-options.js" +// Mirrors OpenAI's `ResponseIncludable` union from the official SDK. Keep this +// in lockstep with `openai-node/src/resources/responses/responses.ts`. +export type OpenAIResponseIncludable = ResponseIncludable +// Mirrors OpenAI's `ServiceTier` union from the official SDK. +export type OpenAIServiceTier = ServiceTier | "scale" | "fast" | "ultrafast" export type OpenAIOptionsInput = Omit & { /** Advanced in-band compaction. The caller owns checkpoint persistence and recovery. */ diff --git a/packages/ai/src/providers/openrouter.ts b/packages/ai/src/providers/openrouter.ts index b4139cf26d7b..1fbdf8399c4c 100644 --- a/packages/ai/src/providers/openrouter.ts +++ b/packages/ai/src/providers/openrouter.ts @@ -8,7 +8,7 @@ import type { ProviderPackage } from "../provider-package.js" import { SystemOne } from "../experimental/system-one.js" import { OpenAIChat } from "../protocols/openai-chat.js" import { cacheControl } from "../protocols/utils/cache.js" -import { isRecord, ProviderShared } from "../protocols/shared.js" +import { isRecord, JsonObject, lenient, ProviderShared } from "../protocols/shared.js" export const id = ProviderID.make("openrouter") const baseURL = "https://openrouter.ai/api/v1" @@ -90,9 +90,26 @@ export type Settings = ProviderPackage.Settings & readonly baseURL?: string } -const OpenRouterBody = Schema.StructWithRest(Schema.Struct(OpenAIChat.bodyFields), [ - Schema.Record(Schema.String, Schema.Any), -]) +const Options = Schema.Struct({ + usage: lenient(Schema.Union([Schema.Boolean, JsonObject])), + models: lenient(Schema.Array(Schema.String)), + provider: lenient(JsonObject), + plugins: lenient(Schema.Array(JsonObject)), + web_search_options: lenient(JsonObject), + debug: lenient(JsonObject), + user: lenient(Schema.String), + reasoning: lenient(JsonObject), +}) +const decodeOptions = Schema.decodeUnknownSync(Options) + +const OpenRouterBody = Schema.StructWithRest( + Schema.Struct({ + ...OpenAIChat.bodyFields, + ...Options.fields, + usage: Schema.optional(JsonObject), + }), + [JsonObject], +) export type OpenRouterBody = Schema.Schema.Type export const protocol = Protocol.make({ @@ -100,33 +117,24 @@ export const protocol = Protocol.make({ body: { schema: OpenRouterBody, from: (request) => - OpenAIChat.fromRequest(request, { cacheControl: cacheControl() }).pipe( - Effect.map((body) => { - const sourceAssistants = request.messages.filter((message) => message.role === "assistant") - let assistantIndex = 0 - const messages = body.messages.map((message) => { - if (message.role !== "assistant") return message - const source = sourceAssistants[assistantIndex++] - const reasoning = source?.content - .filter((part) => part.type === "reasoning") - .map((part) => part.text) - .join("") - const reasoningDetails = Array.isArray(message.reasoning_details) ? message.reasoning_details : undefined - return { - ...message, - reasoning_content: undefined, - reasoning_text: undefined, - reasoning: reasoning && reasoningDetails && reasoningDetails.length > 0 ? reasoning : undefined, - reasoning_details: reasoningDetails, - } - }) + OpenAIChat.fromRequest(request, { + cacheControl: cacheControl(), + assistant: (source, message) => { + if (message.role !== "assistant") return message + const reasoning = source.content + .filter((part) => part.type === "reasoning") + .map((part) => part.text) + .join("") + const reasoningDetails = Array.isArray(message.reasoning_details) ? message.reasoning_details : undefined return { - ...body, - messages, - ...bodyOptions(request.providerOptions, request.generation?.maxTokens), - } as OpenRouterBody - }), - ), + ...message, + reasoning_content: undefined, + reasoning_text: undefined, + reasoning: reasoning && reasoningDetails && reasoningDetails.length > 0 ? reasoning : undefined, + reasoning_details: reasoningDetails, + } + }, + }).pipe(Effect.map((body) => bodyOptions(body, request.providerOptions, request.generation?.maxTokens))), }, stream: OpenAIChat.protocol.stream, }) @@ -138,26 +146,29 @@ const fitReasoning = (reasoning: Record, maxTokens: number | un ? { ...reasoning, max_tokens: ProviderShared.fitThinkingBudget(reasoning.max_tokens, maxTokens, 1_024) } : reasoning -const bodyOptions = (input: unknown, maxTokens: number | undefined) => { +const bodyOptions = (body: OpenAIChat.OpenAIChatBody, input: unknown, maxTokens: number | undefined) => { const openrouter = isRecord(input) ? input : {} - const { usage, models, provider, plugins, web_search_options, debug, user, reasoning, promptCacheKey, ...options } = - openrouter + const options = Object.fromEntries( + Object.entries(openrouter).filter(([key]) => !(key in Options.fields) && key !== "promptCacheKey"), + ) + const known = decodeOptions(openrouter) + const usage = + openrouter.usage === undefined || known.usage === true + ? { include: true } + : known.usage === false + ? { include: false } + : known.usage return { ...options, - ...(usage === undefined || usage === true - ? { usage: { include: true } } - : usage === false - ? { usage: { include: false } } - : isRecord(usage) - ? { usage } - : {}), - ...(Array.isArray(models) ? { models } : {}), - ...(isRecord(provider) ? { provider } : {}), - ...(Array.isArray(plugins) ? { plugins } : {}), - ...(isRecord(web_search_options) ? { web_search_options } : {}), - ...(isRecord(debug) ? { debug } : {}), - ...(typeof user === "string" ? { user } : {}), - ...(isRecord(reasoning) ? { reasoning: fitReasoning(reasoning, maxTokens) } : {}), + ...body, + ...(usage !== undefined ? { usage } : {}), + ...(known.models !== undefined ? { models: known.models } : {}), + ...(known.provider !== undefined ? { provider: known.provider } : {}), + ...(known.plugins !== undefined ? { plugins: known.plugins } : {}), + ...(known.web_search_options !== undefined ? { web_search_options: known.web_search_options } : {}), + ...(known.debug !== undefined ? { debug: known.debug } : {}), + ...(known.user !== undefined ? { user: known.user } : {}), + ...(known.reasoning !== undefined ? { reasoning: fitReasoning(known.reasoning, maxTokens) } : {}), } } diff --git a/packages/ai/src/providers/vercel-ai-gateway.ts b/packages/ai/src/providers/vercel-ai-gateway.ts index 50111c0ab4a0..ff8dd854c5fc 100644 --- a/packages/ai/src/providers/vercel-ai-gateway.ts +++ b/packages/ai/src/providers/vercel-ai-gateway.ts @@ -11,18 +11,18 @@ import { import { AnthropicMessages } from "../protocols/anthropic-messages.js" import { OpenAIChat } from "../protocols/openai-chat.js" import { OpenResponses } from "../protocols/open-responses.js" -import { optionalNull, ProviderShared } from "../protocols/shared.js" +import { JsonObject, optionalNull, ProviderShared } from "../protocols/shared.js" import { gatewayProtocol } from "../protocols/utils/gateway-protocol.js" import type { ProviderPackage } from "../provider-package.js" import { Auth } from "../route/auth.js" import { AuthOptions, type ProviderAuthOption } from "../route/auth-options.js" import { Route, type RouteDefaultsInput } from "../route/client.js" import { Endpoint } from "../route/endpoint.js" +import { RequestExecutor } from "../route/executor.js" import { Framing } from "../route/framing.js" import type { Protocol } from "../route/protocol.js" import { AIError, - HttpContext, HttpOptions, InvalidProviderOutputError, InvalidRequestError, @@ -30,6 +30,7 @@ import { ModelID, ProviderID, ProviderMetadata, + ProviderOptions, ReasoningEffort, Usage, } from "../schema/index.js" @@ -125,7 +126,7 @@ export type Settings = ProviderPackage.Settings & ProviderOptionsInput & { reado const GatewayOptionsSchema = Schema.Struct({ gateway: Schema.optional(Schema.Record(Schema.String, Schema.Unknown)), - upstream: Schema.optional(Schema.Record(Schema.String, Schema.Record(Schema.String, Schema.Unknown))), + upstream: Schema.optional(ProviderMetadata), reasoningEffort: Schema.optional(ReasoningEffort), thinking: Schema.optional( Schema.Struct({ @@ -233,9 +234,9 @@ const Request = Schema.StructWithRest( model: Schema.String, state: EvaluationInput, questions: Schema.Record(Schema.String, EvaluationQuestion), - providerOptions: Schema.optional(Schema.Record(Schema.String, Schema.Unknown)), + providerOptions: Schema.optional(ProviderOptions), }), - [Schema.Record(Schema.String, Schema.Any)], + [JsonObject], ) const Response = Schema.Struct({ model: optionalNull(Schema.String), @@ -310,7 +311,7 @@ export const configure = (input: Options = {}) => { HttpClientRequest.bodyText(body, "application/json"), ), ) - const http = new HttpContext({ url: res.request.url, status: res.status, headers: res.headers }) + const http = RequestExecutor.responseHttp(res) const fail = (message: string, cause: unknown, body?: string) => new AIError({ reason: new InvalidProviderOutputError({ diff --git a/packages/ai/src/route/client.ts b/packages/ai/src/route/client.ts index 7addb1a04266..0cf44a29bcea 100644 --- a/packages/ai/src/route/client.ts +++ b/packages/ai/src/route/client.ts @@ -132,7 +132,7 @@ const makeRouteLanguageModel = { const provider = route.provider ?? ("provider" in mapped ? mapped.provider : undefined) if (!provider) throw new Error(`Route.model(${route.id}) requires a provider`) - if (!endpointBaseURL(route.endpoint)) + if (!route.endpoint.baseURL) throw new ProviderConfigurationError({ provider: ProviderID.make(provider), message: `Route.model(${route.id}) requires an endpoint baseURL — configure it on the route first`, @@ -160,9 +160,6 @@ const mergeRouteDefaults = (base: RouteDefaults | undefined, patch: RouteDefault } } -const endpointBaseURL = (endpoint: Endpoint.Definition) => - typeof endpoint.baseURL === "string" ? endpoint.baseURL : undefined - const mergeHeaders = (...items: ReadonlyArray | undefined>) => { const entries = items.flatMap((item) => item === undefined ? [] : Object.entries(item).filter((entry): entry is [string, string] => entry[1] !== undefined), @@ -647,13 +644,6 @@ export function compact(request: LLMRequest, options?: EndpointCompactOptions | }) } -export const streamRequest = (request: LLMRequest, options?: StreamOptions) => - Stream.unwrap( - Effect.gen(function* () { - return (yield* Service).stream(request, options) - }), - ) - export const layer: Layer.Layer = Layer.effect( Service, Effect.gen(function* () { diff --git a/packages/ai/src/route/executor-service.ts b/packages/ai/src/route/executor-service.ts index 8087867c9bdc..d81d4e20bb7f 100644 --- a/packages/ai/src/route/executor-service.ts +++ b/packages/ai/src/route/executor-service.ts @@ -1,6 +1,6 @@ import { Context, type Effect } from "effect" -import type { HttpClientRequest, HttpClientResponse } from "effect/http" -import type { AIError } from "../schema/errors.js" +import type { Headers, HttpClientRequest, HttpClientResponse } from "effect/http" +import { HttpContext, type AIError } from "../schema/errors.js" // The service tag lives in its own leaf module so `Media.Asset` (imported by the schema layer) can require the // executor without pulling the full executor implementation, and therefore the schema barrel, into a cycle. @@ -24,3 +24,13 @@ export class RequestExecutorService extends Context.Service + Object.fromEntries(Object.entries(headers).map(([name, value]) => [name, String(value)])) + +export const responseHttp = (response: HttpClientResponse.HttpClientResponse) => + new HttpContext({ + url: response.request.url, + status: response.status, + headers: headerDetails(response.headers), + }) diff --git a/packages/ai/src/route/executor.ts b/packages/ai/src/route/executor.ts index ed3d89818753..65850f5e55bc 100644 --- a/packages/ai/src/route/executor.ts +++ b/packages/ai/src/route/executor.ts @@ -9,29 +9,28 @@ import { } from "effect/http" import { HttpContext, HttpRateLimitDetails, AIError, TransportError } from "../schema/index.js" import { classifyProviderFailure, providerErrorMessage } from "../provider-error.js" -import { Service, type HttpMiddleware, type Interface } from "./executor-service.js" +import { headerDetails, responseHttp, Service, type HttpMiddleware, type Interface } from "./executor-service.js" -export { Service } from "./executor-service.js" +export { responseHttp, Service } from "./executor-service.js" export type { HttpHandler, HttpMiddleware, Interface } from "./executor-service.js" -const headerDetails = (headers: Headers.Headers) => - Object.fromEntries(Object.entries(headers).map(([name, value]) => [name, String(value)])) - const normalizedHeaders = (headers: Headers.Headers) => Object.fromEntries(Object.entries(headers).map(([key, value]) => [key.toLowerCase(), value])) +const clampRetryAfterMs = (millis: number) => Math.min(Number.MAX_SAFE_INTEGER, Math.max(0, Math.ceil(millis))) + const retryAfterMs = (headers: Record) => { const millis = Number(headers["retry-after-ms"]) - if (Number.isFinite(millis)) return Math.max(0, millis) + if (Number.isFinite(millis)) return clampRetryAfterMs(millis) const value = headers["retry-after"] if (!value) return undefined const seconds = Number(value) - if (Number.isFinite(seconds)) return Math.max(0, seconds * 1000) + if (Number.isFinite(seconds)) return clampRetryAfterMs(seconds * 1000) const date = Date.parse(value) - if (!Number.isNaN(date)) return Math.max(0, date - Date.now()) + if (!Number.isNaN(date)) return clampRetryAfterMs(date - Date.now()) return undefined } @@ -77,13 +76,6 @@ const rateLimitDetails = (headers: Record, retryAfter: number | }) } -export const responseHttp = (response: HttpClientResponse.HttpClientResponse) => - new HttpContext({ - url: response.request.url, - status: response.status, - headers: headerDetails(response.headers), - }) - const MAX_BODY_CHARS = 2000 // Without a recognized message, show the raw body so the provider's explanation is never dropped. diff --git a/packages/ai/src/route/media-protocol.ts b/packages/ai/src/route/media-protocol.ts index 47514a352c99..615c4fa91e42 100644 --- a/packages/ai/src/route/media-protocol.ts +++ b/packages/ai/src/route/media-protocol.ts @@ -3,11 +3,12 @@ import { HttpClientResponse } from "effect/http" import type { Snapshot, Status } from "../generation.js" import { Media } from "../media.js" import type { AuthInput } from "./auth.js" +import { RequestExecutor } from "./executor.js" import { AIError, AuthenticationError, ContentPolicyError, - HttpContext, + type HttpContext, InvalidProviderOutputError, InvalidRequestError, ProviderID, @@ -200,9 +201,6 @@ const FAILURES = { export type Failure = keyof typeof FAILURES -const context = (response: HttpClientResponse.HttpClientResponse) => - new HttpContext({ url: response.request.url, status: response.status, headers: response.headers }) - /** One protocol's route id, display name, and provider, with the decoders and errors that carry them. */ export const identity = (input: { readonly id: string; readonly name: string; readonly provider: string }) => { const provider = ProviderID.make(input.provider) @@ -217,7 +215,7 @@ export const identity = (input: { readonly id: string; readonly name: string; re * `contentPolicy` is a moderated result. */ const text = Effect.fn("MediaProtocol.text")(function* (response: HttpClientResponse.HttpClientResponse) { - const http = context(response) + const http = RequestExecutor.responseHttp(response) const body = yield* response.text.pipe( Effect.mapError( (cause) => diff --git a/packages/ai/src/route/transport/http.ts b/packages/ai/src/route/transport/http.ts index e8cae2000bc7..a9385338130c 100644 --- a/packages/ai/src/route/transport/http.ts +++ b/packages/ai/src/route/transport/http.ts @@ -32,13 +32,6 @@ export interface HttpPrepared { readonly middleware?: HttpMiddleware } -const applyQuery = (url: string, query: Record | undefined) => { - if (!query) return url - const next = new URL(url) - Object.entries(query).forEach(([key, value]) => next.searchParams.set(key, value)) - return next.toString() -} - const bodyWithOverlay = (body: Body, request: LLMRequest, encodeBody: (body: Body) => string) => Effect.gen(function* () { if (request.http?.body === undefined) return { jsonBody: body, bodyText: encodeBody(body) } @@ -51,10 +44,11 @@ const bodyWithOverlay = (body: Body, request: LLMRequest, encodeBody: (bod export const jsonRequestParts = (input: JsonRequestInput) => Effect.gen(function* () { - const url = applyQuery( - renderEndpoint(input.endpoint, { request: input.request, body: input.body }).toString(), - input.request.http?.query, - ) + const endpoint = renderEndpoint(input.endpoint, { request: input.request, body: input.body }) + if (input.request.http?.query) { + Object.entries(input.request.http.query).forEach(([key, value]) => endpoint.searchParams.set(key, value)) + } + const url = endpoint.toString() const body = yield* bodyWithOverlay(input.body, input.request, input.encodeBody) const headers = yield* Auth.toEffect(input.auth)({ request: input.request, diff --git a/packages/ai/src/route/transport/websocket.ts b/packages/ai/src/route/transport/websocket.ts index f01c6acb63db..4e25c81ee997 100644 --- a/packages/ai/src/route/transport/websocket.ts +++ b/packages/ai/src/route/transport/websocket.ts @@ -8,14 +8,7 @@ import { type HttpContext, type TransportOperation, } from "../../schema/index.js" -import * as HttpTransport from "./http.js" -import type { Transport } from "./index.js" -import type { - ChannelObservation, - WebSocketChannelDriver, - WebSocketChannelExchange, - WebSocketChannelExecutor, -} from "./websocket-channel.js" +import type { ChannelObservation, WebSocketChannelExecutor } from "./websocket-channel.js" export interface WebSocketRequest { readonly url: string @@ -432,81 +425,7 @@ export const direct: Effect.Effect { - readonly toMessage: (body: Body | Record) => Effect.Effect - readonly encodeMessage: (message: Message) => string -} - -export type JsonPatch = Partial> - -export interface JsonTransport extends Transport { - readonly with: (patch: JsonPatch) => JsonTransport -} - -export const json = (input: JsonInput): JsonTransport => ({ - id: "websocket-json", - with: (patch) => json({ ...input, ...patch }), - prepare: (prepareInput) => - Effect.gen(function* () { - const parts = yield* HttpTransport.jsonRequestParts({ - ...prepareInput, - }) - return { - url: yield* toWebSocketUrl(parts.url), - headers: parts.headers, - message: input.encodeMessage(yield* input.toMessage(parts.jsonBody)), - } - }), - execute: (prepared, request, _runtime, options) => { - const webSocket = options?.webSocket - if (!webSocket) { - return Effect.fail( - transportError("WebSocket JSON transport requires StreamOptions.webSocket", { - url: prepared.url, - operation: "request", - code: "unavailable", - phase: "prepare", - delivery: "not-sent", - }), - ) - } - const driver: WebSocketChannelDriver = { - create: () => Effect.succeed({ message: prepared.message, mode: "full" }), - observe: (_create, frame) => Effect.succeed({ type: "frame", frame }), - } - const exchange: WebSocketChannelExchange = { - id: request.id ?? "request", - connect: { url: prepared.url, headers: prepared.headers }, - fallback: () => - Stream.fail( - transportError("WebSocket JSON transport does not provide HTTP fallback", { - url: prepared.url, - operation: "request", - code: "websocket", - phase: "fallback", - delivery: "not-sent", - }), - ), - driver, - } - return webSocket.execute(exchange) - }, -}) - -export const jsonTransport = { - id: "websocket-json", - with: json, -} as const - export const WebSocketTransport = { - json, - jsonTransport, direct, makeDirect, open, diff --git a/packages/ai/src/schema/errors.ts b/packages/ai/src/schema/errors.ts index 76a8e4846146..4c6ab9d1c445 100644 --- a/packages/ai/src/schema/errors.ts +++ b/packages/ai/src/schema/errors.ts @@ -1,4 +1,5 @@ import { Schema } from "effect" +import { NonNegativeInt } from "@opencode/schema" import { Tool } from "@opencode/schema/tool" import { ModelID, ProviderID, RouteID } from "./ids.js" @@ -12,7 +13,7 @@ export class HttpContext extends Schema.Class("AI.HttpContext")({ }) {} export class HttpRateLimitDetails extends Schema.Class("AI.HttpRateLimitDetails")({ - retryAfterMs: Schema.optional(Schema.Number), + retryAfterMs: Schema.optional(NonNegativeInt), limit: Schema.optional(Schema.Record(Schema.String, Schema.String)), remaining: Schema.optional(Schema.Record(Schema.String, Schema.String)), reset: Schema.optional(Schema.Record(Schema.String, Schema.String)), @@ -77,13 +78,16 @@ export class AuthenticationError extends Schema.TaggedError export class RateLimitError extends Schema.TaggedError("AI.Error.RateLimit")("RateLimit", { ...ReasonFields, - retryAfterMs: Schema.optional(Schema.Number), + retryAfterMs: Schema.optional(NonNegativeInt), rateLimit: Schema.optional(HttpRateLimitDetails), }) {} export class QuotaExceededError extends Schema.TaggedError("AI.Error.QuotaExceeded")( "QuotaExceeded", - ReasonFields, + { + ...ReasonFields, + retryAfterMs: Schema.optional(NonNegativeInt), + }, ) {} export class ContentPolicyError extends Schema.TaggedError("AI.Error.ContentPolicy")( @@ -95,7 +99,7 @@ export class ProviderInternalError extends Schema.TaggedError("AI.Error.Timeout")("Timeout", { ...ReasonFields, - timeoutMs: Schema.optional(Schema.Number), + timeoutMs: Schema.optional(NonNegativeInt), }) {} export const AIErrorReason = Schema.Union([ diff --git a/packages/ai/src/schema/messages.ts b/packages/ai/src/schema/messages.ts index cb27dd00e350..a9359bef1d0c 100644 --- a/packages/ai/src/schema/messages.ts +++ b/packages/ai/src/schema/messages.ts @@ -1,4 +1,4 @@ -import { Schema } from "effect" +import { Option, Schema } from "effect" import { withStatics } from "@opencode/schema/schema" import { Tool } from "@opencode/schema/tool" import { @@ -15,6 +15,7 @@ import { } from "./options.js" import { ProviderID } from "./ids.js" import { Media } from "../media.js" +import { Json } from "../utils/json.js" export const MessageRole = Schema.Literals(["system", "user", "assistant", "tool"]) export type MessageRole = Schema.Schema.Type @@ -119,13 +120,11 @@ export const ToolOutput = withStatics( }, ) +const encodeJsonOption = Schema.encodeUnknownOption(Json) + const toolResultText = (value: unknown) => { if (typeof value === "string") return value - try { - return JSON.stringify(value) ?? String(value) - } catch { - return String(value) - } + return Option.getOrElse(encodeJsonOption(value), () => String(value)) } export const ToolCallPart = withStatics( diff --git a/packages/ai/src/speech.ts b/packages/ai/src/speech.ts index 9c07febe72c0..38fc50d07c44 100644 --- a/packages/ai/src/speech.ts +++ b/packages/ai/src/speech.ts @@ -3,7 +3,14 @@ import { ProgressEvent, QueuedEvent } from "./generation.js" import { Media } from "./media.js" import { MediaModel, composeRoute, tryRequest } from "./media-model.js" import { MediaRoute } from "./route/media.js" -import { AIError, HttpOptions, MediaUsage, ProviderMetadata, type OpenString } from "./schema/index.js" +import { + AIError, + HttpOptions, + MediaUsage, + ProviderMetadata, + ProviderOptions, + type OpenString, +} from "./schema/index.js" import { SpeechClient, Service } from "./speech-client.js" // --------------------------------------------------------------------------- @@ -76,7 +83,7 @@ export class SpeechRequest extends Schema.Class("Speech.Request") language: Schema.optional(Schema.String), instructions: Schema.optional(Schema.String), timestamps: Schema.optional(Schema.Boolean), - providerOptions: Schema.optional(Schema.Record(Schema.String, Schema.Unknown)), + providerOptions: Schema.optional(ProviderOptions), http: Schema.optional(HttpOptions), }) { declare protected readonly _SpeechRequest: void diff --git a/packages/ai/src/transcription.ts b/packages/ai/src/transcription.ts index 1302d72fccd6..33f0a8f55af5 100644 --- a/packages/ai/src/transcription.ts +++ b/packages/ai/src/transcription.ts @@ -3,7 +3,7 @@ import { Generation, ProgressEvent, QueuedEvent, type AwaitOptions } from "./gen import { Media } from "./media.js" import { MediaModel, composeRoute, tryRequest } from "./media-model.js" import { MediaRoute } from "./route/media.js" -import { AIError, HttpOptions, MediaUsage, ProviderMetadata } from "./schema/index.js" +import { AIError, HttpOptions, MediaUsage, ProviderMetadata, ProviderOptions } from "./schema/index.js" import { TranscriptionClient, Service } from "./transcription-client.js" // --------------------------------------------------------------------------- @@ -74,7 +74,7 @@ export class TranscriptionRequest extends Schema.Class("Tr timestamps: Schema.optional(TranscriptionTimestamps), diarize: Schema.optional(Schema.Boolean), speakers: Schema.optional(Schema.Int), - providerOptions: Schema.optional(Schema.Record(Schema.String, Schema.Unknown)), + providerOptions: Schema.optional(ProviderOptions), http: Schema.optional(HttpOptions), }) { declare protected readonly _TranscriptionRequest: void diff --git a/packages/ai/src/video.ts b/packages/ai/src/video.ts index c97ac08f43f6..65a2109a4da1 100644 --- a/packages/ai/src/video.ts +++ b/packages/ai/src/video.ts @@ -3,7 +3,14 @@ import { Generation, ProgressEvent, QueuedEvent, type AwaitOptions } from "./gen import { Media } from "./media.js" import { MediaModel, composeRoute, tryRequest } from "./media-model.js" import { MediaRoute } from "./route/media.js" -import { AIError, HttpOptions, MediaUsage, ProviderMetadata, type OpenString } from "./schema/index.js" +import { + AIError, + HttpOptions, + MediaUsage, + ProviderMetadata, + ProviderOptions, + type OpenString, +} from "./schema/index.js" import { VideoClient, Service } from "./video-client.js" // --------------------------------------------------------------------------- @@ -79,7 +86,7 @@ export class VideoRequest extends Schema.Class("Video.Request")({ n: Schema.optional(Schema.Int), seed: Schema.optional(Schema.Number), negativePrompt: Schema.optional(Schema.String), - providerOptions: Schema.optional(Schema.Record(Schema.String, Schema.Unknown)), + providerOptions: Schema.optional(ProviderOptions), http: Schema.optional(HttpOptions), }) { declare protected readonly _VideoRequest: void diff --git a/packages/ai/test/cache-policy.test.ts b/packages/ai/test/cache-policy.test.ts index 97664896bfc2..b8513c810ae7 100644 --- a/packages/ai/test/cache-policy.test.ts +++ b/packages/ai/test/cache-policy.test.ts @@ -1,6 +1,6 @@ import { describe, expect, test } from "bun:test" import { Effect } from "effect" -import { CacheHint, LLM, Message } from "../src/index.js" +import { CacheHint, CompactionPart, LLM, Message, ProviderID, ToolCallPart } from "../src/index.js" import { Auth } from "../src/route.js" import { compileRequest } from "../src/route/client.js" import { @@ -525,6 +525,31 @@ describe("applyCachePolicy", () => { }), ) + test("skips non-cacheable parts when placing message breakpoints", () => { + const compaction = CompactionPart.make({ provider: ProviderID.make("anthropic"), text: "summary" }) + const toolCall = ToolCallPart.make({ id: "tc_1", name: "lookup", input: {} }) + + const mixed = LLM.request({ + model: anthropicModel, + messages: [Message.assistant([toolCall, compaction])], + cache: { messages: "latest-assistant" }, + }) + const appliedMixed = applyCachePolicy(mixed) + expect(appliedMixed.messages[0]?.content[0]).toMatchObject({ + type: "tool-call", + id: "tc_1", + cache: { type: "ephemeral" }, + }) + expect(appliedMixed.messages[0]?.content[1]).toEqual(compaction) + + const compactionOnly = LLM.request({ + model: anthropicModel, + messages: [Message.assistant([compaction])], + cache: { messages: "latest-assistant" }, + }) + expect(applyCachePolicy(compactionOnly)).toBe(compactionOnly) + }) + test("returns the same request reference when policy is a no-op (pure function)", () => { const request = LLM.request({ model: anthropicModel, diff --git a/packages/ai/test/executor.test.ts b/packages/ai/test/executor.test.ts index a0bd3d17ea9c..07baf92bbda9 100644 --- a/packages/ai/test/executor.test.ts +++ b/packages/ai/test/executor.test.ts @@ -64,6 +64,14 @@ describe("RequestExecutor", () => { ) expect(RequestExecutor.httpFailure({ message: "No response", url: request.url }).reason.http).toBeUndefined() expect(RequestExecutor.httpFailure({ message: "No URL", status: 500 }).reason.http).toBeUndefined() + expect( + RequestExecutor.httpFailure({ + message: "Huge retry-after", + url: request.url, + status: 429, + responseHeaders: { "Retry-After": "1e20" }, + }).reason, + ).toMatchObject({ _tag: "RateLimit", retryAfterMs: Number.MAX_SAFE_INTEGER }) }), ) diff --git a/packages/ai/test/generation.test.ts b/packages/ai/test/generation.test.ts index b3ea1be74919..40baff2b41bb 100644 --- a/packages/ai/test/generation.test.ts +++ b/packages/ai/test/generation.test.ts @@ -74,8 +74,17 @@ describe("Generation", () => { expect(error).toBeInstanceOf(AIError) expect(error.reason._tag).toBe("Timeout") + if (error.reason._tag === "Timeout") expect(error.reason.timeoutMs).toBe(5000) expect(error.message).toContain("gen_1") expect(yield* Ref.get(scripted.polls)).toBeGreaterThan(1) + + const subMilliFiber = yield* Effect.forkChild( + generation.await({ poll: { interval: "1 second", timeout: "500 micros" } }).pipe(Effect.flip), + ) + yield* TestClock.adjust("1 milli") + const subMilliError = yield* Fiber.join(subMilliFiber) + + expect(subMilliError.reason).toMatchObject({ _tag: "Timeout", timeoutMs: 1 }) }), ) diff --git a/packages/ai/test/provider/bedrock-converse.test.ts b/packages/ai/test/provider/bedrock-converse.test.ts index 251581254940..887c0150ca69 100644 --- a/packages/ai/test/provider/bedrock-converse.test.ts +++ b/packages/ai/test/provider/bedrock-converse.test.ts @@ -1657,9 +1657,9 @@ describe("Bedrock Converse route", () => { }), ) - it.effect("uses originalMessage from model stream exception frames", () => + it.effect("uses originalMessage and originalStatusCode from model stream exception frames", () => Effect.gen(function* () { - const error = yield* LLMClient.generate(baseRequest).pipe( + const serverError = yield* LLMClient.generate(baseRequest).pipe( Effect.provide( fixedBytes( exceptionFrame("modelStreamErrorException", { @@ -1670,8 +1670,20 @@ describe("Bedrock Converse route", () => { ), Effect.flip, ) + const clientError = yield* LLMClient.generate(baseRequest).pipe( + Effect.provide( + fixedBytes( + exceptionFrame("modelStreamErrorException", { + originalMessage: "Upstream rejected request", + originalStatusCode: 400, + }), + ), + ), + Effect.flip, + ) - expect(error).toMatchObject({ reason: { _tag: "ProviderInternal" }, message: "Upstream model failed" }) + expect(serverError).toMatchObject({ reason: { _tag: "ProviderInternal" }, message: "Upstream model failed" }) + expect(clientError).toMatchObject({ reason: { _tag: "InvalidRequest" }, message: "Upstream rejected request" }) }), ) diff --git a/packages/ai/test/provider/openrouter.test.ts b/packages/ai/test/provider/openrouter.test.ts index 1c374d268aaa..5cd74db2e0e9 100644 --- a/packages/ai/test/provider/openrouter.test.ts +++ b/packages/ai/test/provider/openrouter.test.ts @@ -492,6 +492,7 @@ describe("OpenRouter", () => { model: OpenRouter.configure({ apiKey: "test-key" }).model("anthropic/claude-sonnet-4.6"), cache: "none", messages: [ + Message.assistant(""), Message.assistant([ { type: "reasoning", diff --git a/packages/ai/test/realtime.test.ts b/packages/ai/test/realtime.test.ts index 2f1a663a5abd..ac39877f260c 100644 --- a/packages/ai/test/realtime.test.ts +++ b/packages/ai/test/realtime.test.ts @@ -107,7 +107,7 @@ describe("experimental Realtime", () => { if (req.url === "https://api.openai.com/v1/realtime/client_secrets") { expect(req.headers.get("authorization")).toBe("Bearer sk-openai") expect(JSON.parse(input.text)).toEqual({ - expires_after: { seconds: 60 }, + expires_after: { anchor: "created_at", seconds: 60 }, session: { type: "realtime", instructions: "Brief.", model: "gpt-realtime-2.1" }, }) return input.respond(JSON.stringify({ value: "ek_1", expires_at: 1760000060 })) @@ -138,6 +138,9 @@ describe("experimental Realtime", () => { ) test("routes bare and versioned model names to their engine", () => { + expect(Realtime.engines.map((item) => Realtime.engine(item.model)?.id)).toEqual( + Realtime.engines.map((item) => item.id), + ) expect( [ "gpt-realtime", diff --git a/packages/ai/test/schema.test.ts b/packages/ai/test/schema.test.ts index 7c6f78546d1c..351c0cde2445 100644 --- a/packages/ai/test/schema.test.ts +++ b/packages/ai/test/schema.test.ts @@ -21,6 +21,8 @@ import { QuotaExceededError, RateLimitError, RouteID, + TimeoutError, + ToolOutput, ToolResultValue, TransportError, UnknownProviderError, @@ -115,6 +117,28 @@ describe("ToolResultValue", () => { }), ).toBe(true) }) + + test("fromResultValue encodes text results and falls back to String for non-JSON values", () => { + const circular: Record = {} + circular.self = circular + + expect(ToolOutput.fromResultValue({ type: "text", value: "plain" })).toEqual({ + structured: {}, + content: [{ type: "text", text: "plain" }], + }) + expect(ToolOutput.fromResultValue({ type: "text", value: { ok: true } })).toEqual({ + structured: {}, + content: [{ type: "text", text: '{"ok":true}' }], + }) + expect(ToolOutput.fromResultValue({ type: "text", value: undefined })).toEqual({ + structured: {}, + content: [{ type: "text", text: "undefined" }], + }) + expect(ToolOutput.fromResultValue({ type: "text", value: circular })).toEqual({ + structured: {}, + content: [{ type: "text", text: "[object Object]" }], + }) + }) }) describe("AI.Usage", () => { @@ -326,6 +350,7 @@ test("AI error reasons are tagged Errors with required messages", () => { new TransportError({ message: "Connection failed", transport: "http", operation: "request" }), new InvalidProviderOutputError({ message: "Invalid output" }), new UnknownProviderError({ message: "Unknown failure" }), + new TimeoutError({ message: "Timed out", timeoutMs: 5000 }), ] expect(reasons.map((reason) => reason._tag)).toEqual([ "InvalidRequest", @@ -339,7 +364,14 @@ test("AI error reasons are tagged Errors with required messages", () => { "Transport", "InvalidProviderOutput", "UnknownProvider", + "Timeout", ]) + expect(Schema.decodeUnknownOption(TimeoutError)({ _tag: "Timeout", message: "Timed out", timeoutMs: -1 })._tag).toBe( + "None", + ) + expect(Schema.decodeUnknownOption(TimeoutError)({ _tag: "Timeout", message: "Timed out", timeoutMs: 1.5 })._tag).toBe( + "None", + ) reasons.forEach((reason) => { expect(reason).toBeInstanceOf(Error) const encoded = Schema.encodeSync(AIErrorReason)(reason) diff --git a/packages/cli/package.json b/packages/cli/package.json index ca2b331b4095..eabb4a4c7d9e 100644 --- a/packages/cli/package.json +++ b/packages/cli/package.json @@ -1,7 +1,7 @@ { "$schema": "https://json.schemastore.org/package.json", "name": "@opencode/cli", - "version": "2.1.0", + "version": "2.1.2", "type": "module", "license": "MIT", "bin": { @@ -39,7 +39,7 @@ "@opencode/util": "workspace:*", "@opentui/core": "catalog:", "@opentui/solid": "catalog:", - "@opentunnel/client": "0.4.0", + "@opentunnel/client": "0.4.1", "@parcel/watcher": "2.5.1", "@silvia-odwyer/photon-node": "0.3.4", "diff": "catalog:", diff --git a/packages/cli/src/services/service-config.ts b/packages/cli/src/services/service-config.ts index 0c74f9fd16ae..8c389a51a7de 100644 --- a/packages/cli/src/services/service-config.ts +++ b/packages/cli/src/services/service-config.ts @@ -45,10 +45,11 @@ export function filename(channel = OPENCODE_CHANNEL) { export function defaultPort(channel = OPENCODE_CHANNEL, platform = process.platform) { const port = channelPort(channel) - // Windows hands out 0xC000-0xFFFF for outgoing connections, and Hyper-V and WinNAT reserve blocks of that range at - // boot, so a port there can fail as in use while nothing listens on it. Moving into the 0xB000 block keeps the low - // 12 bits: 0xC0DE becomes 0xB0DE, which is 0xC0DE - 4096, a nod to the v1 port. - if (platform === "win32" && port >= 0xc000) return 0xb000 | (port & 0xfff) + // Windows hands out 0xC000-0xFFFF for outgoing connections, Hyper-V and WinNAT reserve blocks of that range at boot, + // and WSL mirrored networking reserves a 4096-port block that usually covers 44620-48715. A reserved port fails as in + // use while nothing listens on it. Moving into the 0x7000 block, below 32768, avoids all of them and keeps the low + // 12 bits: 0xC0DE becomes 0x70DE. + if (platform === "win32" && port >= 0x8000) return 0x7000 | (port & 0xfff) return port } diff --git a/packages/cli/test/service.test.ts b/packages/cli/test/service.test.ts index e22a7ef11441..400bc1e4c430 100644 --- a/packages/cli/test/service.test.ts +++ b/packages/cli/test/service.test.ts @@ -28,17 +28,17 @@ test("managed service ports are stable per installation channel", () => { expect(ServiceConfig.defaultPort("preview-a", "linux")).not.toBe(ServiceConfig.defaultPort("preview-b", "linux")) }) -test("managed service ports on Windows stay below the range Windows reserves", () => { - expect(ServiceConfig.defaultPort("latest", "win32")).toBe(0xc0de - 4096) - expect(ServiceConfig.defaultPort("dev", "win32")).toBe(0xc0de - 4096) - expect(ServiceConfig.defaultPort("local", "win32")).toBe(0xc0df - 4096) +test("managed service ports on Windows stay below the ranges Windows and WSL reserve", () => { + expect(ServiceConfig.defaultPort("latest", "win32")).toBe(0x70de) + expect(ServiceConfig.defaultPort("dev", "win32")).toBe(0x70de) + expect(ServiceConfig.defaultPort("local", "win32")).toBe(0x70df) const ports = Array.from({ length: 200 }, (_, index) => ({ windows: ServiceConfig.defaultPort(`preview-${index}`, "win32"), linux: ServiceConfig.defaultPort(`preview-${index}`, "linux"), })) - expect(ports.some((port) => port.linux >= 0xc000)).toBe(true) - expect(ports.every((port) => port.windows >= 10_000 && port.windows < 0xc000)).toBe(true) - expect(ports.filter((port) => port.linux < 0xc000 && port.windows !== port.linux)).toEqual([]) + expect(ports.some((port) => port.linux >= 0x8000)).toBe(true) + expect(ports.every((port) => port.windows >= 10_000 && port.windows < 0x8000)).toBe(true) + expect(ports.filter((port) => port.linux < 0x8000 && port.windows !== port.linux)).toEqual([]) }) test("service disabled accepts only booleans without changing configuration on invalid input", async () => { diff --git a/packages/client/package.json b/packages/client/package.json index adcafbee643b..5ca07d6f42f4 100644 --- a/packages/client/package.json +++ b/packages/client/package.json @@ -1,7 +1,7 @@ { "$schema": "https://json.schemastore.org/package.json", "name": "@opencode/client", - "version": "2.1.0", + "version": "2.1.2", "type": "module", "license": "MIT", "repository": { diff --git a/packages/client/src/effect/api/api.ts b/packages/client/src/effect/api/api.ts index 4592c8cf50ca..7b740e56765e 100644 --- a/packages/client/src/effect/api/api.ts +++ b/packages/client/src/effect/api/api.ts @@ -685,6 +685,7 @@ export type SessionLogOutput = readonly message: string readonly status?: number | undefined readonly response?: { readonly body: string } | undefined + readonly retryAfterMs?: number | undefined } } } @@ -852,6 +853,7 @@ export type SessionLogOutput = readonly message: string readonly status?: number | undefined readonly response?: { readonly body: string } | undefined + readonly retryAfterMs?: number | undefined } readonly finish?: "content-filter" | undefined readonly rawFinish?: string | undefined @@ -1040,6 +1042,7 @@ export type SessionLogOutput = readonly message: string readonly status?: number | undefined readonly response?: { readonly body: string } | undefined + readonly retryAfterMs?: number | undefined } readonly content?: | readonly [ @@ -1087,6 +1090,7 @@ export type SessionLogOutput = readonly message: string readonly status?: number | undefined readonly response?: { readonly body: string } | undefined + readonly retryAfterMs?: number | undefined } } } @@ -1148,6 +1152,7 @@ export type SessionLogOutput = readonly message: string readonly status?: number | undefined readonly response?: { readonly body: string } | undefined + readonly retryAfterMs?: number | undefined } readonly inputID?: SessionMessage.ID | undefined readonly cost?: Money.USD | undefined @@ -1413,6 +1418,19 @@ export interface GenerateApi { readonly text: GenerateTextOperation } +export type VoiceModelsOutput = { + readonly models: ReadonlyArray<{ + readonly model: string + readonly provider: string + readonly transports: ReadonlyArray<"webrtc" | "websocket"> + readonly available: boolean + readonly integration?: string | undefined + readonly workspace?: string | undefined + }> + readonly default?: string | undefined +} +export type VoiceModelsOperation = () => Effect.Effect + export type VoiceSessionInput = { readonly model: string readonly session?: { readonly [x: string]: unknown } | undefined @@ -1424,11 +1442,20 @@ export type VoiceSessionOutput = { readonly transport: | { readonly type: "webrtc"; readonly sdp: string } | { readonly type: "websocket"; readonly url: string; readonly token: string; readonly expiresAt: number } + readonly session?: { readonly id: string } | undefined } export type VoiceSessionOperation = (input: VoiceSessionInput) => Effect.Effect +export type VoiceInstructionsInput = { readonly sessionID: string; readonly instructions: string } +export type VoiceInstructionsOutput = void +export type VoiceInstructionsOperation = ( + input: VoiceInstructionsInput, +) => Effect.Effect + export interface VoiceApi { + readonly models: VoiceModelsOperation readonly session: VoiceSessionOperation + readonly instructions: VoiceInstructionsOperation } export type ProviderListInput = { readonly location?: { readonly directory?: string | undefined } | undefined } diff --git a/packages/client/src/effect/generated/client.ts b/packages/client/src/effect/generated/client.ts index bcff357c9da8..6bc3693c1319 100644 --- a/packages/client/src/effect/generated/client.ts +++ b/packages/client/src/effect/generated/client.ts @@ -115,8 +115,11 @@ import type { ModelDefaultOutput, GenerateTextInput, GenerateTextOutput, + VoiceModelsOutput, VoiceSessionInput, VoiceSessionOutput, + VoiceInstructionsInput, + VoiceInstructionsOutput, ProviderListInput, ProviderListOutput, ProviderGetInput, @@ -873,6 +876,14 @@ const EndpointGenerateText = (raw: RawClient["server.generate"]) => (input: Gene const adaptGroupGenerate = (raw: RawClient["server.generate"]) => ({ text: EndpointGenerateText(raw) }) +const EndpointVoiceModels = (raw: RawClient["server.voice"]) => () => + preserveEffect()( + raw["voice.models"]({}).pipe( + Effect.mapError(mapClientError), + Effect.map((value) => value.data), + ), + ) + const EndpointVoiceSession = (raw: RawClient["server.voice"]) => (input: VoiceSessionInput) => preserveEffect()( raw["voice.session"]({ @@ -883,7 +894,19 @@ const EndpointVoiceSession = (raw: RawClient["server.voice"]) => (input: VoiceSe ), ) -const adaptGroupVoice = (raw: RawClient["server.voice"]) => ({ session: EndpointVoiceSession(raw) }) +const EndpointVoiceInstructions = (raw: RawClient["server.voice"]) => (input: VoiceInstructionsInput) => + preserveEffect()( + raw["voice.instructions"]({ + params: { sessionID: input["sessionID"] }, + payload: { instructions: input["instructions"] }, + }).pipe(Effect.mapError(mapClientError)), + ) + +const adaptGroupVoice = (raw: RawClient["server.voice"]) => ({ + models: EndpointVoiceModels(raw), + session: EndpointVoiceSession(raw), + instructions: EndpointVoiceInstructions(raw), +}) const EndpointProviderList = (raw: RawClient["server.provider"]) => (input?: ProviderListInput) => preserveEffect()( diff --git a/packages/client/src/promise/generated/client.ts b/packages/client/src/promise/generated/client.ts index ca3c2ad885c3..c8348a656eba 100644 --- a/packages/client/src/promise/generated/client.ts +++ b/packages/client/src/promise/generated/client.ts @@ -109,8 +109,11 @@ import type { ModelDefaultOutput, GenerateTextInput, GenerateTextOutput, + VoiceModelsOutput, VoiceSessionInput, VoiceSessionOutput, + VoiceInstructionsInput, + VoiceInstructionsOutput, ProviderListInput, ProviderListOutput, ProviderGetInput, @@ -1162,6 +1165,17 @@ export function make(options: ClientOptions) { ).then((value) => value.data), }, voice: { + models: (requestOptions?: RequestOptions) => + request<{ readonly data: VoiceModelsOutput }>( + { + method: "GET", + path: `/api/experimental/voice/models`, + successStatus: 200, + declaredStatuses: [400, 401, 404, 503], + empty: false, + }, + requestOptions, + ).then((value) => value.data), session: (input: VoiceSessionInput, requestOptions?: RequestOptions) => request<{ readonly data: VoiceSessionOutput }>( { @@ -1174,6 +1188,18 @@ export function make(options: ClientOptions) { }, requestOptions, ).then((value) => value.data), + instructions: (input: VoiceInstructionsInput, requestOptions?: RequestOptions) => + request( + { + method: "POST", + path: `/api/experimental/voice/session/${encodeURIComponent(input.sessionID)}/instructions`, + body: { instructions: input["instructions"] }, + successStatus: 204, + declaredStatuses: [400, 401, 404, 503], + empty: true, + }, + requestOptions, + ), }, provider: { list: (input?: ProviderListInput, requestOptions?: RequestOptions) => diff --git a/packages/client/src/promise/generated/types.ts b/packages/client/src/promise/generated/types.ts index be9351b6cca4..c0bbf1cb451b 100644 --- a/packages/client/src/promise/generated/types.ts +++ b/packages/client/src/promise/generated/types.ts @@ -108,7 +108,13 @@ export type ToolTextContent = { type: "text"; text: string } export type ToolFileContent = { type: "file"; uri: string; mime: string; name?: string | null } -export type SessionStructuredError = { type: string; message: string; status?: number; response?: { body: string } } +export type SessionStructuredError = { + type: string + message: string + status?: number + response?: { body: string } + retryAfterMs?: number +} export type SessionActive = { type: "running" } @@ -174,6 +180,8 @@ export type MoneyUSDPerMillionTokens = number export type GenerateTextResponse = { data: { text: string } } +export type VoiceTransportType = "webrtc" | "websocket" + export type VoiceWebRTCTransport = { type: "webrtc"; sdp: string } export type VoiceWebSocketTransport = { type: "websocket"; url: string; token: string; expiresAt: number } @@ -656,6 +664,15 @@ export type ModelCost = { cache: { read: MoneyUSDPerMillionTokens; write: MoneyUSDPerMillionTokens } } +export type VoiceModel = { + model: string + provider: string + transports: Array + available: boolean + integration?: string | undefined + workspace?: string | undefined +} + export type VoiceTransport = VoiceWebRTCTransport | VoiceWebSocketTransport export type IntegrationCommandMethod = { @@ -1979,7 +1996,9 @@ export type ProviderInfo = { body?: { [x: string]: any } } -export type VoiceSessionResponse = { data: { transport: VoiceTransport } } +export type VoiceModelsResponse = { data: { models: Array; default?: string | undefined } } + +export type VoiceSessionResponse = { data: { transport: VoiceTransport; session?: { id: string } | undefined } } export type CredentialValue = CredentialOAuth | CredentialKey | CredentialExternal @@ -2331,6 +2350,7 @@ export type ConfigEntry = portable_shell_scanner?: boolean todo?: boolean subagent_depth?: number + location_idle_timeout?: string | false policies?: Array<{ action: "provider.use" | "tool.use" | "integration.use" resource: string @@ -4959,6 +4979,8 @@ export type GenerateTextInput = { export type GenerateTextOutput = GenerateTextResponse["data"] +export type VoiceModelsOutput = VoiceModelsResponse["data"] + export type VoiceSessionInput = { readonly model: { readonly model: string @@ -4985,6 +5007,13 @@ export type VoiceSessionInput = { export type VoiceSessionOutput = VoiceSessionResponse["data"] +export type VoiceInstructionsInput = { + readonly sessionID: { readonly sessionID: string }["sessionID"] + readonly instructions: { readonly instructions: string }["instructions"] +} + +export type VoiceInstructionsOutput = void + export type ProviderListInput = { readonly location?: { readonly location?: { readonly directory?: string | undefined } | undefined }["location"] } diff --git a/packages/client/test/promise.test.ts b/packages/client/test/promise.test.ts index d2f8b4c28d5d..e06997162bc2 100644 --- a/packages/client/test/promise.test.ts +++ b/packages/client/test/promise.test.ts @@ -46,7 +46,7 @@ test("exposes every standard HTTP API group", () => { expect(Object.keys(client.integration.oauth)).toEqual(["connect", "status", "complete", "cancel"]) expect(Object.keys(client.integration.command)).toEqual(["connect", "status", "cancel"]) expect(Object.keys(client.websearch)).toEqual(["providers", "query"]) - expect(Object.keys(client.voice)).toEqual(["session"]) + expect(Object.keys(client.voice)).toEqual(["models", "session", "instructions"]) expect(Object.keys(client.file)).toEqual(["read", "list", "find", "write"]) expect(Object.keys(client.vcs)).toEqual(["init", "get", "base", "status", "branch", "diff"]) expect(Object.keys(client.pty)).toEqual(["list", "create", "get", "update", "remove", "connect"]) diff --git a/packages/codemode/interpreter-support.md b/packages/codemode/interpreter-support.md index e02a1c4e0fc5..0179ac1e2af6 100644 --- a/packages/codemode/interpreter-support.md +++ b/packages/codemode/interpreter-support.md @@ -274,6 +274,9 @@ Math.floor)` is `"3"`). A detached method loses its receiver, as in JS: `values. unless its cleanup fails, and direct self-resolution rejects with a `TypeError`. - [x] Every `await` (including of plain values and already-settled promises) defers its continuation one reaction turn, so concurrent async functions interleave at await points as in JavaScript. +- [x] Each job runs to completion: code between two awaits, and an async function's body up to its first await, + finishes before another job runs, even when the scheduler preempts a long job. Tool calls still run + concurrently. - [x] Combinators settle one reaction turn after their deciding member (V8-observable ordering): reactions already attached to members run first, and an aggregate cannot beat a plain value settling in the same turn into a `Promise.race`. Exact microtask-count parity beyond this observable ordering is not a documented guarantee. diff --git a/packages/codemode/package.json b/packages/codemode/package.json index 7d6629de2c17..5345f3c9b192 100644 --- a/packages/codemode/package.json +++ b/packages/codemode/package.json @@ -1,7 +1,7 @@ { "$schema": "https://json.schemastore.org/package.json", "name": "@opencode/codemode", - "version": "2.1.0", + "version": "2.1.2", "description": "Effect-native confined code execution over schema-described tools", "type": "module", "license": "MIT", diff --git a/packages/codemode/src/interpreter/execute.ts b/packages/codemode/src/interpreter/execute.ts index c1fdab13c6c7..9643e9393ff9 100644 --- a/packages/codemode/src/interpreter/execute.ts +++ b/packages/codemode/src/interpreter/execute.ts @@ -40,16 +40,21 @@ export const executeProgram = ( const program = parseProgram(code) const pending = new Pending(scope, builtins.Promise) const ctx = new Interpreter({ tools, pending, builtins, logs, globals }) - const result = (yield* toBoundary(ctx, yield* ctx.run(program))) ?? null - returned = { value: result, pending } - const warnings = yield* pending.interrupt() - return { - ok: true, - value: result, - ...(warnings.length > 0 ? { warnings } : {}), - ...logged(), - toolCalls: tools.calls, - } satisfies Result + // The main job holds the turn until leftover promise work is interrupted. + return yield* pending.turn.hold( + Effect.gen(function* () { + const result = (yield* toBoundary(ctx, yield* ctx.run(program))) ?? null + returned = { value: result, pending } + const warnings = yield* pending.interrupt() + return { + ok: true, + value: result, + ...(warnings.length > 0 ? { warnings } : {}), + ...logged(), + toolCalls: tools.calls, + } satisfies Result + }), + ) }), (scope, exit) => Scope.close(scope, exit), ) diff --git a/packages/codemode/src/interpreter/extensions.ts b/packages/codemode/src/interpreter/extensions.ts index 530b4575c86d..a45a435d2b9e 100644 --- a/packages/codemode/src/interpreter/extensions.ts +++ b/packages/codemode/src/interpreter/extensions.ts @@ -155,7 +155,7 @@ export const extensionGlobals = ( return settle(Effect.fail(reason)) } if (!(result instanceof Promise)) return settle(Effect.succeed(result)) - return ctx.pending.create(settle(Effect.tryPromise({ try: () => result, catch: (reason) => reason }))) + return ctx.pending.createHost(settle(Effect.tryPromise({ try: () => result, catch: (reason) => reason }))) }) return extensions.flatMap((extension) => diff --git a/packages/codemode/src/interpreter/interpreter.ts b/packages/codemode/src/interpreter/interpreter.ts index b60ab8404a9b..4527078e634c 100644 --- a/packages/codemode/src/interpreter/interpreter.ts +++ b/packages/codemode/src/interpreter/interpreter.ts @@ -419,9 +419,11 @@ class Frame { }).pipe(Effect.ensuring(Effect.sync(() => self.scopes.pop()))) } - // Fork at the call site so admission and hooks occur when the call is made. + // Fork at the call site so admission and hooks occur when the call is made. Arguments cross on the caller's turn, + // since `toJSON` is program code; the host call runs without it, so tool calls stay concurrent. private createToolCallPromise(path: ReadonlyArray, args: Array): Effect.Effect { - return this.ctx.pending.create(this.ctx.tool((json) => this.ctx.tools.execute(path, json), args)) + const pending = this.ctx.pending + return pending.create(this.ctx.tool((json) => pending.turn.leave(this.ctx.tools.execute(path, json)), args)) } // Fiber exits make settlement idempotent; yielding prevents inline continuation. @@ -429,10 +431,17 @@ class Frame { const pending = this.ctx.pending return Effect.suspend(() => { pending.markObserved(promise) - return Effect.flatMap(pending.await(promise), (exit) => Effect.andThen(Effect.yieldNow, exit)) + return Effect.flatten( + pending.turn.suspend(Effect.flatMap(pending.await(promise), (exit) => Effect.as(Effect.yieldNow, exit))), + ) }) } + // End this job; the program continues in a later one, after the jobs already waiting. + nextJob(): Effect.Effect { + return this.ctx.pending.turn.suspend(Effect.yieldNow) + } + private evaluateStatement(node: Statement | ModuleDeclaration): Effect.Effect { switch (node.type) { case "ExpressionStatement": @@ -730,7 +739,7 @@ class Frame { iterator ? self.closeIterator(iterator, node, awaiting) : awaiting - ? Effect.andThen(cursor?.close ?? Effect.void, Effect.yieldNow) + ? Effect.andThen(cursor?.close ?? Effect.void, self.nextJob()) : (cursor?.close ?? Effect.void) if (left.type === "RestElement" || left.type === "AssignmentPattern") { @@ -887,7 +896,7 @@ class Frame { const called = yield* Effect.exit(self.call(iterator.next, iterator.iterator, [], node)) if (!Exit.isSuccess(called)) { - if (awaiting) yield* Effect.yieldNow + if (awaiting) yield* self.nextJob() return yield* Effect.failCause(called.cause) } const captured = yield* Effect.exit( @@ -897,7 +906,7 @@ class Frame { }), ) if (!Exit.isSuccess(captured)) { - if (awaiting) yield* Effect.yieldNow + if (awaiting) yield* self.nextJob() return yield* Effect.failCause(captured.cause) } return { @@ -915,7 +924,9 @@ class Frame { awaiting = true, ): Effect.Effect { const close = get(iterator.iterator, "return") - if (close === undefined || close === null) return iterator.asynchronous || !awaiting ? Effect.void : Effect.yieldNow + if (close === undefined || close === null) { + return iterator.asynchronous || !awaiting ? Effect.void : this.nextJob() + } const self = this return Effect.gen(function* () { const method = self.requireIteratorMethod(close, "Iterator return", node) @@ -930,14 +941,14 @@ class Frame { const called = yield* Effect.exit(self.call(method, iterator.iterator, [], node)) if (!Exit.isSuccess(called)) { - if (awaiting) yield* Effect.yieldNow + if (awaiting) yield* self.nextJob() return yield* Effect.failCause(called.cause) } const captured = yield* Effect.exit( Effect.sync(() => get(self.requireIteratorObject(called.value, "Iterator return() result", node), "value")), ) if (!Exit.isSuccess(captured)) { - if (awaiting) yield* Effect.yieldNow + if (awaiting) yield* self.nextJob() return yield* Effect.failCause(captured.cause) } if (awaiting) yield* self.awaitValue(captured.value) @@ -1814,14 +1825,22 @@ class Frame { invocation.generatorAsync = asynchronous const builtins = this.ctx.builtins const result = (value: Value, done: boolean) => record(builtins.Object, { value, done }) + const turn = this.ctx.pending.turn const request = (kind: GeneratorRequestKind, value: Value) => { const request = { kind, value, response: Deferred.makeUnsafe() } + // A body that ran on the caller's turn has answered or suspended when it hands the turn back. An answer still + // pending can only be an async generator's: its promise settles when the body answers, without the turn. + const answer = Effect.suspend(() => + Deferred.isDoneUnsafe(request.response) + ? Deferred.await(request.response) + : turn.leave(Deferred.await(request.response)), + ) if (!asynchronous && state.active) return Effect.die(typeError("Generator is already running.")) if (asynchronous && (state.completed || (!state.started && kind !== "next"))) { state.started = true state.completed = true state.pending.push(request) - if (state.draining) return Deferred.await(request.response) + if (state.draining) return answer state.draining = true return Effect.andThen( this.ctx.pending.fork( @@ -1829,7 +1848,7 @@ class Frame { .completeGeneratorRequests(state, true) .pipe(Effect.ensuring(Effect.sync(() => (state.draining = false)))), ), - Deferred.await(request.response), + answer, ) } if (state.completed) { @@ -1844,9 +1863,13 @@ class Frame { state.pending.push(request) if (state.available) { + // The body is parked waiting for a request: it runs this step on the caller's turn. const available = state.available state.available = undefined - Deferred.doneUnsafe(available, Exit.succeed(undefined)) + return Effect.andThen( + turn.handOff(Effect.sync(() => Deferred.doneUnsafe(available, Exit.succeed(undefined)))), + answer, + ) } if (!state.started) { state.started = true @@ -1872,9 +1895,9 @@ class Frame { yield* invocation.completeGeneratorRequests(state, asynchronous) state.completed = true }) - return Effect.andThen(this.ctx.pending.fork(body), Deferred.await(request.response)) + return Effect.andThen(this.ctx.pending.fork(body), answer) } - return Deferred.await(request.response) + return answer } const proto = get(fn, "prototype") return new GeneratorObj(proto instanceof Obj ? proto : builtins.Generator, asynchronous, request) @@ -1912,7 +1935,7 @@ class Frame { if (next) return Effect.succeed(next) state.available = Deferred.makeUnsafe() return Effect.andThen( - Deferred.await(state.available), + this.ctx.pending.turn.park(Deferred.await(state.available)), Effect.sync(() => this.dequeueGeneratorRequest(state)!), ) } diff --git a/packages/codemode/src/interpreter/promises.ts b/packages/codemode/src/interpreter/promises.ts index 4a82667e0768..fde5f60efcf0 100644 --- a/packages/codemode/src/interpreter/promises.ts +++ b/packages/codemode/src/interpreter/promises.ts @@ -16,8 +16,189 @@ const capability = (ctx: Interpreter, name: string, settle: (value: Value) return undefined }) +type Waiter = { + readonly fiber: Fiber.Fiber + readonly resume: (effect: Effect.Effect) => void + next?: Waiter + removed: boolean +} + +// Program fibers that synchronous starts and handbacks may nest on the host stack. Deeper ones start and resume +// through the scheduler, so deep recursion through async calls cannot overflow the host stack. +const MAX_NESTING = 100 + +type Handback = { + readonly fiber: Fiber.Fiber + returned: boolean + resume?: (effect: Effect.Effect) => void +} + +/** + * The right to run program code. JavaScript runs one job at a time: code between two awaits finishes before another + * job starts. Effect preempts a busy fiber so timeouts and host work still run, so only the fiber holding the turn may + * run program code. A synchronous call into another fiber (an async function's body up to its first await, a + * generator step) lends that fiber the turn and gets it back when the callee suspends or ends. An await gives the turn + * to the next waiting job and queues to take it back. Work that only settles a promise or runs host code leaves + * without it, so tool calls stay concurrent and resolvers settle in the call, as in JavaScript. + */ +export class Turn { + private holder: Fiber.Fiber | undefined + // Callers waiting for a callee to hand the turn back, innermost last. + private readonly handbacks: Array = [] + // Jobs waiting to take the turn, in the order they became ready. + private first: Waiter | undefined + private last: Waiter | undefined + private nesting = 0 + + // Take the turn when it is free, after the jobs already waiting. + readonly take: Effect.Effect = Effect.withFiber((fiber) => { + if (this.holder === undefined) { + this.holder = fiber + return Effect.void + } + return Effect.callback((resume) => { + const waiter: Waiter = { fiber, resume, removed: false } + if (this.last === undefined) this.first = waiter + else this.last.next = waiter + this.last = waiter + // Interrupted while waiting, or after the turn was granted but before it ran. + return Effect.sync(() => { + if (this.holder === fiber) this.release(fiber) + else waiter.removed = true + }) + }) + }) + + // Run `effect` holding the turn, as the program's main job does. + hold(effect: Effect.Effect): Effect.Effect { + return Effect.withFiber((fiber) => + Effect.andThen(this.take, effect).pipe(Effect.ensuring(Effect.sync(() => this.release(fiber)))), + ) + } + + // Give up the turn, back to the innermost caller waiting for it or to the next waiting job. + release(fiber: Fiber.Fiber): void { + if (this.holder !== fiber) return + const back = this.handbacks.pop() + if (back !== undefined) { + this.holder = back.fiber + this.handBack(back) + return + } + let next = this.first + while (next !== undefined && next.removed) next = next.next + this.first = next?.next + if (this.first === undefined) this.last = undefined + this.holder = next?.fiber + // Resume in the scheduler, as a yield would: release can run inside another fiber's step. + if (next !== undefined) next.fiber.currentDispatcher.scheduleTask(() => next.resume(Effect.void), 0) + } + + // Wait without the turn, then queue for it like any job that became ready. An interrupted fiber stops without it. + suspend(wait: Effect.Effect): Effect.Effect { + return Effect.uninterruptibleMask((restore) => + Effect.withFiber((fiber) => { + this.release(fiber) + return Effect.flatMap(Effect.exit(restore(wait)), (exit) => + Exit.isFailure(exit) && Cause.hasInterruptsOnly(exit.cause) ? exit : Effect.andThen(restore(this.take), exit), + ) + }), + ) + } + + // Wait without the turn and finish without it, for work that only settles a promise or runs host code. + leave(wait: Effect.Effect): Effect.Effect { + return Effect.withFiber((fiber) => { + this.release(fiber) + return wait + }) + } + + // Run `start`, which wakes a fiber parked with `park`, and continue once it hands the turn back. + handOff(start: Effect.Effect): Effect.Effect { + return Effect.withFiber((fiber) => { + const back: Handback = { fiber, returned: false } + this.handbacks.push(back) + return Effect.andThen(this.nest(start), awaitHandback(back)) + }) + } + + // Wait without the turn for a caller to hand it over with `handOff`. + park(wait: Effect.Effect): Effect.Effect { + return Effect.uninterruptibleMask((restore) => + Effect.withFiber((fiber) => { + this.release(fiber) + // The wake runs inside the caller's step; when that is deep, yield so the body continues on its own stack. + return Effect.tap(restore(wait), () => { + this.holder = fiber + return this.nesting > MAX_NESTING ? Effect.yieldNow : Effect.void + }) + }), + ) + } + + // Fork a fiber that runs program code. It starts with the caller's turn, and the caller continues once the fiber + // suspends or ends; the fiber gives up the turn whenever it ends. + fork(effect: Effect.Effect, scope: Scope.Scope): Effect.Effect, never, R> { + return Effect.withFiber((caller) => { + const back: Handback = { fiber: caller, returned: false } + this.handbacks.push(back) + let started = false + const body = Effect.withFiber((fiber) => { + started = true + this.holder = fiber + return effect + }) + const forked = + this.nesting < MAX_NESTING + ? this.nest(Effect.forkIn(body, scope, { startImmediately: true })) + : Effect.forkIn(body, scope) + return Effect.flatMap(forked, (fiber) => { + fiber.addObserver(() => { + if (started) return this.release(fiber) + // Forked into a closing scope: it never took the turn, so the caller keeps it. + const index = this.handbacks.lastIndexOf(back) + if (index >= 0) this.handbacks.splice(index, 1) + this.handBack(back) + }) + // A callee that suspends without blocking hands the turn back before `forkIn` returns. + return back.returned ? Effect.succeed(fiber) : Effect.as(awaitHandback(back), fiber) + }) + }) + } + + // Resume a parked caller in the scheduler: resumed inline, it would run to its next await and queue before the callee + // that just suspended has queued its own continuation. + private handBack(back: Handback): void { + back.returned = true + const resume = back.resume + if (resume !== undefined) back.fiber.currentDispatcher.scheduleTask(() => resume(Effect.void), 0) + } + + // Run `effect`, which can run another program fiber inside the current step, counting the nesting. + private nest(effect: Effect.Effect): Effect.Effect { + return Effect.suspend(() => { + this.nesting++ + return Effect.ensuring( + effect, + Effect.sync(() => { + this.nesting-- + }), + ) + }) + } +} + +// The check runs when the caller suspends: it can be preempted after deciding to wait. +const awaitHandback = (back: Handback) => + Effect.callback((resume) => { + if (back.returned) resume(Effect.void) + else back.resume = resume + }) + // Observation only controls rejection reporting; program completion interrupts all promise work. export class Pending { + readonly turn = new Turn() private readonly active = new Set() private readonly ids = new WeakMap() private readonly observed = new WeakSet() @@ -40,7 +221,20 @@ export class Pending { }) } + // A promise whose work runs program code, on the turn. Host work that runs no program code uses `createHost`. create(effect: Effect.Effect): Effect.Effect { + return this.register(effect, (body) => this.turn.fork(body, this.scope)) + } + + // Starts now, without the turn, so host work stays concurrent and attaches to host promises at once. + createHost(effect: Effect.Effect): Effect.Effect { + return this.register(effect, (body) => Effect.forkIn(body, this.scope, { startImmediately: true })) + } + + private register( + effect: Effect.Effect, + fork: (body: Effect.Effect) => Effect.Effect, never, R>, + ): Effect.Effect { return Effect.flatMap(CallSite, (site) => { if (this.active.size >= MAX_PENDING_PROMISES) { throw rangeError( @@ -50,7 +244,7 @@ export class Pending { // Allocate before forking so reruns get distinct IDs and diagnostics retain creation order. const id = this.nextID++ const body = Effect.catchDefect(effect, (defect) => Effect.die(locate(defect, site.node))) - return Effect.map(Effect.forkIn(body, this.scope, { startImmediately: true }), (fiber) => { + return Effect.map(fork(body), (fiber) => { const promise = new PromiseObj(this.proto, fiber) this.active.add(promise) this.ids.set(promise, id) @@ -84,7 +278,7 @@ export class Pending { } fork(effect: Effect.Effect): Effect.Effect { - return Effect.asVoid(Effect.forkIn(effect, this.scope, { startImmediately: true })) + return Effect.asVoid(this.turn.fork(effect, this.scope)) } diagnostics(): Array { @@ -118,7 +312,7 @@ export const resolvePromiseValue = ( return Effect.gen(function* () { // Promise resolution invokes a thenable's method in a later job. - yield* Effect.yieldNow + yield* ctx.pending.turn.suspend(Effect.yieldNow) const deferred = Deferred.makeUnsafe() const resolve = capability(ctx, "resolve", (result) => Deferred.doneUnsafe(deferred, Exit.succeed(result))) const reject = capability(ctx, "reject", (reason) => Deferred.doneUnsafe(deferred, Exit.fail(new Throw(reason)))) @@ -127,7 +321,7 @@ export const resolvePromiseValue = ( if (Cause.hasInterruptsOnly(executed.cause)) return yield* Effect.failCause(executed.cause) Deferred.doneUnsafe(deferred, Exit.fail(Cause.squash(executed.cause))) } - return yield* resolvePromiseValue(ctx, yield* Deferred.await(deferred), own) + return yield* resolvePromiseValue(ctx, yield* ctx.pending.turn.leave(Deferred.await(deferred)), own) }) } @@ -184,6 +378,7 @@ const invokePromiseMethod = ( return new Arr( ctx.builtins.Array, yield* settleAfterTurn( + ctx, Effect.all( items.map((item) => Effect.flatten(ctx.pending.await(item))), { concurrency: "unbounded" }, @@ -192,29 +387,33 @@ const invokePromiseMethod = ( ) } if (name === "allSettled") { - const outcomes: Array = [] - for (const item of items) { - const exit = yield* ctx.pending.await(item) - if (Exit.isSuccess(exit)) { - outcomes.push(record(ctx.builtins.Object, { status: "fulfilled", value: exit.value })) - continue - } - if (Cause.hasInterruptsOnly(exit.cause)) return yield* Effect.failCause(exit.cause) - outcomes.push( - record(ctx.builtins.Object, { - status: "rejected", - reason: materialize(ctx, Cause.squash(exit.cause)), - }), - ) - } - yield* Effect.yieldNow - return new Arr(ctx.builtins.Array, outcomes) + return yield* ctx.pending.turn.suspend( + Effect.gen(function* () { + const outcomes: Array = [] + for (const item of items) { + const exit = yield* ctx.pending.await(item) + if (Exit.isSuccess(exit)) { + outcomes.push(record(ctx.builtins.Object, { status: "fulfilled", value: exit.value })) + continue + } + if (Cause.hasInterruptsOnly(exit.cause)) return yield* Effect.failCause(exit.cause) + outcomes.push( + record(ctx.builtins.Object, { + status: "rejected", + reason: materialize(ctx, Cause.squash(exit.cause)), + }), + ) + } + yield* Effect.yieldNow + return new Arr(ctx.builtins.Array, outcomes) + }), + ) } if (name === "race") { if (items.length === 0) { throw typeError("Promise.race([]) would never settle; provide at least one promise or value.") } - return yield* settleAfterTurn(Effect.flatten(Effect.raceAll(items.map((item) => ctx.pending.await(item))))) + return yield* settleAfterTurn(ctx, Effect.flatten(Effect.raceAll(items.map((item) => ctx.pending.await(item))))) } const flipped = items.map((item) => Effect.flatMap(ctx.pending.await(item), (exit) => { @@ -224,6 +423,7 @@ const invokePromiseMethod = ( }), ) return yield* settleAfterTurn( + ctx, Effect.all(flipped, { concurrency: "unbounded" }).pipe( Effect.flatMap((reasons) => Effect.fail(new Throw(createAggregateErrorValue(ctx, reasons, "All promises were rejected"))), @@ -259,7 +459,9 @@ const promiseCapability = (ctx: Interpreter) => Effect.gen(function* () { const deferred = Deferred.makeUnsafe() const promise = yield* ctx.pending.createWithSelf((self) => - Effect.flatMap(Deferred.await(deferred), (value) => resolvePromiseValue(ctx, value, self)), + Effect.flatMap(ctx.pending.turn.leave(Deferred.await(deferred)), (value) => + resolvePromiseValue(ctx, value, self), + ), ) const resolve = capability(ctx, "resolve", (value) => Deferred.doneUnsafe(deferred, Exit.succeed(value))) const reject = capability(ctx, "reject", (value) => Deferred.doneUnsafe(deferred, Exit.fail(new Throw(value)))) @@ -281,9 +483,12 @@ const constructPromise = (ctx: Interpreter, executor: Value): Effect.Effec }) } -// Settle one reaction turn after the deciding member, after its existing reactions. -const settleAfterTurn = (body: Effect.Effect): Effect.Effect => - Effect.flatMap(Effect.exit(body), (exit) => Effect.andThen(Effect.yieldNow, exit)) +// Settle one reaction turn after the deciding member, after its existing reactions. That reaction is a job: it queues +// for the turn behind the jobs ready before it and settles holding it. +const settleAfterTurn = (ctx: Interpreter, body: Effect.Effect): Effect.Effect => + Effect.flatten( + ctx.pending.turn.suspend(Effect.flatMap(Effect.exit(body), (exit) => Effect.as(Effect.yieldNow, exit))), + ) class PromiseAnyFulfilled { constructor(readonly value: Value) {} @@ -304,12 +509,14 @@ const reactionExit = ( ctx: Interpreter, source: PromiseObj, ): Effect.Effect, unknown, R> => - Effect.gen(function* () { - const exit = yield* ctx.pending.await(source) - if (!Exit.isSuccess(exit) && Cause.hasInterruptsOnly(exit.cause)) return yield* Effect.failCause(exit.cause) - yield* Effect.yieldNow - return exit - }) + ctx.pending.turn.suspend( + Effect.gen(function* () { + const exit = yield* ctx.pending.await(source) + if (!Exit.isSuccess(exit) && Cause.hasInterruptsOnly(exit.cause)) return yield* Effect.failCause(exit.cause) + yield* Effect.yieldNow + return exit + }), + ) const chainReaction = ( ctx: Interpreter, diff --git a/packages/codemode/test/promise.test.ts b/packages/codemode/test/promise.test.ts index 7d6bfda7472e..5a36c37ea41b 100644 --- a/packages/codemode/test/promise.test.ts +++ b/packages/codemode/test/promise.test.ts @@ -1342,3 +1342,61 @@ describe("promise construction", () => { ).toEqual({ kind: "function" }) }) }) + +describe("jobs run to completion", () => { + // Effect preempts a fiber after 2,048 operations; these loops are long enough to be preempted mid-job. + test("updates to shared state between awaits are not lost", async () => { + expect( + await value(` + let total = 0 + const work = async () => { + await null + for (let i = 0; i < 1000; i++) total = total + 1 + } + await Promise.all([work(), work()]) + return total + `), + ).toBe(2000) + }) + + test("an async function runs to its first await before its caller continues", async () => { + expect( + await value(` + let count = 0 + const work = async () => { + for (let i = 0; i < 1000; i++) count = count + 1 + await null + } + work() + return count + `), + ).toBe(1000) + }) + + test("a combinator's reaction runs after the continuations a preempted job queued first", async () => { + for (const op of ["race", "any", "all", "allSettled"]) { + expect( + await value(` + const burn = () => { for (let i = 0; i < 400; i++) {} } + const log = [] + const settled = Promise.${op}([{ then(resolve) { burn(); resolve(1) } }]).then(() => log.push("result")) + const job = (async () => { await null; burn(); log.push("b0"); await null; log.push("b1") })() + await Promise.all([settled, job]) + return log + `), + ).toEqual(["b0", "b1", "result"]) + } + }) + + test("a job that never awaits still times out while other jobs wait for it", async () => { + const failure = await error( + ` + const spin = async () => { await null; while (true) {} } + const quick = async () => { await null; return 1 } + return await Promise.all([spin(), quick(), quick()]) + `, + { limits: { timeoutMs: 100 } }, + ) + expect(failure.kind).toBe("TimeoutExceeded") + }) +}) diff --git a/packages/containers/bun-node/Dockerfile b/packages/containers/bun-node/Dockerfile index f4ca864629b7..5418d522e7bb 100644 --- a/packages/containers/bun-node/Dockerfile +++ b/packages/containers/bun-node/Dockerfile @@ -4,7 +4,7 @@ FROM ${REGISTRY}/build/base:24.04 SHELL ["/bin/bash", "-lc"] ARG NODE_VERSION=24.4.0 -ARG BUN_VERSION=1.4.2 +ARG BUN_VERSION=1.4.4 ENV BUN_INSTALL=/opt/bun ENV PATH=/opt/bun/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin diff --git a/packages/core/package.json b/packages/core/package.json index d063983c8cf6..72588e801a4a 100644 --- a/packages/core/package.json +++ b/packages/core/package.json @@ -1,6 +1,6 @@ { "$schema": "https://json.schemastore.org/package.json", - "version": "2.1.0", + "version": "2.1.2", "name": "@opencode/core", "type": "module", "license": "MIT", diff --git a/packages/core/src/agent.ts b/packages/core/src/agent.ts index 14732fdd7522..d574e7528b04 100644 --- a/packages/core/src/agent.ts +++ b/packages/core/src/agent.ts @@ -30,8 +30,8 @@ type Data = { } export type Editor = { - list: () => readonly Info[] - get: (id: ID) => Info | undefined + list: () => readonly Types.DeepMutable[] + get: (id: ID) => Types.DeepMutable | undefined default: (id: ID | undefined) => void update: (id: ID, fn: (agent: Types.DeepMutable) => void) => void remove: (id: ID) => void @@ -40,7 +40,7 @@ export type Editor = { export interface Interface extends State.Transformable { readonly get: (id: ID) => Effect.Effect readonly resolve: (id?: ID | string) => Effect.Effect - readonly select: (id?: ID | string) => Effect.Effect + readonly select: (id?: ID) => Effect.Effect readonly list: () => Effect.Effect } @@ -54,7 +54,7 @@ const layer = Layer.effect( name: "agent", initial: () => ({ agents: new Map() }), editor: (editor) => ({ - list: () => Array.fromIterable(editor.agents.values()) as Info[], + list: () => Array.fromIterable(editor.agents.values()), get: (id) => editor.agents.get(id), default: (id) => { editor.default = id @@ -77,7 +77,7 @@ const layer = Layer.effect( const data = state.get() const configured = data.default ? selectable(data.agents.get(data.default)) : undefined if (configured) return configured - const build = selectable(data.agents.get(ID.make("build"))) + const build = selectable(data.agents.get(defaultID)) if (build) return build for (const agent of data.agents.values()) { const fallback = selectable(agent) @@ -96,10 +96,7 @@ const layer = Layer.effect( return selectedDefault() }), select: Effect.fn("Agent.select")(function* (id) { - if (id !== undefined) { - const selected = ID.make(id) - return { id: selected, info: state.get().agents.get(selected) } - } + if (id !== undefined) return { id, info: state.get().agents.get(id) } const info = selectedDefault() return { id: info?.id ?? defaultID, info } }), diff --git a/packages/core/src/aisdk.ts b/packages/core/src/aisdk.ts index a392273dad70..4904eedd38ba 100644 --- a/packages/core/src/aisdk.ts +++ b/packages/core/src/aisdk.ts @@ -32,6 +32,7 @@ import { } from "@opencode/ai" import { Auth, Endpoint, RequestExecutor, type AnyRoute, type HttpMiddleware } from "@opencode/ai/route" import { ProviderShared } from "@opencode/ai/protocols/shared" +import { isRecord } from "@opencode/ai/utils/record" import { Cause, Context, Effect, Layer, Option, Schema, Scope, Stream } from "effect" import { makeParser } from "effect/encoding/Sse" import { HttpClientRequest, HttpClientResponse } from "effect/http" @@ -39,25 +40,26 @@ import { AsyncLocalStorage } from "node:async_hooks" import type { ID, RuntimeInfo } from "./model.js" import { Provider } from "./provider.js" import { State } from "./state.js" +import { stringify } from "./tool/runtime.js" type SDK = any type UserContent = Extract["content"] type AssistantContent = Extract["content"] type ToolResultContent = Extract -const decodeJson = Schema.decodeUnknownOption(Schema.fromJsonString(Schema.Unknown)) +const decodeJsonRecord = Schema.decodeUnknownOption(Schema.fromJsonString(Schema.Record(Schema.String, Schema.Json))) export interface SDKEvent { readonly model: RuntimeInfo readonly package: string - readonly options: Record + readonly options: Record sdk?: SDK } export interface LanguageEvent { readonly model: RuntimeInfo readonly sdk: SDK - readonly options: Record + readonly options: Record language?: LanguageModelV3 } @@ -120,7 +122,7 @@ function wrapSSE(res: Response, ms: number, ctl: AbortController) { function prepareOptions(model: RuntimeInfo, pkg: string) { const projected = mapBodyToProviderOptions(model, pkg) - const options: Record = { + const options: Record = { name: model.canonical ?? model.providerID, ...(model.settings ?? {}), headers: model.headers, @@ -153,13 +155,13 @@ function prepareOptions(model: RuntimeInfo, pkg: string) { ) if (typeof opts.body === "string" && model.body !== undefined) { - const decoded = Option.getOrUndefined(decodeJson(opts.body)) - if (Schema.is(Schema.Record(Schema.String, Schema.Json))(decoded)) { + const decoded = Option.getOrUndefined(decodeJsonRecord(opts.body)) + if (decoded) { opts.body = JSON.stringify(Provider.mergeOverlay(decoded, model.body)) } } - const send: Fetch = typeof customFetch === "function" ? customFetch : fetch + const send: Fetch = isFetch(customFetch) ? customFetch : fetch const middleware = httpMiddleware.getStore() const res = await ( middleware @@ -175,6 +177,8 @@ function prepareOptions(model: RuntimeInfo, pkg: string) { type Fetch = (input: Parameters[0], init?: BunFetchRequestInit) => Promise +const isFetch = (value: unknown): value is Fetch => typeof value === "function" + // HTTP hook middleware is scoped to one model request, but the SDK's fetch is baked into the // cached language model, so the active middleware rides along in async context instead. const httpMiddleware = new AsyncLocalStorage<{ http: HttpMiddleware; context: Context.Context }>() @@ -420,11 +424,8 @@ function modelFromLanguage(info: RuntimeInfo, language: LanguageModelV3) { }) } -function gatewayProviderOptions(modelID: ID, settings: Readonly>) { - const gateway = - typeof settings.gateway === "object" && settings.gateway !== null && !Array.isArray(settings.gateway) - ? Object.fromEntries(Object.entries(settings.gateway)) - : undefined +function gatewayProviderOptions(modelID: ID, settings: Readonly>) { + const gateway = isRecord(settings.gateway) ? { ...settings.gateway } : undefined const model = Object.fromEntries(Object.entries(settings).filter(([key]) => key !== "gateway")) if (Object.keys(model).length === 0) return gateway === undefined ? undefined : { gateway } @@ -666,7 +667,7 @@ function toolOutput(result: ToolResultValue) { switch (result.type) { case "text": case "error": - return { type: "text" as const, value: messageValue(result.value) } + return { type: "text" as const, value: stringify(result.value) } case "content": return { type: "content" as const, @@ -924,16 +925,7 @@ function jsonValue(input: unknown): JSONValue { const encoded = JSON.stringify(input) return encoded === undefined ? null : (JSON.parse(encoded) as JSONValue) } catch { - return messageValue(input) - } -} - -function messageValue(input: unknown) { - if (typeof input === "string") return input - try { - return JSON.stringify(input) ?? String(input) - } catch { - return String(input) + return stringify(input) } } diff --git a/packages/core/src/codemode/catalog.ts b/packages/core/src/codemode/catalog.ts index 3c8f2b9f19cc..07dfe59c5adb 100644 --- a/packages/core/src/codemode/catalog.ts +++ b/packages/core/src/codemode/catalog.ts @@ -2,14 +2,13 @@ export * as CodeModeCatalog from "./catalog.js" import { Schema } from "effect" -export const Tool = Schema.Struct({ - type: Schema.Literal("tool"), - name: Schema.String, - description: Schema.String, - signature: Schema.String, - pinned: Schema.optionalKey(Schema.Boolean), -}) -export type Tool = typeof Tool.Type +export type Tool = { + readonly type: "tool" + readonly name: string + readonly description: string + readonly signature: string + readonly pinned?: boolean +} export type Namespace = { readonly type: "namespace" diff --git a/packages/core/src/codemode/web.ts b/packages/core/src/codemode/web.ts index 0a6b5b84f8e6..fb7dc4ebd588 100644 --- a/packages/core/src/codemode/web.ts +++ b/packages/core/src/codemode/web.ts @@ -1,6 +1,7 @@ export * as CodeModeWeb from "./web.js" import { Extension } from "@opencode/codemode" +import { isRecord } from "@opencode/ai/utils/record" const TIMEOUT_MS = 30_000 @@ -43,6 +44,7 @@ export const extension = Extension.make({ name: "web", globals: { fetch } }) /** What to show for a fetch call: its method and URL. */ export const display = (args: ReadonlyArray) => { - const [input, init] = args as [string | URL, Init | undefined] - return { method: init?.method?.toUpperCase() ?? "GET", url: String(input) } + const [input, init] = args + const method = isRecord(init) && typeof init.method === "string" ? init.method.toUpperCase() : "GET" + return { method, url: String(input) } } diff --git a/packages/core/src/config/normalize.ts b/packages/core/src/config/normalize.ts index 35d702a051f7..4ef02b96e5e3 100644 --- a/packages/core/src/config/normalize.ts +++ b/packages/core/src/config/normalize.ts @@ -485,6 +485,15 @@ function normalizeExperimental( ) if (value !== undefined) result.subagent_depth = value } + if (own(experimental, "location_idle_timeout")) { + const value = decodeEncoded( + ConfigExperimental.Info.fields.location_idle_timeout, + experimental.location_idle_timeout, + ["experimental", "location_idle_timeout"], + diagnostics, + ) + if (value !== undefined) result.location_idle_timeout = value + } } } if (Object.keys(result).length || (isRecord(input.experimental) && !Object.keys(input.experimental).length)) diff --git a/packages/core/src/config/plugin/agent.ts b/packages/core/src/config/plugin/agent.ts index 1d3949d015ea..57abeef1f580 100644 --- a/packages/core/src/config/plugin/agent.ts +++ b/packages/core/src/config/plugin/agent.ts @@ -26,7 +26,6 @@ const legacySources = [ const sourceDirectories = ["agent", "agents", "mode", "modes"] as const const decodeAgent = Schema.decodeUnknownOption(ConfigAgent.Info) const decodeLegacyAgent = Schema.decodeUnknownOption(ConfigAgentV1.Info) -const decodeConfig = Schema.decodeUnknownOption(Info) type PathAction = FileAccess.ExternalDirectoryAuthorization["action"] | typeof ReadTool.name | typeof EditTool.name const pathActions = ["external_directory", "read", "edit"] as const satisfies readonly PathAction[] const agentKeys = new Set(["variant", ...Object.keys(ConfigAgent.Info.fields)]) @@ -202,11 +201,8 @@ function decode(file: { directory: string; filepath: string; primary: boolean }, ) : Option.getOrUndefined(decodeAgent({ ...data, system: body }, { errors: "all" })) if (!agent) return - const info = Option.getOrUndefined( - decodeConfig({ - agents: { [name]: file.primary ? { ...agent, mode: "primary" } : agent }, - }), - ) - if (!info) return + const info = Info.make({ + agents: { [name]: file.primary ? ConfigAgent.Info.make({ ...agent, mode: "primary" }) : agent }, + }) return new Document({ type: "document", path: AbsolutePath.make(file.filepath), info }) } diff --git a/packages/core/src/config/plugin/command.ts b/packages/core/src/config/plugin/command.ts index a029fb8adeff..5dc6c140acfe 100644 --- a/packages/core/src/config/plugin/command.ts +++ b/packages/core/src/config/plugin/command.ts @@ -3,8 +3,6 @@ export * as ConfigCommandPlugin from "./command.js" import { define } from "@opencode/plugin/effect/plugin" import { Info, type Entry } from "@opencode/schema/config" import { ConfigCommand } from "@opencode/schema/config/command" -import { Model } from "@opencode/schema/model" -import { Provider } from "@opencode/schema/provider" import { AppProcess } from "@opencode/util/process" import path from "path" import { Effect, Option, PubSub, Schema, Stream } from "effect" @@ -84,11 +82,9 @@ export const Plugin = define({ command.model === undefined ? commandAgent?.model : { - id: Model.ID.make(command.model.model), - providerID: Provider.ID.make(command.model.providerID), - ...(command.model.variant === undefined - ? {} - : { variant: Model.VariantID.make(command.model.variant) }), + id: command.model.model, + providerID: command.model.providerID, + ...(command.model.variant === undefined ? {} : { variant: command.model.variant }), } const text = yield* evaluateTemplate(command.template, input.prompt.text, { location, diff --git a/packages/core/src/config/plugin/provider.ts b/packages/core/src/config/plugin/provider.ts index 5ea1c70a58e7..21a1a87edb84 100644 --- a/packages/core/src/config/plugin/provider.ts +++ b/packages/core/src/config/plugin/provider.ts @@ -85,9 +85,16 @@ export const Plugin = define({ providers.models.update(providerID, id, (model) => { if (base) Object.assign(model, structuredClone(base)) if (item.package !== undefined) model.package = undefined - if (item.settings?.baseURL !== undefined && model.settings) delete model.settings.baseURL }) } + // A configured provider baseURL replaces every catalog per-model endpoint, not only those of configured + // models. Configured per-model settings are applied later and still win. + if (item.settings?.baseURL !== undefined) + for (const [id, model] of providers.get(providerID)?.models ?? []) + if (model.settings?.baseURL !== undefined) + providers.models.update(providerID, id, (model) => { + delete model.settings?.baseURL + }) next.set(item, { providerID, models: definitions }) } for (const [item, definition] of next) { diff --git a/packages/core/src/config/plugin/skill-file.ts b/packages/core/src/config/plugin/skill-file.ts index 95782512b9fc..6d1672bdb8e5 100644 --- a/packages/core/src/config/plugin/skill-file.ts +++ b/packages/core/src/config/plugin/skill-file.ts @@ -1,6 +1,7 @@ export * as SkillFile from "./skill-file.js" import path from "path" +import { isRecord } from "@opencode/ai/utils/record" import { Result, Schema, type SchemaIssue, SchemaParser } from "effect" import { ConfigMarkdown } from "../markdown.js" import { AbsolutePath } from "../../schema.js" @@ -20,10 +21,8 @@ export type ParseResult = | { readonly _tag: "Skipped"; readonly reason: "frontmatter"; readonly issue: SchemaIssue.Issue } const metadataBoolean = (metadata: unknown, key: string) => { - if (metadata === undefined || metadata === null || typeof metadata !== "object" || Array.isArray(metadata)) { - return undefined - } - return booleanValue((metadata as Record)[key]) + if (!isRecord(metadata)) return undefined + return booleanValue(metadata[key]) } const booleanValue = (value: unknown) => { diff --git a/packages/core/src/database/path.ts b/packages/core/src/database/path.ts index 66a130b46d0c..8ddef89e8070 100644 --- a/packages/core/src/database/path.ts +++ b/packages/core/src/database/path.ts @@ -12,12 +12,15 @@ function isWindowsStoragePath(input: string) { return /^[A-Za-z]:\//.test(input) || input.startsWith("//") } -function absolute(input: string) { +export function storedAbsolutePath(input: string) { const result = storagePath(input) // Persisted projects and sessions can move between operating systems during migration. - if (!nodePath.posix.isAbsolute(result) && !isWindowsStoragePath(result)) { - throw new Error(`Path is not absolute: ${input}`) - } + if (nodePath.posix.isAbsolute(result) || isWindowsStoragePath(result)) return result +} + +function absolute(input: string) { + const result = storedAbsolutePath(input) + if (result === undefined) throw new Error(`Path is not absolute: ${input}`) return result } diff --git a/packages/core/src/database/v1-migration.bun.ts b/packages/core/src/database/v1-migration.bun.ts index 591c7924fb99..4d7c15df7b69 100644 --- a/packages/core/src/database/v1-migration.bun.ts +++ b/packages/core/src/database/v1-migration.bun.ts @@ -2,6 +2,7 @@ export * as V1Migration from "./v1-migration.js" import { Cause, Effect, Layer, Option, Schema, Semaphore } from "effect" import { Database } from "./database.js" +import { storedAbsolutePath } from "./path.js" import { SessionMessageTable, SessionTable } from "../session/sql.js" import { SessionV1 } from "@opencode/schema/session-v1" import { SessionMessage } from "../session/message.js" @@ -14,6 +15,7 @@ import { existsSync } from "node:fs" import path from "node:path" import type { Database as SQLiteDatabase } from "bun:sqlite" import { Project } from "@opencode/schema/project" +import { Agent } from "@opencode/schema/agent" export type SourceMessage = { readonly id: string @@ -468,7 +470,7 @@ export function transformSession(input: TransformInput): TransformResult { return { messages: migrated, session: { - agent: input.session.agent ?? (latestUser?.value.role === "user" ? latestUser.value.agent : null), + agent: input.session.agent ?? (latestUser?.value.role === "user" ? Agent.ID.make(latestUser.value.agent) : null), model: input.session.model ?? (latestUser?.value.role === "user" @@ -580,16 +582,19 @@ export function run(options: Options = {}): Effect.Effect(sql`SELECT id FROM project`)).map((project) => project.id), + const projects = new Map( + (yield* db.all<{ id: string; worktree: string }>(sql`SELECT id, worktree FROM project`)).map((project) => [ + project.id, + project.worktree, + ]), ) while (true) { const state = yield* readState(db) const cursorValue = state?.phase === "sessions" ? state.cursor : undefined - const nextID = yield* db.get<{ id: string; project_id: string }>( + const nextID = yield* db.get<{ id: string; project_id: string; directory: string }>( cursorValue === undefined - ? sql`SELECT id, project_id FROM session ORDER BY id DESC LIMIT 1` - : sql`SELECT id, project_id FROM session WHERE id < ${cursorValue} ORDER BY id DESC LIMIT 1`, + ? sql`SELECT id, project_id, directory FROM session ORDER BY id DESC LIMIT 1` + : sql`SELECT id, project_id, directory FROM session WHERE id < ${cursorValue} ORDER BY id DESC LIMIT 1`, ) if (!nextID) break yield* db @@ -609,6 +614,18 @@ export function run(options: Options = {}): Effect.Effect Effect.Effect + ) => Effect.Effect /** Resolve a read target and obtain external-directory approval before read approval. */ readonly authorizeRead: ( file: string, context: Invocation, options?: ReadOptions, - ) => Effect.Effect + ) => Effect.Effect } export class Service extends Context.Service()("@opencode/FileAccess") {} diff --git a/packages/core/src/filesystem.ts b/packages/core/src/filesystem.ts index bd81f0b160eb..ab421a7737a6 100644 --- a/packages/core/src/filesystem.ts +++ b/packages/core/src/filesystem.ts @@ -10,17 +10,15 @@ import { FileSystemSearch } from "./filesystem/search.js" import { Entry, FileSystem, FindInput, Write } from "@opencode/schema/filesystem" export { Entry, Match, Submatch } from "@opencode/schema/filesystem" -export const ReadInput = Schema.Struct({ - path: RelativePath, -}) -export type ReadInput = typeof ReadInput.Type +export interface ReadInput { + readonly path: RelativePath +} -export const WriteInput = Schema.Struct({ +export interface WriteInput { /** Absolute, or relative to the location directory. */ - path: Schema.String, - data: Schema.Uint8Array, -}) -export type WriteInput = typeof WriteInput.Type + readonly path: string + readonly data: Uint8Array +} export class NotFoundError extends Schema.TaggedError()("FileSystem.NotFoundError", { path: RelativePath, @@ -59,10 +57,9 @@ export const Content = Schema.Struct({ }).annotate({ identifier: "FileSystem.Content" }) export type Content = typeof Content.Type -export const ListInput = Schema.Struct({ - path: Schema.String.pipe(Schema.optional), -}) -export type ListInput = typeof ListInput.Type +export interface ListInput { + readonly path?: string +} export { FindInput, Write } diff --git a/packages/core/src/filesystem/search.ts b/packages/core/src/filesystem/search.ts index 7b32acee3db5..22cfb74d98f8 100644 --- a/packages/core/src/filesystem/search.ts +++ b/packages/core/src/filesystem/search.ts @@ -2,7 +2,7 @@ export * as FileSystemSearch from "./search.js" import { makeLocationNode } from "@opencode/util/effect/app-node" import path from "path" -import { Clock, Context, Deferred, Duration, Effect, Layer, Schema, Scope } from "effect" +import { Clock, Context, Deferred, Duration, Effect, Layer, Scope } from "effect" import { Fff } from "#fff" import fuzzysort from "fuzzysort" import { FileSystem } from "@opencode/schema/filesystem" @@ -15,10 +15,9 @@ export interface Interface { readonly find: (input: FileSystem.FindInput) => Effect.Effect } -export const Options = Schema.Struct({ - fff: Schema.optional(Schema.Boolean), -}) -export type Options = typeof Options.Type +export interface Options { + readonly fff?: boolean +} export class Service extends Context.Service()("@opencode/FileSystem/Search") {} @@ -50,7 +49,7 @@ function search(index: ReturnType, input: FileSystem.FindInpu }) } -export const ripgrepLayer = Layer.effect( +const ripgrepLayer = Layer.effect( Service, Effect.gen(function* () { const location = yield* Location.Service @@ -181,7 +180,7 @@ export const fffLayer = Layer.effect( }), ) -export const layer = (options?: Options) => +const layer = (options?: Options) => Layer.unwrap( Effect.gen(function* () { const location = yield* Location.Service diff --git a/packages/core/src/filesystem/watcher.ts b/packages/core/src/filesystem/watcher.ts index b5c438661d0b..f33413bfb023 100644 --- a/packages/core/src/filesystem/watcher.ts +++ b/packages/core/src/filesystem/watcher.ts @@ -3,17 +3,15 @@ export * as Watcher from "./watcher.js" // @ts-ignore import { createWrapper } from "@parcel/watcher/wrapper.js" import type ParcelWatcher from "@parcel/watcher" -import { FileSystem } from "@opencode/schema/filesystem" import { makeGlobalNode } from "@opencode/util/effect/app-node" import { FSUtil } from "@opencode/util/fs-util" -import { Cause, Context, Effect, Layer, PubSub, RcMap, Schema, Stream } from "effect" +import { Cause, Context, Effect, Layer, PubSub, RcMap, Stream } from "effect" import { lazy } from "../util/lazy.js" import { watch } from "node:fs" import path from "path" import loadBinding from "./watcher-binding.js" const SUBSCRIBE_TIMEOUT_MS = 10_000 -export const Event = { Updated: FileSystem.Event.Changed } const watcher = lazy((): typeof ParcelWatcher | undefined => { try { @@ -37,7 +35,7 @@ export type WatchInput = | { readonly path: string; readonly type: "entries"; readonly names: readonly string[] } | { readonly path: string; readonly type: "directory"; readonly ignore?: readonly string[] } -export type Subscription = { +type Subscription = { readonly unsubscribe: () => Promise /** Backend name for logging, e.g. "node" or "fs-events". */ readonly backend?: string @@ -65,14 +63,13 @@ export interface Interface { readonly subscribe: (input: WatchInput, onReady?: Effect.Effect) => Effect.Effect> } -export const Options = Schema.Struct({ - enabled: Schema.optional(Schema.Boolean), -}) -export type Options = typeof Options.Type +export interface Options { + readonly enabled?: boolean +} export class Service extends Context.Service()("@opencode/Watcher") {} -export interface TestInterface extends Interface { +interface TestInterface extends Interface { /** Delivers one update to every active watch whose target covers `update.path`. */ readonly emit: (update: Update) => Effect.Effect /** Returns every subscribe call observed so far, in order. */ @@ -226,7 +223,7 @@ export const nativeLayer = Layer.succeed( }), ) -export const nativeNode = makeGlobalNode({ service: Native, layer: nativeLayer, deps: [] }) +const nativeNode = makeGlobalNode({ service: Native, layer: nativeLayer, deps: [] }) export function configured(options?: Options) { return makeGlobalNode({ service: Service, layer: layer(options), deps: [nativeNode] }) diff --git a/packages/core/src/formatter/builtins.ts b/packages/core/src/formatter/builtins.ts index 86b3714add58..8609451c66b4 100644 --- a/packages/core/src/formatter/builtins.ts +++ b/packages/core/src/formatter/builtins.ts @@ -1,5 +1,6 @@ import { Effect } from "effect" import { ChildProcess } from "effect/process" +import { isRecord } from "@opencode/ai/utils/record" import { FSUtil } from "@opencode/util/fs-util" import { Npm } from "@opencode/util/npm" import { AppProcess } from "@opencode/util/process" @@ -291,7 +292,3 @@ function hasRecordKey(input: unknown, field: string, key: string) { if (!isRecord(input)) return false return isRecord(input[field]) && key in input[field] } - -function isRecord(input: unknown): input is Record { - return Boolean(input && typeof input === "object" && !Array.isArray(input)) -} diff --git a/packages/core/src/integration.ts b/packages/core/src/integration.ts index 8e75588ce648..dc929331db3a 100644 --- a/packages/core/src/integration.ts +++ b/packages/core/src/integration.ts @@ -147,8 +147,8 @@ export type Ref = Integration.Ref type Entry = { ref: Types.DeepMutable - methods: Types.DeepMutable[] - implementations: Map> + methods: Method[] + implementations: Map } type Data = { @@ -156,8 +156,8 @@ type Data = { } export type Editor = { - list: () => readonly Ref[] - get: (id: ID) => Ref | undefined + list: () => readonly Types.DeepMutable[] + get: (id: ID) => Types.DeepMutable | undefined update: (id: ID, update: (integration: Types.DeepMutable) => void) => void remove: (id: ID) => void method: { @@ -318,8 +318,8 @@ const layer = Layer.effect( name: "integration", initial: () => ({ integrations: new Map() }), editor: (editor) => ({ - list: () => Array.from(editor.integrations.values(), (entry) => entry.ref) as Ref[], - get: (id) => editor.integrations.get(id)?.ref as Ref | undefined, + list: () => Array.from(editor.integrations.values(), (entry) => entry.ref), + get: (id) => editor.integrations.get(id)?.ref, update: (id, update) => { const current = editor.integrations.get(id) ?? { ref: { id, name: id }, @@ -332,7 +332,7 @@ const layer = Layer.effect( }, remove: (id) => editor.integrations.delete(id), method: { - list: (integrationID) => (editor.integrations.get(integrationID)?.methods as Method[] | undefined) ?? [], + list: (integrationID) => editor.integrations.get(integrationID)?.methods ?? [], update: (implementation) => { const current = editor.integrations.get(implementation.integrationID) ?? { ref: { @@ -340,7 +340,7 @@ const layer = Layer.effect( name: implementation.integrationID, }, methods: [], - implementations: new Map>(), + implementations: new Map(), } if (!editor.integrations.has(implementation.integrationID)) { editor.integrations.set(implementation.integrationID, current) @@ -355,13 +355,10 @@ const layer = Layer.effect( return method.id === implementation.method.id return true }) - if (index === -1) current.methods.push(implementation.method as Types.DeepMutable) - else current.methods[index] = implementation.method as Types.DeepMutable - if (implementation.method.type === "oauth") { - current.implementations.set( - implementation.method.id, - implementation as Types.DeepMutable, - ) + if (index === -1) current.methods.push(implementation.method) + else current.methods[index] = implementation.method + if ("authorize" in implementation) { + current.implementations.set(implementation.method.id, implementation) } }, remove: (integrationID, method) => { diff --git a/packages/core/src/job.ts b/packages/core/src/job.ts index 771c5a79722e..7afc0f17f229 100644 --- a/packages/core/src/job.ts +++ b/packages/core/src/job.ts @@ -133,6 +133,7 @@ export interface Interface { readonly backgroundAll: (input: BackgroundAllInput) => Effect.Effect readonly cancel: (id: string) => Effect.Effect readonly pendingBackground: Effect.Effect + readonly runningBackground: Effect.Effect readonly completeBackground: (notificationID: SessionMessage.ID) => Effect.Effect } @@ -449,6 +450,15 @@ export const make = Effect.gen(function* () { return recovered }).pipe(Effect.withSpan("Job.pendingBackground")) + const runningBackground: Interface["runningBackground"] = SynchronizedRef.get(state.jobs).pipe( + Effect.map((jobs) => + [...jobs.values()].flatMap((job) => + job.isBackgrounded && job.info.status === "running" && job.recovery ? [job.recovery] : [], + ), + ), + Effect.withSpan("Job.runningBackground"), + ) + const completeBackground: Interface["completeBackground"] = Effect.fn("Job.completeBackground")((notificationID) => SynchronizedRef.updateEffect(state.jobs, (jobs) => Effect.gen(function* () { @@ -471,6 +481,7 @@ export const make = Effect.gen(function* () { backgroundAll, cancel, pendingBackground, + runningBackground, completeBackground, }) }) diff --git a/packages/core/src/location-activity.ts b/packages/core/src/location-activity.ts index 0425e9acab04..c98319027a20 100644 --- a/packages/core/src/location-activity.ts +++ b/packages/core/src/location-activity.ts @@ -1,7 +1,10 @@ export * as LocationActivity from "./location-activity.js" -import { Clock, Context, Duration, Effect, Layer, RcMap, Schema } from "effect" +import { Clock, Context, Duration, Effect, Layer, Option, RcMap, Schema } from "effect" +import type { Document } from "@opencode/schema/config" import { Bus } from "./bus.js" +import { Config } from "./config.js" +import { Job } from "./job.js" import { Location } from "./location.js" import { LocationServiceMap } from "./location-service-map.js" import { SessionEvent } from "./session/event.js" @@ -21,15 +24,32 @@ export function layer(options: { readonly timeToLive?: Duration.Input; readonly const bus = yield* Bus.Service const locations = yield* LocationServiceMap.Service const execution = yield* SessionExecution.Service + const jobs = yield* Job.Service const sessions = yield* SessionStore.Service - const timeToLive = Duration.toMillis(options.timeToLive ?? "60 minutes") - const entries = new Map() + const timeToLive = Duration.fromInputUnsafe(options.timeToLive ?? "60 minutes") + const entries = new Map() const key = (ref: Location.Ref) => `${LocationServiceMap.canonical(ref).directory}\0${ref.workspaceID ?? ""}` const touch = (ref: Location.Ref) => Effect.sync(() => { - entries.set(key(ref), { ref, expiresAt: clock.currentTimeMillisUnsafe() + timeToLive }) + entries.set(key(ref), { ref, touchedAt: clock.currentTimeMillisUnsafe() }) }) + // Read per location so project config can override the global value. + const idleTimeout = (ref: Location.Ref) => + Effect.gen(function* () { + const context = yield* locations.contextEffectOption(ref) + if (Option.isNone(context)) return timeToLive + const entries = yield* Context.get(context.value, Config.Service).entries() + const configured = entries.findLast( + (entry): entry is Document => + entry.type === "document" && entry.info.experimental?.location_idle_timeout !== undefined, + ) + return configured?.info.experimental?.location_idle_timeout ?? timeToLive + }).pipe( + Effect.scoped, + Effect.orElseSucceed(() => timeToLive), + ) + const unsubscribe = yield* bus.listen((event) => { if (!isSessionEvent(event)) return Effect.void const location = event.location @@ -48,39 +68,35 @@ export function layer(options: { readonly timeToLive?: Duration.Input; readonly if (!cached.has(id)) entries.delete(id) } const now = clock.currentTimeMillisUnsafe() - const expired = Array.from(entries.values()).filter((entry) => entry.expiresAt <= now) + const expired = yield* Effect.filter(Array.from(entries.values()), (entry) => + idleTimeout(entry.ref).pipe( + Effect.map((timeout) => timeout !== false && entry.touchedAt + Duration.toMillis(timeout) <= now), + ), + ) if (expired.length === 0) return - const active = yield* Effect.forEach(yield* execution.active, (sessionID) => sessions.get(sessionID)) + // Waiting on a question, a long tool call, or a background job emits no durable activity. + const owners = [ + ...(yield* execution.active), + ...(yield* jobs.runningBackground).flatMap((recovery) => + recovery.kind === "shell" ? [recovery.sessionID] : [recovery.childSessionID, recovery.parentSessionID], + ), + ] + const occupied = new Set( + (yield* Effect.forEach(owners, (sessionID) => sessions.get(sessionID))).flatMap((session) => + session ? [key(session.location)] : [], + ), + ) yield* Effect.forEach( expired, - (entry) => - Effect.gen(function* () { - const owners = active.flatMap((session) => - session && key(session.location) === key(entry.ref) ? [session] : [], - ) - // Invalidation only detaches the cache entry; borrowers retain the old - // graph. Stop its executions and settle tool cleanup before detaching it. - yield* Effect.forEach( - owners, - (session) => execution.interrupt(session.id, { reason: "inactivity", awaitSettlement: true }), - { - discard: true, - concurrency: "unbounded", - }, - ) - const remaining = yield* Effect.forEach(yield* execution.active, (sessionID) => sessions.get(sessionID)) - // New work admitted during cleanup may now own the cached graph. - if (remaining.some((session) => session && key(session.location) === key(entry.ref))) { - yield* touch(entry.ref) - return - } - entries.delete(key(entry.ref)) - yield* Effect.logInfo("location services evicted", { - directory: entry.ref.directory, - workspaceID: entry.ref.workspaceID, - }).pipe(Effect.andThen(locations.invalidate(entry.ref))) - }), - { discard: true, concurrency: "unbounded" }, + (entry) => { + if (occupied.has(key(entry.ref))) return touch(entry.ref) + entries.delete(key(entry.ref)) + return Effect.logInfo("location services evicted", { + directory: entry.ref.directory, + workspaceID: entry.ref.workspaceID, + }).pipe(Effect.andThen(locations.invalidate(entry.ref))) + }, + { discard: true }, ) }).pipe(Effect.forever, Effect.forkScoped) @@ -92,5 +108,5 @@ export function layer(options: { readonly timeToLive?: Duration.Input; readonly export const node = makeGlobalNode({ service: Service, layer: layer(), - deps: [Bus.node, LocationServiceMap.node, SessionExecution.node, SessionStore.node], + deps: [Bus.node, Job.node, LocationServiceMap.node, SessionExecution.node, SessionStore.node], }) diff --git a/packages/core/src/mcp/client.ts b/packages/core/src/mcp/client.ts index 0848eb4cad59..f652e741cbb1 100644 --- a/packages/core/src/mcp/client.ts +++ b/packages/core/src/mcp/client.ts @@ -81,7 +81,7 @@ export class SessionExpiredError extends Schema.TaggedError export type CallToolContent = | { readonly type: "text"; readonly text: string } - | { readonly type: "media"; readonly data: string; readonly mimeType: string } + | { readonly type: "media"; readonly data: string; readonly mimeType: string; readonly uri?: string } export interface CallToolResult { readonly isError: boolean @@ -437,7 +437,7 @@ function toCallToolResult(result: SdkCallToolResult): CallToolResult { const resource = part.resource if ("text" in resource && typeof resource.text === "string") return [{ type: "text", text: resource.text }] if ("blob" in resource && typeof resource.blob === "string" && typeof resource.mimeType === "string") - return [{ type: "media", data: resource.blob, mimeType: resource.mimeType }] + return [{ type: "media", data: resource.blob, mimeType: resource.mimeType, uri: resource.uri }] return [{ type: "text", text: resource.uri }] } return [] diff --git a/packages/core/src/mcp/index.ts b/packages/core/src/mcp/index.ts index 0c021a4c95ef..f8fe29bafb5a 100644 --- a/packages/core/src/mcp/index.ts +++ b/packages/core/src/mcp/index.ts @@ -213,8 +213,7 @@ export const layer = (options?: Options) => }) .pipe(Scope.provide(scope)) }) - const requireServer = Effect.fnUntraced(function* (server: ServerName | string) { - const name = ServerName.make(server) + const requireServer = Effect.fnUntraced(function* (name: ServerName) { const entry = entries.get(name) if (!entry) return yield* new NotFoundError({ server: name }) return { name, entry } @@ -667,7 +666,7 @@ export const layer = (options?: Options) => .toSorted((a, b) => a.server.localeCompare(b.server) || a.name.localeCompare(b.name)) }), callTool: Effect.fn("MCP.callTool")(function* (input) { - const target = yield* requireServer(input.server) + const target = yield* requireServer(ServerName.make(input.server)) yield* target.entry.startup.await if (!target.entry.client) return yield* new ToolCallError({ @@ -704,7 +703,7 @@ export const layer = (options?: Options) => .toSorted((a, b) => a.server.localeCompare(b.server) || a.name.localeCompare(b.name)) }), prompt: Effect.fn("MCP.prompt")(function* (input) { - const target = yield* requireServer(input.server) + const target = yield* requireServer(ServerName.make(input.server)) yield* target.entry.startup.await if (!target.entry.client) return undefined const result = yield* recovering(target.name, target.entry, target.entry.client, (connection) => @@ -726,13 +725,13 @@ export const layer = (options?: Options) => return mergeCatalogs(catalogs) }), resources: Effect.fn("MCP.resources")(function* (input) { - const target = yield* requireServer(input.server) + const target = yield* requireServer(ServerName.make(input.server)) yield* target.entry.startup.await if (!target.entry.client) return ResourceCatalog.make({ resources: [], templates: [] }) return mergeCatalogs([yield* loadCatalog(target.name, target.entry, target.entry.client)]) }), readResource: Effect.fn("MCP.readResource")(function* (input) { - const target = yield* requireServer(input.server) + const target = yield* requireServer(ServerName.make(input.server)) yield* target.entry.startup.await if (!target.entry.client) return undefined const result = yield* recovering(target.name, target.entry, target.entry.client, (connection) => diff --git a/packages/core/src/mcp/instructions.ts b/packages/core/src/mcp/instructions.ts index 3517b74ce9d6..b6cfd31bef99 100644 --- a/packages/core/src/mcp/instructions.ts +++ b/packages/core/src/mcp/instructions.ts @@ -35,8 +35,8 @@ const update = (previous: ReadonlyArray, current: ReadonlyArray server.server, (before, after) => before.instructions !== after.instructions || before.codemode !== after.codemode, ) - // Additions and removals render as small deltas; anything else restates the full list. - if (diff.changed.length > 0 || (diff.added.length === 0 && diff.removed.length === 0)) + // Added, changed, and removed servers render as deltas; the full list is restated only when no entry differs. + if (diff.added.length === 0 && diff.changed.length === 0 && diff.removed.length === 0) return [ "The available MCP server instructions have changed. This list supersedes the previous one.", render(current), @@ -45,6 +45,12 @@ const update = (previous: ReadonlyArray, current: ReadonlyArray change.current)), + ]), ...(diff.removed.length === 0 ? [] : [ diff --git a/packages/core/src/mime.ts b/packages/core/src/mime.ts index 20b469271af5..35f449ca86fd 100644 --- a/packages/core/src/mime.ts +++ b/packages/core/src/mime.ts @@ -16,6 +16,16 @@ export function detect(bytes: Uint8Array) { return isText(bytes) ? "text/plain" : "application/octet-stream" } +export const isTextual = (mime: string) => + !mime || + mime.startsWith("text/") || + mime === "application/json" || + mime.endsWith("+json") || + mime === "application/xml" || + mime.endsWith("+xml") || + mime === "application/javascript" || + mime === "application/x-javascript" + function startsWith(bytes: Uint8Array, prefix: number[]) { return prefix.every((value, index) => bytes[index] === value) } diff --git a/packages/core/src/model.ts b/packages/core/src/model.ts index c0246ee9bd3d..7a199b6c5d72 100644 --- a/packages/core/src/model.ts +++ b/packages/core/src/model.ts @@ -92,6 +92,7 @@ const layer = Layer.effect( const providers = yield* Provider.Service const bus = yield* Bus.Service const location = yield* Location.Service + const disableFree = ["1", "true"].includes(process.env.OPENCODE_DISABLE_FREE_MODELS?.toLowerCase() ?? "") let input: Provider.Snapshot | undefined const state: State.Interface = State.create({ name: "model", @@ -220,6 +221,11 @@ const layer = Layer.effect( ] }), ) + if (disableFree) + for (const providerID of [Provider.ID.opencode, Provider.ID.make("opencode-go")]) + byProvider.get(providerID)?.forEach((model, id, models) => { + if (!model.cost.some((cost) => cost.input > 0)) models.delete(id) + }) const all = Array.from(byProvider.values()) .flatMap((models) => Array.from(models.values())) .sort((left, right) => right.time.released - left.time.released) diff --git a/packages/core/src/models-dev/snapshot.txt b/packages/core/src/models-dev/snapshot.txt index ac61d7d29d54..8edc62d56bfc 100644 --- a/packages/core/src/models-dev/snapshot.txt +++ b/packages/core/src/models-dev/snapshot.txt @@ -1 +1 @@ -{"model-oracle-ai":{"id":"model-oracle-ai","env":["MODEL_ORACLE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.modeloracle.com/api/v1","name":"Model Oracle AI","doc":"https://modeloracle.com/setup/","models":{"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"canonical_model_id":"openai/o4-mini"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"canonical_model_id":"openai/gpt-5.5"},"auto":{"id":"auto","name":"Auto","description":"Model Oracle AI decision engine that selects and routes among configured coding-agent models","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-29","last_updated":"2026-07-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000}},"claude-opus-4.8":{"id":"claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"canonical_model_id":"anthropic/claude-opus-4-8"},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"canonical_model_id":"openai/gpt-4.1-mini"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"canonical_model_id":"deepseek/deepseek-v4-pro"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"canonical_model_id":"anthropic/claude-fable-5"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"canonical_model_id":"openai/gpt-5"},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"canonical_model_id":"openai/gpt-4.1"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"canonical_model_id":"openai/gpt-5.4"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"canonical_model_id":"openai/gpt-5.4-nano"},"claude-haiku-4.5":{"id":"claude-haiku-4.5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"canonical_model_id":"anthropic/claude-haiku-4-5"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"canonical_model_id":"openai/gpt-5.4-mini"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"canonical_model_id":"zhipuai/glm-5.2"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"canonical_model_id":"anthropic/claude-sonnet-5"}}},"venice":{"id":"venice","env":["VENICE_API_KEY"],"npm":"venice-ai-sdk-provider","name":"Venice AI","doc":"https://docs.venice.ai","models":{"venice-uncensored-role-play":{"id":"venice-uncensored-role-play","name":"Venice Role Play Uncensored","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"venice","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-02-20","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.5,"output":2}},"qwen3-235b-a22b-thinking-2507":{"id":"qwen3-235b-a22b-thinking-2507","name":"Qwen 3 235B A22B Thinking 2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"release_date":"2025-04-29","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.45,"output":3.5}},"openai-gpt-oss-120b":{"id":"openai-gpt-oss-120b","name":"OpenAI GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-11-06","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.07,"output":0.3},"canonical_model_id":"openai/gpt-oss-120b"},"kimi-k2-6":{"id":"kimi-k2-6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.75,"output":3.5,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"inkling":{"id":"inkling","name":"Inkling","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-17","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"cost":{"input":1.25,"output":5.0625,"cache_read":0.2125},"canonical_model_id":"thinkingmachines/inkling"},"deepseek-v4-flash-0731-fast":{"id":"deepseek-v4-flash-0731-fast","name":"DeepSeek V4 Flash 0731 Fast","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-08-09","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":0.35,"output":0.7,"cache_read":0.0875},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"qwen3-vl-235b-a22b":{"id":"qwen3-vl-235b-a22b","name":"Qwen3 VL 235B","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-01-16","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.21,"output":1.9,"cache_read":0.1}},"openai-gpt-54-pro":{"id":"openai-gpt-54-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"hy4-preview":{"id":"hy4-preview","name":"Hy4 Preview","description":"Preview model for early access evaluation, prototyping, and compatibility testing","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-10-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":1.0425,"output":3.12625,"cache_read":0.0525},"canonical_model_id":"tencent/hy4-preview"},"olafangensan-glm-4.7-flash-heretic":{"id":"olafangensan-glm-4.7-flash-heretic","name":"GLM 4.7 Flash Heretic","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-02-04","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":24000},"cost":{"input":0.07,"output":0.4,"cache_read":0.035}},"google-gemma-4-26b-a4b-it":{"id":"google-gemma-4-26b-a4b-it","name":"Google Gemma 4 26B A4B Instruct","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-06-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.13,"output":0.4,"cache_read":0.05},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"gemini-3-1-pro-preview":{"id":"gemini-3-1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-06-11","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":2.5,"output":15,"cache_read":0.5,"cache_write":0.5,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"cache_write":0.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5,"cache_write":0.5}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"grok-4-5":{"id":"grok-4-5","name":"Grok 4.5","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-07","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":32000},"cost":{"input":2.27,"output":6.8,"cache_read":0.34,"tiers":[{"input":4.53,"output":13.6,"cache_read":0.68,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4.53,"output":13.6,"cache_read":0.68}},"canonical_model_id":"xai/grok-4.5"},"zai-org-glm-4.7":{"id":"zai-org-glm-4.7","name":"GLM 4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-24","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":16384},"cost":{"input":0.55,"output":2.65,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"google-gemma-3-27b-it":{"id":"google-gemma-3-27b-it","name":"Google Gemma 3 27B Instruct","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-04","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":16384},"cost":{"input":0.12,"output":0.2},"canonical_model_id":"google/gemma-3-27b-it"},"nvidia-nemotron-3-ultra-550b-a55b":{"id":"nvidia-nemotron-3-ultra-550b-a55b","name":"NVIDIA Nemotron 3 Ultra","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"cost":{"input":0.625,"output":3.125,"cache_read":0.1875},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"gemma-4-uncensored":{"id":"gemma-4-uncensored","name":"Gemma 4 Uncensored","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-04-13","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.1625,"output":0.5}},"mercury-2":{"id":"mercury-2","name":"Mercury 2","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-20","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":50000},"cost":{"input":0.3125,"output":0.9375,"cache_read":0.03125},"canonical_model_id":"inception/mercury-2"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-19","last_updated":"2026-06-11","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"cost":{"input":0.7,"output":3.75,"cache_read":0.07},"canonical_model_id":"google/gemini-3-flash-preview"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":1.65,"output":3.301,"cache_read":0.33},"canonical_model_id":"deepseek/deepseek-v4-pro"},"mistral-small-3-2-24b-instruct":{"id":"mistral-small-3-2-24b-instruct","name":"Mistral Small 3.2 24B Instruct","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2026-01-15","last_updated":"2026-06-11","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.09375,"output":0.25},"canonical_model_id":"mistral/mistral-small-3.2-24b-instruct"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-10","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"grok-4-7":{"id":"grok-4-7","name":"Grok 4.7","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-16","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":200000},"cost":{"input":2.27,"output":6.8,"cache_read":0.57,"tiers":[{"input":4.53,"output":13.6,"cache_read":1.13,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4.53,"output":13.6,"cache_read":1.13}},"canonical_model_id":"xai/grok-4.7"},"qwen3-5-9b":{"id":"qwen3-5-9b","name":"Qwen 3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-05","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"cost":{"input":0.1,"output":0.15},"canonical_model_id":"alibaba/qwen3.5-9b"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-08-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":0.13,"output":0.26,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"z-ai-glm-5-turbo":{"id":"z-ai-glm-5-turbo","name":"GLM 5 Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-15","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32768},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"openai-gpt-6-luna":{"id":"openai-gpt-6-luna","name":"GPT-6 Luna","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"muse-wonder":{"id":"muse-wonder","name":"Muse Wonder","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-10-12","last_updated":"2026-10-15","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":5,"output":25,"cache_read":0.25}},"qwen-3-6-plus":{"id":"qwen-3-6-plus","name":"Qwen 3.6 Plus Uncensored","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-06","last_updated":"2026-06-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.625,"output":3.75,"cache_read":0.0625,"cache_write":0.78,"tiers":[{"input":2.5,"output":7.5,"cache_read":0.0625,"cache_write":0.78,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2.5,"output":7.5,"cache_read":0.0625,"cache_write":0.78}},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen3-6-27b":{"id":"qwen3-6-27b","name":"Qwen 3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-24","last_updated":"2026-06-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.325,"output":3.25},"canonical_model_id":"alibaba/qwen3.6-27b"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3.75,"output":18.75,"cache_read":0.375},"canonical_model_id":"moonshotai/kimi-k3"},"deepseek-v4-1-flash":{"id":"deepseek-v4-1-flash","name":"DeepSeek V4.1 Flash","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.0075},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"xiaomi-mimo-v2-6-flash":{"id":"xiaomi-mimo-v2-6-flash","name":"MiMo-V2.6-Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.175,"output":0.35,"cache_read":0.00375},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"qwen3-235b-a22b-instruct-2507":{"id":"qwen3-235b-a22b-instruct-2507","name":"Qwen 3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-29","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.75},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"qwen-3-8-27b":{"id":"qwen-3-8-27b","name":"Qwen 3.8 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-17","last_updated":"2026-08-18","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.45,"output":3.2},"canonical_model_id":"alibaba/qwen3.8-27b"},"zai-org-glm-4.6":{"id":"zai-org-glm-4.6","name":"GLM 4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2024-04-01","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":16384},"cost":{"input":0.43,"output":1.75,"cache_read":0.08},"canonical_model_id":"zhipuai/glm-4.6"},"kimi-k3-fast-api":{"id":"kimi-k3-fast-api","name":"Kimi K3 Fast","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":4.5,"output":22.5,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"openai-gpt-61-sol":{"id":"openai-gpt-61-sol","name":"GPT-6.1 Sol","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"nvidia-nemotron-3-nano-30b-a3b":{"id":"nvidia-nemotron-3-nano-30b-a3b","name":"NVIDIA Nemotron 3 Nano 30B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2026-01-27","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.075,"output":0.3},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"openai-gpt-6-sol":{"id":"openai-gpt-6-sol","name":"GPT-6 Sol","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai-gpt-56-sol":{"id":"openai-gpt-56-sol","name":"GPT-5.6 Sol","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"grok-4-20":{"id":"grok-4-20","name":"Grok 4.20","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-03-12","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":128000},"cost":{"input":1.42,"output":2.83,"cache_read":0.23,"tiers":[{"input":2.83,"output":5.67,"cache_read":0.45,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.83,"output":5.67,"cache_read":0.45}}},"qwen-3-8-2-4t-a95b":{"id":"qwen-3-8-2-4t-a95b","name":"Qwen 3.8 2.4T","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":2.5,"output":7.5,"cache_read":0.3125},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"seed-2-1-turbo":{"id":"seed-2-1-turbo","name":"Seed 2.1 Turbo","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-28","last_updated":"2026-07-24","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"cost":{"input":0.625,"output":3.125,"cache_read":0.125},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-08-29","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"venice-uncensored-1-2":{"id":"venice-uncensored-1-2","name":"Venice Uncensored 1.2","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"venice","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-04-01","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.2,"output":0.9}},"deepseek-v4-1-flash-1017":{"id":"deepseek-v4-1-flash-1017","name":"DeepSeek v4.1 Flash (1017)","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.0075},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"openai-gpt-54-mini":{"id":"openai-gpt-54-mini","name":"GPT-5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-27","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"qwen3-5-35b-a3b":{"id":"qwen3-5-35b-a3b","name":"Qwen 3.5 35B A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-25","last_updated":"2026-06-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.15,"output":1,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"hermes-3-llama-3.1-405b":{"id":"hermes-3-llama-3.1-405b","name":"Hermes 3 Llama 3.1 405b","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"hermes","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2025-09-25","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":1.1,"output":3},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-405b"},"qwen-3-7-plus":{"id":"qwen-3-7-plus","name":"Qwen 3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":2,"cache_read":0.05,"cache_write":0.625,"tiers":[{"input":1.5,"output":6,"cache_read":0.15,"cache_write":1.875,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.5,"output":6,"cache_read":0.15,"cache_write":1.875}},"canonical_model_id":"alibaba/qwen3.7-plus"},"zai-org-glm-5-1":{"id":"zai-org-glm-5-1","name":"GLM 5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":80000},"cost":{"input":1.54,"output":4.84,"cache_read":0.286},"canonical_model_id":"zhipuai/glm-5.1"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":1.65,"output":4.95,"cache_read":0.165},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"abliteration-abliterated-model-large-v2":{"id":"abliteration-abliterated-model-large-v2","name":"Abliterated Large V2","description":"Flagship model for demanding analysis, coding, and production agent workflows","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":3,"output":5,"cache_read":0.3}},"xiaomi-mimo-v2-5":{"id":"xiaomi-mimo-v2-5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-06-11","last_updated":"2026-06-11","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0.4,"output":2,"cache_read":0.08},"canonical_model_id":"xiaomi/mimo-v2.5"},"kimi-k2-7-code":{"id":"kimi-k2-7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-13","last_updated":"2026-06-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.75,"output":3.5,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"openai-gpt-56-terra-pro":{"id":"openai-gpt-56-terra-pro","name":"GPT-5.6 Terra Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"zai-org-glm-5":{"id":"zai-org-glm-5","name":"GLM 5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-11","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":32000},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"grok-4-6":{"id":"grok-4-6","name":"Grok 4.6","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-10","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":200000},"cost":{"input":2.27,"output":6.8,"cache_read":0.57,"tiers":[{"input":4.53,"output":13.6,"cache_read":1.13,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4.53,"output":13.6,"cache_read":1.13}},"canonical_model_id":"xai/grok-4.6"},"openai-gpt-52":{"id":"openai-gpt-52","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-13","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":272000,"output":65536},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai-gpt-56-luna-pro":{"id":"openai-gpt-56-luna-pro","name":"GPT-5.6 Luna Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"qwen3-5-397b-a17b":{"id":"qwen3-5-397b-a17b","name":"Qwen 3.5 397B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-16","last_updated":"2026-06-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.75,"output":4.5},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen-3-8-flash":{"id":"qwen-3-8-flash","name":"Qwen 3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.14,"output":0.49,"cache_read":0.014},"canonical_model_id":"alibaba/qwen3.8-flash"},"minimax-m3-1-flash":{"id":"minimax-m3-1-flash","name":"MiniMax M3.1 Flash","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-10-18","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1875,"output":0.75,"cache_read":0.01875},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"aion-labs-aion-3-5":{"id":"aion-labs-aion-3-5","name":"Aion 3.5","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":3.75,"output":7.5,"cache_read":0.9375}},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Gemini multimodal model for text, image, audio, video, and document tasks","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-14","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":2.5,"output":12.5,"cache_read":0.125},"canonical_model_id":"google/gemini-4-argon"},"openai-gpt-55-pro":{"id":"openai-gpt-55-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-24","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-01-15","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":198000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"z-ai-glm-5-3":{"id":"z-ai-glm-5-3","name":"GLM 5.3","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-18","last_updated":"2026-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.75,"output":5.5,"cache_read":0.325},"canonical_model_id":"zhipuai/glm-5.3"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash 0423","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":0.138,"output":0.275,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"minimax-m25":{"id":"minimax-m25","name":"MiniMax M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":32768},"cost":{"input":0.27,"output":0.95,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"openai-gpt-6-astra-pro":{"id":"openai-gpt-6-astra-pro","name":"GPT-6 Astra Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-05","last_updated":"2026-09-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}}},"qwen3-coder-480b-a35b-instruct-turbo":{"id":"qwen3-coder-480b-a35b-instruct-turbo","name":"Qwen 3 Coder 480B Turbo","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-01-27","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.35,"output":1.5,"cache_read":0.04}},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"openai-gpt-56-sol-pro":{"id":"openai-gpt-56-sol-pro","name":"GPT-5.6 Sol Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"google-gemma-4-31b-it":{"id":"google-gemma-4-31b-it","name":"Google Gemma 4 31B Instruct","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-03","last_updated":"2026-06-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.12,"output":0.36,"cache_read":0.09},"canonical_model_id":"google/gemma-4-31b-it"},"grok-build-0-1":{"id":"grok-build-0-1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}},"canonical_model_id":"xai/grok-build-0.1"},"z-ai-glm-5-3-flash":{"id":"z-ai-glm-5-3-flash","name":"GLM 5.3 Flash","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"mercury-3":{"id":"mercury-3","name":"Mercury 3","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-10-19","last_updated":"2026-10-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1,"output":0.35,"cache_read":0.01}},"claude-opus-5-5-fast":{"id":"claude-opus-5-5-fast","name":"Claude Opus 5.5 Fast","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-26","last_updated":"2026-09-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":8,"output":40,"cache_read":0.4,"cache_write":10},"canonical_model_id":"anthropic/claude-opus-5-5"},"qwen-3-7-max":{"id":"qwen-3-7-max","name":"Qwen 3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-05-22","last_updated":"2026-06-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":2.7,"output":8.05,"cache_read":0.27,"cache_write":3.35},"canonical_model_id":"alibaba/qwen3.7-max"},"mercury-2-5":{"id":"mercury-2-5","name":"Mercury 2.5","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":260000,"output":65536},"cost":{"input":0.25,"output":0.9375,"cache_read":0.025},"canonical_model_id":"inception/mercury-2.5"},"nvidia-nemotron-3-5-super-vl-120b-a12b":{"id":"nvidia-nemotron-3-5-super-vl-120b-a12b","name":"NVIDIA Nemotron 3.5 Super VL 120B","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.25,"output":1.25,"cache_read":0.125},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"llama-3.2-3b":{"id":"llama-3.2-3b","name":"Llama 3.2 3B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-10-03","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"meta/llama-3.2-3b"},"gemini-3-6-flash":{"id":"gemini-3-6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-09","last_updated":"2026-07-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.9375,"output":4.6875,"cache_read":0.09375},"canonical_model_id":"google/gemini-3.6-flash"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"qwen3-6-35b-a3b":{"id":"qwen3-6-35b-a3b","name":"Qwen 3.6 35B A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-20","last_updated":"2026-07-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.1,"output":1},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"llama-3.3-70b":{"id":"llama-3.3-70b","name":"Llama 3.3 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"release_date":"2025-04-06","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.7,"output":2.8}},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-23","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"openai-gpt-56-luna":{"id":"openai-gpt-56-luna","name":"GPT-5.6 Luna","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-04","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":160000,"output":32768},"cost":{"input":0.33,"output":0.48,"cache_read":0.16},"canonical_model_id":"deepseek/deepseek-v3.2"},"z-ai-glm-5v-turbo":{"id":"z-ai-glm-5v-turbo","name":"GLM 5V Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32768},"cost":{"input":1.5,"output":5,"cache_read":0.3},"canonical_model_id":"zhipuai/glm-5v-turbo"},"openai-gpt-56-terra":{"id":"openai-gpt-56-terra","name":"GPT-5.6 Terra","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"grok-4-20-multi-agent":{"id":"grok-4-20-multi-agent","name":"Grok 4.20 Multi-Agent","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"release_date":"2026-03-12","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":128000},"cost":{"input":1.42,"output":2.83,"cache_read":0.23,"tiers":[{"input":2.83,"output":5.67,"cache_read":0.45,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.83,"output":5.67,"cache_read":0.45}}},"gemini-3-8-flash":{"id":"gemini-3-8-flash","name":"Gemini 3.8 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.9375,"output":4.6875,"cache_read":0.09375},"canonical_model_id":"google/gemini-3.8-flash"},"qwen-3-8-max":{"id":"qwen-3-8-max","name":"Qwen 3.8 Max","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-07-22","last_updated":"2026-07-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.3125,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"gemini-3-5-flash":{"id":"gemini-3-5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-22","last_updated":"2026-06-11","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.55,"output":9.45,"cache_read":0.155,"cache_write":0.086},"canonical_model_id":"google/gemini-3.5-flash"},"muse-spark-open":{"id":"muse-spark-open","name":"Muse Spark Open","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.5625,"output":5.3125,"cache_read":0.1875},"canonical_model_id":"meta/muse-spark-open"},"openai-gpt-61-sol-ultrafast":{"id":"openai-gpt-61-sol-ultrafast","name":"GPT-6.1 Sol Ultrafast","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-10-08","last_updated":"2026-10-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":12,"output":60,"cache_read":0.6,"cache_write":15,"tiers":[{"input":24,"output":90,"cache_read":1.2,"cache_write":30,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":24,"output":90,"cache_read":1.2,"cache_write":30}}},"openai-gpt-53-codex":{"id":"openai-gpt-53-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-24","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"aion-labs-aion-3-5-mini":{"id":"aion-labs-aion-3-5-mini","name":"Aion 3.5 Mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.875,"output":1.75,"cache_read":0.225}},"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-26","last_updated":"2026-09-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"minimax-m3-preview":{"id":"minimax-m3-preview","name":"MiniMax M3 Preview","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax-m3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"release_date":"2026-06-12","last_updated":"2026-06-13","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"cost":{"input":0.3,"output":1.2,"cache_read":0.06}},"kimi-k2-5":{"id":"kimi-k2-5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2024-04","release_date":"2026-01-27","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.56,"output":3.5,"cache_read":0.22},"canonical_model_id":"moonshotai/kimi-k2.5"},"openai-gpt-4o-mini-2024-07-18":{"id":"openai-gpt-4o-mini-2024-07-18","name":"GPT-4o Mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2026-02-28","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"minimax-m27":{"id":"minimax-m27","name":"MiniMax M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":32768},"cost":{"input":0.375,"output":1.5,"cache_read":0.06875},"canonical_model_id":"minimax/MiniMax-M2.7"},"openai-gpt-4o-2024-11-20":{"id":"openai-gpt-4o-2024-11-20","name":"GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2026-02-28","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-18","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"grok-4-3":{"id":"grok-4-3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-18","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":1.42,"output":2.83,"cache_read":0.23,"tiers":[{"input":2.83,"output":5.67,"cache_read":0.45,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.83,"output":5.67,"cache_read":0.45}},"canonical_model_id":"xai/grok-4.3"},"openai-gpt-54":{"id":"openai-gpt-54","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":131072},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"claude-opus-4-8-fast":{"id":"claude-opus-4-8-fast","name":"Claude Opus 4.8 Fast","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-opus-4-8"},"openai-gpt-55":{"id":"openai-gpt-55","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":131072},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-29","last_updated":"2026-07-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-12-06","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":198000,"output":32768},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"openai-gpt-6-astra":{"id":"openai-gpt-6-astra","name":"GPT-6 Astra","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-05","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"aion-labs-aion-3-0-mini":{"id":"aion-labs-aion-3-0-mini","name":"Aion 3.0 Mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":0.875,"output":1.75,"cache_read":0.225},"canonical_model_id":"aion-labs/aion-3.0-mini"},"qwen3-next-80b":{"id":"qwen3-next-80b","name":"Qwen 3 Next 80b","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2025-04-29","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.35,"output":1.9}},"gemini-3-5-flash-lite":{"id":"gemini-3-5-flash-lite","name":"Gemini 3.5 Flash-Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-09","last_updated":"2026-07-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.375,"output":3.125,"cache_read":0.0375},"canonical_model_id":"google/gemini-3.5-flash-lite"},"claude-opus-5-fast":{"id":"claude-opus-5-fast","name":"Claude Opus 5 Fast","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-23","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-opus-5"},"zai-org-glm-5-2":{"id":"zai-org-glm-5-2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-16","last_updated":"2026-06-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"aion-labs-aion-3-0":{"id":"aion-labs-aion-3-0","name":"Aion 3.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":3.75,"output":7.5,"cache_read":0.9375},"canonical_model_id":"aion-labs/aion-3.0"},"gemini-3-7-flash":{"id":"gemini-3-7-flash","name":"Gemini 3.7 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.9375,"output":4.6875,"cache_read":0.09375},"canonical_model_id":"google/gemini-3.7-flash"}}},"jiekou":{"id":"jiekou","env":["JIEKOU_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.jiekou.ai/openai","name":"Jiekou.AI","doc":"https://docs.jiekou.ai/docs/support/quickstart?utm_source=github_models.dev","models":{"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4}},"gemini-2.5-pro-preview-06-05":{"id":"gemini-2.5-pro-preview-06-05","name":"gemini-2.5-pro-preview-06-05","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":200000},"cost":{"input":1.125,"output":9}},"grok-code-fast-1":{"id":"grok-code-fast-1","name":"grok-code-fast-1","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.18,"output":1.35}},"claude-opus-4-1-20250805":{"id":"claude-opus-4-1-20250805","name":"claude-opus-4-1-20250805","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":13.5,"output":67.5}},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"gemini-3-flash-preview","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3}},"gpt-5.2":{"id":"gpt-5.2","name":"gpt-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.575,"output":12.6}},"gpt-5-pro":{"id":"gpt-5-pro","name":"gpt-5-pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":272000},"cost":{"input":13.5,"output":108}},"grok-4-0709":{"id":"grok-4-0709","name":"grok-4-0709","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":8192},"cost":{"input":2.7,"output":13.5}},"claude-opus-4-5-20251101":{"id":"claude-opus-4-5-20251101","name":"claude-opus-4-5-20251101","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":65536},"cost":{"input":4.5,"output":22.5}},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"gemini-2.5-flash-lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.09,"output":0.36}},"grok-4-fast-non-reasoning":{"id":"grok-4-fast-non-reasoning","name":"grok-4-fast-non-reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":0.18,"output":0.45}},"o3-mini":{"id":"o3-mini","name":"o3-mini","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":1.1,"output":4.4}},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"gemini-2.5-pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":1.125,"output":9}},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"claude-opus-4-6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024,"max":127999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02","last_updated":"2026-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25}},"gpt-5-codex":{"id":"gpt-5-codex","name":"gpt-5-codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.125,"output":9}},"claude-sonnet-4-5-20250929":{"id":"claude-sonnet-4-5-20250929","name":"claude-sonnet-4-5-20250929","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":2.7,"output":13.5}},"gpt-5-mini":{"id":"gpt-5-mini","name":"gpt-5-mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.225,"output":1.8}},"grok-4-fast-reasoning":{"id":"grok-4-fast-reasoning","name":"grok-4-fast-reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":0.18,"output":0.45}},"claude-sonnet-4-20250514":{"id":"claude-sonnet-4-20250514","name":"claude-sonnet-4-20250514","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":2.7,"output":13.5}},"gemini-2.5-flash-lite-preview-09-2025":{"id":"gemini-2.5-flash-lite-preview-09-2025","name":"gemini-2.5-flash-lite-preview-09-2025","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.09,"output":0.36}},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"gpt-5.2-codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.75,"output":14}},"o3":{"id":"o3","name":"o3","description":"O-series reasoning model for hard analysis, math, coding, and planning","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":10,"output":40}},"gpt-5.1-codex-mini":{"id":"gpt-5.1-codex-mini","name":"gpt-5.1-codex-mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.225,"output":1.8}},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"gpt-5.1-codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.125,"output":9}},"gpt-5.2-pro":{"id":"gpt-5.2-pro","name":"gpt-5.2-pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":18.9,"output":151.2}},"gemini-2.5-flash-preview-05-20":{"id":"gemini-2.5-flash-preview-05-20","name":"gemini-2.5-flash-preview-05-20","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":200000},"cost":{"input":0.135,"output":3.15}},"gpt-5.1":{"id":"gpt-5.1","name":"gpt-5.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02","last_updated":"2026-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.125,"output":9}},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"claude-haiku-4-5-20251001","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":20000,"output":64000},"cost":{"input":0.9,"output":4.5}},"grok-4-1-fast-non-reasoning":{"id":"grok-4-1-fast-non-reasoning","name":"grok-4-1-fast-non-reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":0.18,"output":0.45}},"gpt-5.1-codex-max":{"id":"gpt-5.1-codex-max","name":"gpt-5.1-codex-max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.125,"output":9}},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"gemini-2.5-flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.27,"output":2.25}},"gemini-2.5-flash-lite-preview-06-17":{"id":"gemini-2.5-flash-lite-preview-06-17","name":"gemini-2.5-flash-lite-preview-06-17","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","video","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.09,"output":0.36}},"gpt-5-nano":{"id":"gpt-5-nano","name":"gpt-5-nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.045,"output":0.36}},"claude-opus-4-20250514":{"id":"claude-opus-4-20250514","name":"claude-opus-4-20250514","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":13.5,"output":67.5}},"grok-4-1-fast-reasoning":{"id":"grok-4-1-fast-reasoning","name":"grok-4-1-fast-reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":0.18,"output":0.45}},"gpt-5-chat-latest":{"id":"gpt-5-chat-latest","name":"gpt-5-chat-latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.125,"output":9}},"gemini-3-pro-preview":{"id":"gemini-3-pro-preview","name":"gemini-3-pro-preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.8,"output":10.8}},"xiaomimimo/mimo-v2-flash":{"id":"xiaomimimo/mimo-v2-flash","name":"XiaomiMiMo/MiMo-V2-Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0,"output":0}},"deepseek/deepseek-r1-0528":{"id":"deepseek/deepseek-r1-0528","name":"DeepSeek R1 0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.7,"output":2.5}},"deepseek/deepseek-v3-0324":{"id":"deepseek/deepseek-v3-0324","name":"DeepSeek V3 0324","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.28,"output":1.14}},"deepseek/deepseek-v3.1":{"id":"deepseek/deepseek-v3.1","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":32767}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.27,"output":1}},"qwen/qwen3-235b-a22b-thinking-2507":{"id":"qwen/qwen3-235b-a22b-thinking-2507","name":"Qwen3 235B A22b Thinking 2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.3,"output":3}},"qwen/qwen3-32b-fp8":{"id":"qwen/qwen3-32b-fp8","name":"Qwen3 32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":20000},"cost":{"input":0.1,"output":0.45}},"qwen/qwen3-235b-a22b-instruct-2507":{"id":"qwen/qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.15,"output":0.8}},"qwen/qwen3-235b-a22b-fp8":{"id":"qwen/qwen3-235b-a22b-fp8","name":"Qwen3 235B A22B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":20000},"cost":{"input":0.2,"output":0.8}},"qwen/qwen3-next-80b-a3b-instruct":{"id":"qwen/qwen3-next-80b-a3b-instruct","name":"Qwen3 Next 80B A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.15,"output":1.5}},"qwen/qwen3-next-80b-a3b-thinking":{"id":"qwen/qwen3-next-80b-a3b-thinking","name":"Qwen3 Next 80B A3B Thinking","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.15,"output":1.5}},"qwen/qwen3-coder-480b-a35b-instruct":{"id":"qwen/qwen3-coder-480b-a35b-instruct","name":"Qwen3 Coder 480B A35B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.29,"output":1.2}},"qwen/qwen3-coder-next":{"id":"qwen/qwen3-coder-next","name":"qwen/qwen3-coder-next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02","last_updated":"2026-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.2,"output":1.5}},"qwen/qwen3-30b-a3b-fp8":{"id":"qwen/qwen3-30b-a3b-fp8","name":"Qwen3 30B A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":20000},"cost":{"input":0.09,"output":0.45}},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 0905","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":2.5}},"moonshotai/kimi-k2-instruct":{"id":"moonshotai/kimi-k2-instruct","name":"Kimi K2 Instruct","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.57,"output":2.3}},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":262143}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3}},"baidu/ernie-4.5-vl-424b-a47b":{"id":"baidu/ernie-4.5-vl-424b-a47b","name":"ERNIE 4.5 VL 424B A47B","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ernie","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":123000,"output":16000},"cost":{"input":0.42,"output":1.25}},"baidu/ernie-4.5-300b-a47b-paddle":{"id":"baidu/ernie-4.5-300b-a47b-paddle","name":"ERNIE 4.5 300B A47B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"ernie","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":123000,"output":12000},"cost":{"input":0.28,"output":1.1}},"minimaxai/minimax-m1-80k":{"id":"minimaxai/minimax-m1-80k","name":"MiniMax M1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":40000},"cost":{"input":0.55,"output":2.2}},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"Minimax M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":131071}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2}},"zai-org/glm-4.5v":{"id":"zai-org/glm-4.5v","name":"GLM 4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glmv","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":0.6,"output":1.8}},"zai-org/glm-4.7":{"id":"zai-org/glm-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2}},"zai-org/glm-4.7-flash":{"id":"zai-org/glm-4.7-flash","name":"GLM-4.7-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.07,"output":0.4}},"zai-org/glm-4.5":{"id":"zai-org/glm-4.5","name":"GLM-4.5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.6,"output":2.2}}}},"aihubmix":{"id":"aihubmix","env":["AIHUBMIX_API_KEY"],"npm":"@aihubmix/ai-sdk-provider","name":"AIHubMix","doc":"https://docs.aihubmix.com","models":{"glm-4.6v":{"id":"glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.137,"output":0.411,"cache_read":0.0274,"tiers":[{"input":0.274,"output":0.822,"cache_read":0.0548,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6v"},"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"qwen3.7-flash":{"id":"qwen3.7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"input":991808,"output":64000},"cost":{"input":0.0282,"output":0.1128,"cache_read":0.00564,"cache_write":0.03525},"canonical_model_id":"alibaba/qwen3.7-flash"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.11268,"output":0.39438,"cache_read":0.02817},"canonical_model_id":"zhipuai/glm-5.3-flash"},"coding-glm-5.1":{"id":"coding-glm-5.1","name":"Coding GLM 5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-11","last_updated":"2026-04-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.06,"output":0.22,"cache_read":0.013}},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}}},"deepseek-v4-flash-0731-fast":{"id":"deepseek-v4-flash-0731-fast","name":"DeepSeek V4 Flash 0731 Fast","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.28,"output":1.4,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"claude-opus-4-6-think":{"id":"claude-opus-4-6-think","name":"Claude Opus 4.6 Thinking","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}}},"doubao-seed-2-0-lite-260428":{"id":"doubao-seed-2-0-lite-260428","name":"Doubao Seed 2.0 Lite 260428","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.08,"output":0.51,"cache_read":0.01692,"input_audio":1.269,"tiers":[{"input":0.13,"output":0.76,"cache_read":0.02536,"input_audio":1.902,"tier":{"type":"context","size":32000}},{"input":0.25,"output":1.52,"cache_read":0.05072,"input_audio":3.804,"tier":{"type":"context","size":128000}}]}},"coding-glm-5.3-flash-free":{"id":"coding-glm-5.3-flash-free","name":"GLM-5.3-Flash (Coding, free)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"coding-minimax-m2.7":{"id":"coding-minimax-m2.7","name":"Coding MiniMax M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128100},"cost":{"input":0.2,"output":0.2}},"hy4-preview":{"id":"hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":64000},"cost":{"input":0.845,"output":2.535,"cache_read":0.04225},"canonical_model_id":"tencent/hy4-preview"},"gemini-3.1-flash-image":{"id":"gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"interleaved":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","video","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image"},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375}},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5"},"longcat-2.0":{"id":"longcat-2.0","name":"LongCat-2.0","description":"Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.7746,"output":3.0984,"cache_read":0.015492},"canonical_model_id":"meituan/longcat-2.0"},"nemotron-3-nano-omni-30b-a3b-reasoning-free":{"id":"nemotron-3-nano-omni-30b-a3b-reasoning-free","name":"Nemotron 3 Nano Omni 30B A3B Reasoning (free)","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"glm-4.7-flash-free":{"id":"glm-4.7-flash-free","name":"GLM-4.7-Flash (free)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"claude-sonnet-4-5-think":{"id":"claude-sonnet-4-5-think","name":"Claude Sonnet 4.5 Thinking","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125,"tiers":[{"input":6.6,"output":24.75,"cache_read":0.66,"cache_write":8.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6.6,"output":24.75,"cache_read":0.66,"cache_write":8.25}},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"gpt-image-2.5-sunburst":{"id":"gpt-image-2.5-sunburst","type":"image","name":"GPT Image 2.5 Sunburst","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-sunburst"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"tiers":[{"input":0.5,"output":3,"cache_read":0.05,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":0.5,"output":3,"cache_read":0.05}}},"coding-glm-5.3-free":{"id":"coding-glm-5.3-free","name":"GLM-5.3 (Coding, free)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3"},"qwen3.5-122b-a10b":{"id":"qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.1126,"output":0.9008,"tiers":[{"input":0.2818,"output":2.2544,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"gpt-image-2.5-flare":{"id":"gpt-image-2.5-flare","type":"image","name":"GPT Image 2.5 Flare","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-flare"},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"alicloud-deepseek-v4-pro":{"id":"alicloud-deepseek-v4-pro","name":"DeepSeek V4 Pro (Alibaba Cloud)","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.69,"output":3.38,"cache_read":0.13}},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.48,"output":0.96,"cache_read":0.00384},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.142,"output":0.284,"cache_read":0.0284},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"claude-opus-4-8-think":{"id":"claude-opus-4-8-think","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"nemotron-3.5-content-safety-free":{"id":"nemotron-3.5-content-safety-free","name":"Nemotron 3.5 Content Safety (free)","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"ling-3.0-flash-free":{"id":"ling-3.0-flash-free","name":"Ling 3.0 Flash (free)","description":"Hybrid-linear MoE reasoning model with 124B total and 5.1B active parameters for efficient agents","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"inclusionai/ling-3.0-flash"},"qwen3.5-plus":{"id":"qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1096,"output":0.6576,"cache_read":0.01096,"cache_write":0.137,"tiers":[{"input":0.274,"output":1.644,"cache_read":0.0274,"cache_write":0.3425,"tier":{"type":"context","size":128000}},{"input":0.548,"output":3.288,"cache_read":0.0548,"cache_write":0.685,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.254,"output":1.524},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.95,"output":4,"cache_read":0.16}},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"coding-glm-5.1-free":{"id":"coding-glm-5.1-free","name":"Coding GLM 5.1 (free)","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm-free","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-11","last_updated":"2026-04-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0,"output":0}},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"qwen3.8-omni-flash":{"id":"qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1126,"output":0.380025,"cache_read":0.014075,"cache_write":0.175937},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"mimo-v2.6-pro-ultraspeed":{"id":"mimo-v2.6-pro-ultraspeed","name":"MiMo-V2.6-Pro-UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.8,"output":9.6,"cache_read":0.0384},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"doubao-seed-2-0-mini-260428":{"id":"doubao-seed-2-0-mini-260428","name":"Doubao Seed 2.0 Mini 260428","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.03,"output":0.28,"cache_read":0.00564,"input_audio":0.423,"tiers":[{"input":0.06,"output":0.56,"cache_read":0.01128,"input_audio":0.846,"tier":{"type":"context","size":32000}},{"input":0.11,"output":1.13,"cache_read":0.02256,"input_audio":1.692,"tier":{"type":"context","size":128000}}]}},"alicloud-glm-5.1":{"id":"alicloud-glm-5.1","name":"GLM-5.1 (Alibaba Cloud)","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-27","last_updated":"2026-03-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.84,"output":3.38,"cache_read":0.169,"cache_write":1.05625}},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-05-09","last_updated":"2026-05-09","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":64000},"cost":{"input":0.28,"output":1.69,"cache_read":0.0282,"cache_write":0.3525,"tiers":[{"input":1.13,"output":6.77,"cache_read":0.1128,"cache_write":1.41,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.13,"output":6.77,"cache_read":0.1128,"cache_write":1.41}}},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"glm-4.5v":{"id":"glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":16384},"cost":{"input":0.274,"output":1.096,"tiers":[{"input":0.548,"output":1.644,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.5v"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"openai/gpt-5.6-terra"},"hy3-preview":{"id":"hy3-preview","name":"Hy3 Preview","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":128000},"cost":{"input":0.17,"output":0.566661,"cache_read":0.051},"canonical_model_id":"tencent/hy3-preview"},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":64000},"cost":{"input":0.17,"output":1.01,"cache_read":0.0169,"cache_write":0.21125,"tiers":[{"input":0.68,"output":4.06,"cache_read":0.0676,"cache_write":0.845,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.68,"output":4.06,"cache_read":0.0676,"cache_write":0.845}}},"qwen3.5-flash":{"id":"qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.0282,"output":0.282,"cache_read":0.00282,"cache_write":0.03525,"tiers":[{"input":0.1126,"output":1.126,"cache_read":0.01126,"cache_write":0.14075,"tier":{"type":"context","size":128000}},{"input":0.169,"output":1.69,"cache_read":0.0169,"cache_write":0.21125,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.5-flash"},"kimi-k2.7-code-highspeed":{"id":"kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":1.9,"output":7.999,"cache_read":0.32167},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"qwen3.8-max-preview":{"id":"qwen3.8-max-preview","name":"Qwen3.8 Max Preview","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-07-19","last_updated":"2026-07-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.338,"output":1.014,"cache_read":0.0676,"cache_write":0.4225},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-03-20","last_updated":"2025-06-05","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}}},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}}},"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.155,"output":0.31,"cache_read":0.0031},"canonical_model_id":"xiaomi/mimo-v2.5"},"gpt-image-1.5":{"id":"gpt-image-1.5","type":"image","name":"GPT-Image-1.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":32},"canonical_model_id":"openai/gpt-image-1.5"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1126,"output":0.380025,"cache_read":0.014075,"cache_write":0.175937},"canonical_model_id":"alibaba/qwen3.8-flash"},"gemini-3.1-pro-preview-customtools":{"id":"gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}}},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":0.275,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5-1"},"step-3.7-flash":{"id":"step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.22,"output":1.32,"cache_read":0.044},"canonical_model_id":"stepfun/step-3.7-flash"},"qwen3.8-2.4t-a95b":{"id":"qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"hy3":{"id":"hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0.1562,"output":0.6248,"cache_read":0.03905},"canonical_model_id":"tencent/hy3"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":1}},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":64000},"cost":{"input":0.282,"output":1.128,"cache_read":0.0564,"cache_write":0.3525},"canonical_model_id":"alibaba/qwen3.7-plus"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.6918,"output":2.0754,"cache_read":0.023058},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":1,"output":6,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"openai/gpt-5.6-luna"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}}},"zai-glm-5.1":{"id":"zai-glm-5.1","name":"GLM-5.1 (Z.ai)","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-27","last_updated":"2026-03-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.845,"output":3.38,"cache_read":0.183112}},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.1268,"output":3.9438,"cache_read":0.2817},"canonical_model_id":"zhipuai/glm-5.3"},"qwen3.5-27b":{"id":"qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.0846,"output":0.6768,"tiers":[{"input":0.2536,"output":2.0288,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-27b"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.8-flash"},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.273974,"output":1.095896,"cache_read":0.0547948,"tiers":[{"input":0.547946,"output":2.191784,"cache_read":0.1095892,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.7"},"gpt-image-2":{"id":"gpt-image-2","type":"image","name":"GPT-Image-2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2"},"doubao-seed-2-0-pro":{"id":"doubao-seed-2-0-pro","name":"Doubao Seed 2.0 Pro","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.48,"output":2.41,"cache_read":0.09644,"tiers":[{"input":0.72,"output":3.62,"cache_read":0.144656,"tier":{"type":"context","size":32000}},{"input":1.45,"output":7.23,"cache_read":0.28932,"tier":{"type":"context","size":128000}}]}},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}}},"gemini-3-pro-image":{"id":"gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"interleaved":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":120},"canonical_model_id":"google/gemini-3-pro-image"},"xiaomi-mimo-v2.5-pro-free":{"id":"xiaomi-mimo-v2.5-pro-free","name":"Xiaomi MiMo-V2.5-Pro (free)","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo-v2.5-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-05-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":1.5,"output":9,"cache_read":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"o3":{"id":"o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.155,"output":0.31,"cache_read":0.0031},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"alicloud-deepseek-v4-flash":{"id":"alicloud-deepseek-v4-flash","name":"DeepSeek V4 Flash (Alibaba Cloud)","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.028}},"gpt-5.1-codex-mini":{"id":"gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"gemini-3.1-flash-lite-image":{"id":"gemini-3.1-flash-lite-image","type":"image","name":"Nano Banana 2 Lite","description":"Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":4096},"cost":{"input":0.25,"output":1.5,"cache_read":0.025},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"qwen3.5-35b-a3b":{"id":"qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.0564,"output":0.4512,"tiers":[{"input":0.2254,"output":1.8032,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1,"input_audio":2},"canonical_model_id":"google/gemini-4-argon"},"coding-glm-5.2-free":{"id":"coding-glm-5.2-free","name":"GLM-5.2 (Coding, free)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4-flash-vision-exp":{"id":"deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.155,"output":0.62,"cache_read":0.0031},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125,"tiers":[{"input":6.6,"output":24.75,"cache_read":0.66,"cache_write":8.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6.6,"output":24.75,"cache_read":0.66,"cache_write":8.25}},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2.2,"output":6.6,"cache_read":0.55,"tiers":[{"input":4.4,"output":13.2,"cache_read":1.1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4.4,"output":13.2,"cache_read":1.1}},"canonical_model_id":"xai/grok-4.7"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"grok-build-0.1":{"id":"grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2},"canonical_model_id":"xai/grok-build-0.1"},"muse-spark-1.3":{"id":"muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.375,"output":4.675,"cache_read":0.165},"canonical_model_id":"meta/muse-spark-1.3"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}}},"doubao-seed-2-0-code-preview":{"id":"doubao-seed-2-0-code-preview","name":"Doubao Seed 2.0 Code Preview","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.48,"output":2.41,"cache_read":0.09644,"tiers":[{"input":0.72,"output":3.62,"cache_read":0.144656,"tier":{"type":"context","size":32000}},{"input":1.45,"output":7.23,"cache_read":0.28932,"tier":{"type":"context","size":128000}}]}},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.5"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.95,"output":3.9995,"cache_read":0.160835},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"coding-xiaomi-mimo-v2.5-pro":{"id":"coding-xiaomi-mimo-v2.5-pro","name":"Coding Xiaomi MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo-v2.5-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-05-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.2,"output":0.6,"cache_read":0.04,"tiers":[{"input":0.4,"output":1.2,"cache_read":0.08,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.4,"output":1.2,"cache_read":0.08}},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"openai/gpt-5.6-sol"},"muse-spark-1.1":{"id":"muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.375,"output":4.675},"canonical_model_id":"meta/muse-spark-1.1"},"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":1000000},"cost":{"input":1.08,"output":3.0888,"cache_read":0.054},"canonical_model_id":"stepfun/step-5-preview"},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.1644,"output":0.9864,"tiers":[{"input":0.411,"output":2.466,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}}},"nemotron-3.5-lightning-free":{"id":"nemotron-3.5-lightning-free","name":"Nemotron 3.5 Lightning 30B A3B (free)","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"coding-minimax-m2.7-highspeed":{"id":"coding-minimax-m2.7-highspeed","name":"Coding MiniMax M2.7 Highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128100},"cost":{"input":0.2,"output":0.2}},"deep-deepseek-v4-flash":{"id":"deep-deepseek-v4-flash","name":"DeepSeek V4 Flash (DeepSeek)","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.154,"output":0.308,"cache_read":0.0308}},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"xiaomi-mimo-v2.5-free":{"id":"xiaomi-mimo-v2.5-free","name":"Xiaomi MiMo-V2.5 (free)","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo-v2.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-05-13","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"glm-5v-turbo":{"id":"glm-5v-turbo","name":"GLM 5 Vision Turbo","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glmv","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-09","last_updated":"2026-05-09","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":0.7042,"output":3.09848,"cache_read":0.169008}},"qwen3.6-max-preview":{"id":"qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"qwen3.6","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-05-09","last_updated":"2026-05-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":240000,"output":64000},"cost":{"input":1.27,"output":7.61,"cache_read":0.1268,"cache_write":1.585,"tiers":[{"input":2.11,"output":12.67,"cache_read":0.2112,"cache_write":2.64,"tier":{"type":"context","size":128000}}]}},"ox-alpha":{"id":"ox-alpha","name":"Ox Alpha","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"deep-deepseek-v4-pro":{"id":"deep-deepseek-v4-pro","name":"DeepSeek V4 Pro (DeepSeek)","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.478,"output":0.956,"cache_read":0.004302}},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.288,"output":1.152},"canonical_model_id":"minimax/MiniMax-M3"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.499999,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.13}},"nemotron-3-super-120b-a12b-free":{"id":"nemotron-3-super-120b-a12b-free","name":"Nemotron 3 Super 120B A12B (free)","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.48,"output":0.96,"cache_read":0.00384},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":64000},"cost":{"input":1.69,"output":5.07,"cache_read":0.169,"cache_write":2.1125},"canonical_model_id":"alibaba/qwen3.7-max"},"claude-sonnet-4-6-think":{"id":"claude-sonnet-4-6-think","name":"Claude Sonnet 4.6 Thinking","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}}},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-03-20","last_updated":"2025-06-05","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1}},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":1.5,"output":9,"cache_read":0.15},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.75,"output":4.5,"cache_read":0.075}},"glm-4.6":{"id":"glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.273974,"output":1.095896,"cache_read":0.0547948,"tiers":[{"input":0.547946,"output":2.191784,"cache_read":0.1095892,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.1268,"output":3.9438,"cache_read":0.2817},"canonical_model_id":"zhipuai/glm-5.2"},"claude-opus-4-7-think":{"id":"claude-opus-4-7-think","name":"Claude Opus 4.7 Thinking","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}}},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.155,"output":0.62,"cache_read":0.0031},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.6-flash"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.6,"output":3,"cache_read":0.1}},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-5.5-pro":{"id":"gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"qwen3-vl-plus":{"id":"qwen3-vl-plus","name":"Qwen3-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.137,"output":1.37,"cache_read":0,"tiers":[{"input":0.2054,"output":2.054,"cache_read":0,"tier":{"type":"context","size":32000}},{"input":0.411,"output":4.11,"cache_read":0,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3-vl-plus"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":128000},"cost":{"input":1.69,"output":5.07,"cache_read":0.169,"cache_write":2.1125},"canonical_model_id":"alibaba/qwen3.8-max"},"coding-xiaomi-mimo-v2.5":{"id":"coding-xiaomi-mimo-v2.5","name":"Coding Xiaomi MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo-v2.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-05-13","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.08,"output":0.4,"cache_read":0.016,"tiers":[{"input":0.16,"output":0.8,"cache_read":0.032,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.16,"output":0.8,"cache_read":0.032}},"canonical_model_id":"xiaomi/mimo-v2.5"},"qwen3-max":{"id":"qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.4508,"output":2.7048,"cache_read":0.09016,"cache_write":0.5635,"tiers":[{"input":0.902,"output":5.412,"cache_read":0.1804,"cache_write":1.1275,"tier":{"type":"context","size":32000}},{"input":1.3522,"output":8.1132,"cache_read":0.27044,"cache_write":1.69025,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3-max"},"gemini-2.5-flash-image":{"id":"gemini-2.5-flash-image","type":"image","name":"Nano Banana","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":30,"cache_read":0.03},"canonical_model_id":"google/gemini-2.5-flash-image"},"claude-opus-4-5-think":{"id":"claude-opus-4-5-think","name":"Claude Opus 4.5 Thinking","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"xiaomi-mimo-v2.5-pro":{"id":"xiaomi-mimo-v2.5-pro","name":"Xiaomi MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo-v2.5-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-05-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.1,"output":3.3,"cache_read":0.22,"tiers":[{"input":2.2,"output":6.6,"cache_read":0.44,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2.2,"output":6.6,"cache_read":0.44}},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"nemotron-3-ultra-550b-a55b-free":{"id":"nemotron-3-ultra-550b-a55b-free","name":"Nemotron 3 Ultra 550B A55B (free)","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"muse-spark-1.2":{"id":"muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.375,"output":4.675},"canonical_model_id":"meta/muse-spark-1.2"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"xiaomi-mimo-v2.5":{"id":"xiaomi-mimo-v2.5","name":"Xiaomi MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo-v2.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-05-13","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.44,"output":2.2,"cache_read":0.088,"tiers":[{"input":0.88,"output":4.4,"cache_read":0.176,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.88,"output":4.4,"cache_read":0.176}},"canonical_model_id":"xiaomi/mimo-v2.5"},"coding-minimax-m2.7-free":{"id":"coding-minimax-m2.7-free","name":"Coding MiniMax M2.7 (Free)","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax-free","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128100},"cost":{"input":0,"output":0}}}},"kilo":{"id":"kilo","env":["KILO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.kilo.ai/api/gateway","name":"Kilo Gateway","doc":"https://kilo.ai","models":{"meituan/longcat-2.0":{"id":"meituan/longcat-2.0","name":"Meituan: LongCat 2.0","description":"LongCat 2.0 is a sparse mixture-of-experts language model from Meituan, with 48B active parameters out of 1.6T total. It is suited for coding, repository-level changes, long-horizon problem solving, and agentic...","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048756,"output":262144},"cost":{"input":0.75,"output":3,"cache_read":0.015},"canonical_model_id":"meituan/longcat-2.0"},"~deepseek/deepseek-pro-latest":{"id":"~deepseek/deepseek-pro-latest","name":"DeepSeek: DeepSeek Pro Latest","description":"This model always redirects to the latest model in the DeepSeek Pro family.","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-14","last_updated":"2026-09-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":393216},"cost":{"input":0.225,"output":8,"cache_read":0.23}},"~deepseek/deepseek-flash-latest":{"id":"~deepseek/deepseek-flash-latest","name":"DeepSeek: DeepSeek Flash Latest","description":"This model always redirects to the latest model in the DeepSeek Flash family.","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-14","last_updated":"2026-09-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.07,"output":0.5,"cache_read":0.008}},"~deepseek/deepseek-v4-flash-latest":{"id":"~deepseek/deepseek-v4-flash-latest","name":"DeepSeek: DeepSeek V4 Flash Latest","description":"This model always redirects to the latest model in the DeepSeek V4 Flash family.","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-01","last_updated":"2026-08-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.01208,"output":0.12,"cache_read":0.00336}},"undi95/remm-slerp-l2-13b":{"id":"undi95/remm-slerp-l2-13b","name":"ReMM SLERP 13B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-06-30","release_date":"2023-07-22","last_updated":"2023-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":6144,"output":5529},"cost":{"input":0.35,"output":0.65},"canonical_model_id":"undi95/remm-slerp-l2-13b"},"aion-labs/aion-3.0-mini":{"id":"aion-labs/aion-3.0-mini","name":"AionLabs: Aion-3.0-Mini","description":"Aion-3.0 Mini is a multi-model roleplaying and storytelling system from AionLabs, built on the DeepSeek family of models. It uses a collaborative generation process in which multiple specialized models each...","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-14","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.7,"output":1.4,"cache_read":0.18},"canonical_model_id":"aion-labs/aion-3.0-mini"},"aion-labs/aion-rp-llama-3.1-8b":{"id":"aion-labs/aion-rp-llama-3.1-8b","name":"AionLabs: Aion-RP 1.0 (8B)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12-31","release_date":"2025-02-04","last_updated":"2025-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":29491},"cost":{"input":0.8,"output":1.6},"canonical_model_id":"aion-labs/aion-rp-llama-3.1-8b"},"aion-labs/aion-3.5":{"id":"aion-labs/aion-3.5","name":"AionLabs: Aion 3.5","description":"Aion 3.5 is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models. It uses a collaborative generation process in which multiple specialized models each...","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":3,"output":6,"cache_read":0.75}},"aion-labs/aion-3.5-mini":{"id":"aion-labs/aion-3.5-mini","name":"AionLabs: Aion 3.5 Mini","description":"Aion 3.5 Mini is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models. It is the smaller, lower-cost sibling of Aion 3.5 and uses...","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.7,"output":1.4,"cache_read":0.18}},"aion-labs/aion-2.0":{"id":"aion-labs/aion-2.0","name":"AionLabs: Aion-2.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-21","last_updated":"2025-12-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.8,"output":1.6,"cache_read":0.2},"canonical_model_id":"aion-labs/aion-2.0"},"aion-labs/aion-3.0":{"id":"aion-labs/aion-3.0","name":"AionLabs: Aion-3.0","description":"Aion-3.0 is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models. It uses a collaborative generation process in which multiple specialized models each contribute...","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-05","last_updated":"2026-05-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":3,"output":6,"cache_read":0.75},"canonical_model_id":"aion-labs/aion-3.0"},"inception/mercury-2":{"id":"inception/mercury-2","name":"Inception: Mercury 2","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":50000},"cost":{"input":0.25,"output":0.75,"cache_read":0.025},"canonical_model_id":"inception/mercury-2"},"inception/mercury-2.5":{"id":"inception/mercury-2.5","name":"Inception: Mercury 2.5","description":"Mercury 2.5 is the fastest reasoning LLM, and the latest diffusion LLM (dLLM) from Inception. Instead of generating tokens sequentially, Mercury 2.5 produces and refines multiple tokens in parallel, achieving...","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":260000,"output":65536},"cost":{"input":0.2,"output":0.75,"cache_read":0.02},"canonical_model_id":"inception/mercury-2.5"},"inception/mercury-3":{"id":"inception/mercury-3","name":"Inception: Mercury 3","description":"Mercury 3 is a diffusion-based language model from Inception, and the lab's most capable model to date. It is suited for agentic workflows and real-time applications where latency compounds, such...","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.2,"output":0.7,"cache_read":0.02}},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Claude Opus 4.8 is Anthropic's most capable generally available model in the Opus family. It supports text, image, and file inputs with text output, with reasoning support and a 1M-token...","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude Fable 5 is a Mythos-class model from Anthropic, built for autonomous knowledge work and coding. It supports text, image, and file inputs with text output, with reasoning support and...","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude Opus 5.5 is Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work, succeeding Claude Opus 5. It is particularly strong at multi-step changes in large codebases, code...","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Claude Sonnet 5.5 is Anthropic's Sonnet-class model for well-scoped everyday work, succeeding Claude Sonnet 5 as a direct upgrade. It is especially strong at building features, fixing bugs, and producing...","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Claude Haiku 5.5 is Anthropic's small, fast model for high-volume, cost-sensitive work such as summarization, subagents, and browser use. It succeeds Claude Haiku 4.5 with stronger coding, computer use, and...","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Claude Opus 5 is Anthropic’s flagship model for demanding reasoning, coding, and long-horizon agentic work. It is particularly strong at end-to-end software tasks, code review and bug finding, visual analysis...","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-opus-4.1":{"id":"anthropic/claude-opus-4.1","name":"Claude Opus 4.1 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Sonnet 5 is Anthropic's most capable Sonnet-class model, with frontier performance across coding, agents, and professional work. It supports adaptive thinking with selectable reasoning effort levels (low, medium, high, max,...","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude Fable 5.1 improves on Claude Fable 5 across the board, with the biggest gains in agentic coding, long-running agentic workflows, and knowledge work: long code refactors, front-end and visual...","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"nex-agi/nex-n2.5-pro":{"id":"nex-agi/nex-n2.5-pro","name":"Nex AGI: Nex-N2.5-Pro","description":"Nex-N2.5 is an agentic model built to turn goals into working, verified outcomes. Its core strength is agentic coding within a visual feedback loop: it can explore codebases, implement multi-file...","family":"agi","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0.075,"output":0.25,"cache_read":0.015}},"nex-agi/nex-n2.5-mini":{"id":"nex-agi/nex-n2.5-mini","name":"Nex AGI: Nex-N2.5-Mini","description":"Nex-N2.5 is an agentic model built to turn goals into working, verified outcomes. Its core strength is agentic coding within a visual feedback loop: it can explore codebases, implement multi-file...","family":"agi","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0.025,"output":0.1,"cache_read":0.0025}},"amazon/nova-premier-v1":{"id":"amazon/nova-premier-v1","name":"Amazon: Nova Premier 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2025-04-30","last_updated":"2025-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"status":"deprecated","cost":{"input":2.5,"output":12.5,"cache_read":0.625},"canonical_model_id":"amazon/nova-premier"},"amazon/nova-2-lite-v1":{"id":"amazon/nova-2-lite-v1","name":"Amazon: Nova 2 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.3,"output":2.5},"canonical_model_id":"amazon/nova-2-lite-v1"},"amazon/nova-lite-v1":{"id":"amazon/nova-lite-v1","name":"Amazon: Nova Lite 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":5120},"cost":{"input":0.06,"output":0.24},"canonical_model_id":"amazon/nova-lite-v1"},"amazon/nova-pro-v1":{"id":"amazon/nova-pro-v1","name":"Amazon: Nova Pro 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":5120},"cost":{"input":0.8,"output":3.2},"canonical_model_id":"amazon/nova-pro-v1"},"amazon/nova-micro-v1":{"id":"amazon/nova-micro-v1","name":"Amazon: Nova Micro 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":5120},"cost":{"input":0.035,"output":0.14},"canonical_model_id":"amazon/nova-micro-v1"},"~z-ai/glm-flash-latest":{"id":"~z-ai/glm-flash-latest","name":"Z.ai: GLM Flash Latest","description":"This model always redirects to the latest model in the GLM Flash family.","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.037476,"output":0.16,"cache_read":0.006}},"~z-ai/glm-latest":{"id":"~z-ai/glm-latest","name":"Z.ai: GLM Latest","description":"This model always redirects to the latest GLM model from Z.ai.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-19","last_updated":"2026-08-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.079,"output":1.4,"cache_read":0.078}},"scaledown/summarize":{"id":"scaledown/summarize","name":"ScaleDown: Summarize","description":"Summarize is an abstractive summarization model from ScaleDown. It condenses long text and documents into a fluent, faithful rewrite rather than lifting sentences from the source, and is suited for...","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":4000},"cost":{"input":0.05,"output":0}},"scaledown/extract":{"id":"scaledown/extract","name":"ScaleDown: Extract","description":"Extract is a description-driven entity extraction model from ScaleDown. Instead of a fixed set of entity types, callers describe the fields they want in plain English through `provider.options.scaledown.entities`, and the...","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":900000},"cost":{"input":0.05,"output":0}},"anthracite-org/magnum-v4-72b":{"id":"anthracite-org/magnum-v4-72b","name":"Magnum v4 72B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":4096},"cost":{"input":2.5,"output":5},"canonical_model_id":"anthracite-org/magnum-v4-72b"},"mancer/weaver":{"id":"mancer/weaver","name":"Mancer: Weaver (alpha)","description":"An attempt to recreate Claude-style verbosity, but don't expect the same level of coherence or memory. Meant for use in roleplay/narrative situations.","family":"alpha","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-06-30","release_date":"2023-08-02","last_updated":"2023-08-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":6000},"cost":{"input":0.4,"output":0.75},"canonical_model_id":"mancer/weaver"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":384000},"cost":{"input":1.6,"output":3.2,"cache_read":0.135},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows.","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.44,"output":1.32,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-r1-0528":{"id":"deepseek/deepseek-r1-0528","name":"DeepSeek R1 0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.5,"output":2.15,"cache_read":0.35},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek/deepseek-v3.2-exp":{"id":"deepseek/deepseek-v3.2-exp","name":"DeepSeek: DeepSeek V3.2 Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":147456},"cost":{"input":0.27,"output":0.41}},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro 0813 is a large-scale mixture-of-experts model from DeepSeek. This is the GA release of DeepSeek V4 Pro.","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-chat-v3.1":{"id":"deepseek/deepseek-chat-v3.1","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.25,"output":0.95,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-chat-v3.1"},"deepseek/deepseek-v3.1-terminus":{"id":"deepseek/deepseek-v3.1-terminus","name":"DeepSeek: DeepSeek V3.1 Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-22","last_updated":"2025-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":147456},"cost":{"input":0.27,"output":1}},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"DeepSeek V4 Flash Vision Exp is an experimental vision-enabled version of [DeepSeek V4 Flash 0731](https://openrouter.ai/deepseek/deepseek-v4-flash-0731) from DeepSeek, adding image understanding while matching the base model on text capabilities including agents,...","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0.2156,"output":0.6468,"cache_read":0.00686},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048575,"output":943717},"cost":{"input":0.091,"output":0.182,"cache_read":0.0182},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-chat":{"id":"deepseek/deepseek-chat","name":"DeepSeek Chat","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16000},"cost":{"input":0.2574,"output":1.0287},"canonical_model_id":"deepseek/deepseek-chat"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.28,"output":0.42,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-chat-v3-0324":{"id":"deepseek/deepseek-chat-v3-0324","name":"DeepSeek V3 0324","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07-31","release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":115200},"cost":{"input":0.29,"output":1.14,"cache_read":0.11},"canonical_model_id":"deepseek/deepseek-chat-v3-0324"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"DeepSeek V4.1 Flash 1017 is the October 17, 2026 snapshot of DeepSeek V4.1 Flash, served by DeepSeek under the `deepseek-flash` API model name. V4.1 Flash is a sparse mixture-of-experts model...","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the cost-efficient tier of the V4.1 family. DeepSeek reports that it exceeds V4 Pro on performance, speed, and task...","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"microsoft/phi-4":{"id":"microsoft/phi-4","name":"Microsoft: Phi 4","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"phi","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-01-10","last_updated":"2025-01-10","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":14745},"cost":{"input":0.056,"output":0.112}},"gryphe/mythomax-l2-13b":{"id":"gryphe/mythomax-l2-13b","name":"MythoMax 13B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2023-08-10","last_updated":"2023-08-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":3686},"cost":{"input":0.08,"output":0.11},"canonical_model_id":"gryphe/mythomax-l2-13b"},"nvidia/nemotron-3.5-content-safety":{"id":"nvidia/nemotron-3.5-content-safety","name":"Nemotron 3.5 Content Safety","description":"NVIDIA Nemotron 3.5 Content Safety is a compact 4B-parameter multimodal guardrail model from NVIDIA, fine-tuned from Google Gemma-3-4B. It moderates both inputs to and responses from LLMs and VLMs, accepting...","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.2,"output":0.2},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"nvidia/nemotron-3.5-lightning":{"id":"nvidia/nemotron-3.5-lightning","name":"Nemotron 3.5 Lightning 30B A3B","description":"NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that...","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.025,"output":0.12,"cache_read":0.0125},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-3.5-super-vl":{"id":"nvidia/nemotron-3.5-super-vl","name":"NVIDIA: Nemotron 3.5 Super VL","description":"NVIDIA Nemotron 3.5 Super VL is an open multimodal mixture-of-experts model from NVIDIA, with 12B active parameters out of 120B total. It takes text, images, documents, GUI screenshots, and video...","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.1,"output":0.4,"cache_read":0.08},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/nemotron-3-ultra-550b-a55b":{"id":"nvidia/nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it...","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":182520},"cost":{"input":0.4,"output":1.76,"cache_read":0.08},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/nemotron-3.5-super-vl:free":{"id":"nvidia/nemotron-3.5-super-vl:free","name":"NVIDIA: Nemotron 3.5 Super VL (free)","description":"NVIDIA Nemotron 3.5 Super VL is an open multimodal mixture-of-experts model from NVIDIA, with 12B active parameters out of 120B total. It takes text, images, documents, GUI screenshots, and video... **Terms of service** For NVIDIA free endpoints (Super/Ultra/etc): Trial use only - do not submit personal or confidential data. Your use is logged for security purposes and to improve NVIDIA products and services. The logged session data for improvement purposes is not linked to your identity or any persistent identifier. For more information about our data processing practices, see our [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/). By interacting with this endpoint, you consent to our collection, recording, and use of such information and the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/nemotron-3.5-content-safety:free":{"id":"nvidia/nemotron-3.5-content-safety:free","name":"NVIDIA: Nemotron 3.5 Content Safety (free)","description":"NVIDIA Nemotron 3.5 Content Safety is a compact 4B-parameter multimodal guardrail model from NVIDIA, fine-tuned from Google Gemma-3-4B. It moderates both inputs to and responses from LLMs and VLMs, accepting... **Terms of service** For NVIDIA free endpoints (Super/Ultra/etc): Trial use only - do not submit personal or confidential data. Your use is logged for security purposes and to improve NVIDIA products and services. The logged session data for improvement purposes is not linked to your identity or any persistent identifier. For more information about our data processing practices, see our [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/). By interacting with this endpoint, you consent to our collection, recording, and use of such information and the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"nvidia/nemotron-3.5-lightning:free":{"id":"nvidia/nemotron-3.5-lightning:free","name":"NVIDIA: Nemotron 3.5 Lightning (free)","description":"NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that... **Terms of service** For NVIDIA free endpoints (Super/Ultra/etc): Trial use only - do not submit personal or confidential data. Your use is logged for security purposes and to improve NVIDIA products and services. The logged session data for improvement purposes is not linked to your identity or any persistent identifier. For more information about our data processing practices, see our [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/). By interacting with this endpoint, you consent to our collection, recording, and use of such information and the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-3-nano-30b-a3b":{"id":"nvidia/nemotron-3-nano-30b-a3b","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.05,"output":0.2},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/nemotron-3-super-120b-a12b":{"id":"nvidia/nemotron-3-super-120b-a12b","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.08,"output":0.45},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/nemotron-3-ultra-550b-a55b:free":{"id":"nvidia/nemotron-3-ultra-550b-a55b:free","name":"NVIDIA: Nemotron 3 Ultra (free)","description":"NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it... **Terms of service** For NVIDIA free endpoints (Super/Ultra/etc): Trial use only - do not submit personal or confidential data. Your use is logged for security purposes and to improve NVIDIA products and services. The logged session data for improvement purposes is not linked to your identity or any persistent identifier. For more information about our data processing practices, see our [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/). By interacting with this endpoint, you consent to our collection, recording, and use of such information and the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/nemotron-3-super-120b-a12b:free":{"id":"nvidia/nemotron-3-super-120b-a12b:free","name":"NVIDIA: Nemotron 3 Super (free)","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free":{"id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free","name":"NVIDIA: Nemotron 3 Nano Omni (free)","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"z-ai/glm-4.6v":{"id":"z-ai/glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.055},"canonical_model_id":"zhipuai/glm-4.6v"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks. Its hybrid sparse and linear attention architecture maintains accurate long-context behavior while...","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.3-prime":{"id":"z-ai/glm-5.3-prime","name":"Z.ai: GLM 5.3 Prime","description":"GLM-5.3-Prime is the high-speed variant of Z.ai's GLM-5.3, inheriting its full capabilities while delivering 1.5–2× the output throughput through inference acceleration. It supports text input and output with a 1M-token...","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.8,"output":8.8,"cache_read":0.56}},"z-ai/glm-4.5v":{"id":"z-ai/glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":0.6,"output":1.8,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5v"},"z-ai/glm-4.5-air":{"id":"z-ai/glm-4.5-air","name":"GLM-4.5-Air","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.13,"output":0.85,"cache_read":0.025},"canonical_model_id":"zhipuai/glm-4.5-air"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"GLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks. It supports text input and output with a 1M-token context window, and improves...","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.54,"output":1.98,"cache_read":0.099},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-4.7-flash":{"id":"z-ai/glm-4.7-flash","name":"GLM-4.7-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.0605,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":128000},"cost":{"input":0.6,"output":1.92,"cache_read":0.12},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-5v-turbo":{"id":"z-ai/glm-5v-turbo","name":"GLM-5V-Turbo","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"z-ai/glm-5.3-flashx":{"id":"z-ai/glm-5.3-flashx","name":"Z.ai: GLM 5.3 FlashX","description":"GLM-5.3-FlashX is the high-speed variant of Z.ai's GLM-5.3-Flash, a native multimodal model delivering inference speeds of up to 200 tokens/s. Built on the same hybrid sparse and linear attention architecture...","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.37,"output":1.25,"cache_read":0.09}},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM-5-Turbo","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"z-ai/glm-4.6":{"id":"z-ai/glm-4.6","name":"GLM-4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.55,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"GLM 5.2 is a large-scale reasoning model from Z.ai. It supports text input and output with a 1M-token context window, and is suited for long-horizon agent workflows, project-level software engineering,...","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-4.5":{"id":"z-ai/glm-4.5","name":"GLM-4.5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5"},"inclusionai/ling-3.0-flash-vl":{"id":"inclusionai/ling-3.0-flash-vl","name":"inclusionAI: Ling 3.0 Flash VL","description":"Ling 3.0 Flash VL builds on Ling 3.0 Flash (124B total / 5.5B active MoE from InclusionAI), further strengthening its language capabilities while adding native visual perception and advanced visual...","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012},"canonical_model_id":"inclusionai/ling-3.0-flash-vl"},"inclusionai/ling-3.0-flash":{"id":"inclusionai/ling-3.0-flash","name":"inclusionAI: Ling 3.0 Flash","description":"*Ling-3.0-flash* is a *124B-parameter Mixture-of-Experts (MoE) model*, with approximately *5.1B parameters activated per token*. The model is designed with *token efficiency and production-scale agentic inference* as key priorities, enabling developers...","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012},"canonical_model_id":"inclusionai/ling-3.0-flash"},"inclusionai/ling-3.1-flash":{"id":"inclusionai/ling-3.1-flash","name":"inclusionAI: Ling 3.1 Flash","description":"Ling 3.1 Flash is a hybrid reasoning mixture-of-experts model from inclusionAI, with 25B active parameters out of 560B total.","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.3,"output":0.9,"cache_read":0.06},"canonical_model_id":"inclusionai/ling-3.1-flash"},"meta/muse-spark-1.3-contributor":{"id":"meta/muse-spark-1.3-contributor","name":"Meta: Muse Spark 1.3 Contributor","description":"Muse Spark 1.3 Contributor is the cost-efficient contributor tier of Meta’s multimodal reasoning model for experimentation, learning, and early-stage agentic, multi-agent, and coding workflows. It is designed to track information...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/muse-wonder":{"id":"meta/muse-wonder","name":"Meta: Muse Wonder","description":"Muse Wonder is a frontier reasoning model from Meta, designed for complex personal and professional agentic tasks. It accepts text, images, video, audio, and PDF documents and returns text, with...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":4,"output":20,"cache_read":0.2}},"meta/muse-glimmer-30b":{"id":"meta/muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer 30B is a dense, open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark and optimized for autonomous agents on consumer hardware. It is suited for long-horizon...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.3,"output":1.2,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"meta/muse-spark-1.2-contributor":{"id":"meta/muse-spark-1.2-contributor","name":"Meta: Muse Spark 1.2 Contributor","description":"Muse Spark 1.2 contributor tier is a reasoning model from Meta designed for developers who want to start building at an even lower cost. It’s meaningfully cheaper than Muse Spark...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-08-21","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It is designed to keep track of information across extended tasks, work through...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark 1.1 is a multimodal reasoning model from Meta, built for agentic tasks. It accepts text, images, video, audio, and PDF documents and returns text, with a 1M-token context...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-open":{"id":"meta/muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is a coding-optimized open-weight model. Trained specifically for agentic coding workflows, it delivers higher first-attempt accuracy, reliable tool calling, and deep codebase understanding across the repos developers...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a reasoning model from Meta, designed for complex agentic tasks. It accepts text, images, video, audio, and PDF documents, returns text, and offers a 1M-token context...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"bytedance-seed/seed-1.6-flash":{"id":"bytedance-seed/seed-1.6-flash","name":"ByteDance Seed: Seed 1.6 Flash (retires Nov 11)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.075,"output":0.3},"canonical_model_id":"bytedance-seed/seed-1-6-flash"},"bytedance-seed/seed-2-1-turbo":{"id":"bytedance-seed/seed-2-1-turbo","name":"ByteDance Seed: Seed 2.1 Turbo","description":"Seed 2.1 Turbo is a multimodal model from ByteDance Seed for coding and long-horizon agent workflows. It is suited for end-to-end software delivery, multi-step task execution, and understanding visual and...","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0.5,"output":2.5},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"bytedance-seed/seed-2.0-mini":{"id":"bytedance-seed/seed-2.0-mini","name":"Seed 2.0 Mini","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.1,"output":0.4},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"bytedance-seed/seed-1.6":{"id":"bytedance-seed/seed-1.6","name":"ByteDance Seed: Seed 1.6 (retires Nov 11)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.25,"output":2},"canonical_model_id":"bytedance-seed/seed-1-6"},"bytedance-seed/seed-2.0-code":{"id":"bytedance-seed/seed-2.0-code","name":"Seed 2.0 Code","description":"Seed 2.0 Code is a model from ByteDance Seed optimized for agentic coding. It is suited for frontend development, multilingual programming tasks, and coding-agent workflows in tools such as Claude...","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"status":"deprecated","cost":{"input":0.5,"output":3},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"bytedance-seed/seed-2.0-lite":{"id":"bytedance-seed/seed-2.0-lite","name":"Seed 2.0 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.25,"output":2},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"meta-llama/llama-3.1-8b-instruct":{"id":"meta-llama/llama-3.1-8b-instruct","name":"Llama-3.1-8B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.02,"output":0.04},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta-llama/llama-3.2-3b-instruct":{"id":"meta-llama/llama-3.2-3b-instruct","name":"Meta: Llama 3.2 3B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":80000,"output":72000},"cost":{"input":0.05,"output":0.33}},"meta-llama/llama-3.3-70b-instruct":{"id":"meta-llama/llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.1,"output":0.32},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/llama-4-maverick":{"id":"meta-llama/llama-4-maverick","name":"Meta: Llama 4 Maverick","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.1875,"output":0.6525,"cache_read":0.05},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"meta-llama/llama-4-scout":{"id":"meta-llama/llama-4-scout","name":"Meta: Llama 4 Scout","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":327680,"output":16384},"cost":{"input":0.08,"output":0.24},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"meta-llama/llama-3.2-1b-instruct":{"id":"meta-llama/llama-3.2-1b-instruct","name":"Llama 3.2 1B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":60000,"output":54000},"cost":{"input":0.027,"output":0.201},"canonical_model_id":"meta/llama-3.2-1b-instruct"},"meta-llama/llama-guard-4-12b":{"id":"meta-llama/llama-guard-4-12b","name":"Meta: Llama Guard 4 12B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-30","last_updated":"2025-04-30","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":16384},"cost":{"input":0.144,"output":0.144}},"meta-llama/llama-3.1-70b-instruct":{"id":"meta-llama/llama-3.1-70b-instruct","name":"Llama-3.1-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.4,"output":0.4},"canonical_model_id":"meta/llama-3.1-70b-instruct"},"qwen/qwen3.5-plus-02-15":{"id":"qwen/qwen3.5-plus-02-15","name":"Qwen: Qwen3.5 Plus 2026-02-15","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.26,"output":1.56}},"qwen/qwen3-235b-a22b-thinking-2507":{"id":"qwen/qwen3-235b-a22b-thinking-2507","name":"Qwen: Qwen3 235B A22B Thinking 2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.45,"output":3.5}},"qwen/qwen3.7-flash":{"id":"qwen/qwen3.7-flash","name":"Qwen3.7 Flash","description":"Qwen3.7 Flash is a vision-language reasoning model from Alibaba. It is suited for multimodal agents, visual coding, search, and computer interaction, with strengths in object recognition, spatial understanding, and real-world...","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":65536},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.038},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen/qwen3-vl-30b-a3b-instruct":{"id":"qwen/qwen3-vl-30b-a3b-instruct","name":"Qwen3 VL 30B A3B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"qwen/qwen3.5-9b":{"id":"qwen/qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"status":"deprecated","cost":{"input":0.1,"output":0.15},"canonical_model_id":"alibaba/qwen3.5-9b"},"qwen/qwen-2.5-7b-instruct":{"id":"qwen/qwen-2.5-7b-instruct","name":"Qwen: Qwen2.5 7B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-10-16","last_updated":"2024-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":29491},"cost":{"input":0.1,"output":0.2}},"qwen/qwen3-vl-235b-a22b-instruct":{"id":"qwen/qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.21,"output":1.9,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"qwen/qwen-plus":{"id":"qwen/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.26,"output":0.78,"cache_read":0.052,"cache_write":0.325},"canonical_model_id":"alibaba/qwen-plus"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.26,"output":2.08},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3.6-35b-a3b":{"id":"qwen/qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.1,"output":0.95,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3.8-omni-flash":{"id":"qwen/qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen3.8 Omni Flash is an omni-modal reasoning model from Alibaba, the first Qwen model built around agentic capabilities with native audio-video understanding. It is suited for audio-video analysis and summarization,...","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"qwen/qwen3-235b-a22b-2507":{"id":"qwen/qwen3-235b-a22b-2507","name":"Qwen: Qwen3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.09,"output":0.55},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.325,"output":1.95,"cache_write":0.40625},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1875,"output":1.125,"cache_write":0.234375},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3-coder-flash":{"id":"qwen/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.195,"output":0.975,"cache_read":0.039,"cache_write":0.24375},"canonical_model_id":"alibaba/qwen3-coder-flash"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen3.8 Flash is a multimodal reasoning model from Alibaba. It is suited for coding assistance, agentic workflows, visual understanding, document and codebase analysis, desktop interaction, chart analysis, and long-video analysis.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3.8-max-prime":{"id":"qwen/qwen3.8-max-prime","name":"Qwen 3.8 Max Prime","description":"Qwen3.8 Max Prime is a higher-throughput variant of Qwen3.8 Max from Alibaba's Qwen team, served as a separate SKU at a higher price point. It accepts text, image, and video...","family":"qwen3.8-max","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":4,"output":12,"cache_read":0.5},"canonical_model_id":"alibaba/qwen3.8-max-prime"},"qwen/qwen-2.5-coder-32b-instruct":{"id":"qwen/qwen-2.5-coder-32b-instruct","name":"Qwen2.5 Coder 32B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":29491},"cost":{"input":0.66,"output":1},"canonical_model_id":"alibaba/qwen2.5-coder-32b-instruct"},"qwen/qwen3-30b-a3b":{"id":"qwen/qwen3-30b-a3b","name":"Qwen3 30B A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.5},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"qwen/qwen3.8-2.4t-a95b":{"id":"qwen/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Qwen3.7-Plus is a cost-effective model in Alibaba's Qwen3.7 series. It supports text and image input with text output, building on the series' text capabilities with a comprehensive upgrade to its...","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.32,"output":1.28,"cache_read":0.064,"cache_write":0.4},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3-vl-30b-a3b-thinking":{"id":"qwen/qwen3-vl-30b-a3b-thinking","name":"Qwen: Qwen3 VL 30B A3B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0.29,"output":1}},"qwen/qwen3-next-80b-a3b-instruct":{"id":"qwen/qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.09,"output":1.1},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.195,"output":1.56},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3-next-80b-a3b-thinking":{"id":"qwen/qwen3-next-80b-a3b-thinking","name":"Qwen3-Next 80B-A3B (Thinking)","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"qwen/qwen3.5-plus-20260420":{"id":"qwen/qwen3.5-plus-20260420","name":"Qwen: Qwen3.5 Plus 2026-04-20","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":1.8,"cache_write":0.375}},"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.1625,"output":1.3},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3-30b-a3b-instruct-2507":{"id":"qwen/qwen3-30b-a3b-instruct-2507","name":"Qwen: Qwen3 30B A3B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-29","last_updated":"2025-07-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0.09,"output":0.3}},"qwen/qwen3-32b":{"id":"qwen/qwen3-32b","name":"Qwen3 32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.08,"output":0.28},"canonical_model_id":"alibaba/qwen3-32b"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Qwen3.8 27B is an open-weight dense vision-language model from Qwen. It is suited for coding, professional workflows, research, multimodal interaction, and long-running agent tasks, with flexible thinking that can be...","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.425,"output":2.55,"cache_read":0.085,"cache_write":0.53125},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3-coder-30b-a3b-instruct":{"id":"qwen/qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.07,"output":0.28},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.39,"output":2.34},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3-vl-8b-instruct":{"id":"qwen/qwen3-vl-8b-instruct","name":"Qwen: Qwen3 VL 8B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-14","last_updated":"2025-10-14","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0.25,"output":0.75,"cache_read":0.12}},"qwen/qwen3-coder-next":{"id":"qwen/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.12,"output":0.8,"cache_read":0.07},"canonical_model_id":"alibaba/qwen3-coder-next"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.45,"output":2.7},"canonical_model_id":"alibaba/qwen3.6-27b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.475,"output":4.425,"cache_read":0.295,"cache_write":1.84375},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3-coder":{"id":"qwen/qwen3-coder","name":"Qwen: Qwen3 Coder 480B A35B","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":1,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"qwen/qwen2.5-vl-72b-instruct":{"id":"qwen/qwen2.5-vl-72b-instruct","name":"Qwen: Qwen2.5 VL 72B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-01-26","last_updated":"2025-01-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":115200},"cost":{"input":0.8,"output":1,"cache_read":0.4},"canonical_model_id":"alibaba/qwen2-5-vl-72b-instruct"},"qwen/qwen-2.5-72b-instruct":{"id":"qwen/qwen-2.5-72b-instruct","name":"Qwen2.5 72B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-09-19","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":16384},"cost":{"input":0.36,"output":0.4}},"qwen/qwen3-14b":{"id":"qwen/qwen3-14b","name":"Qwen: Qwen3 14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.24}},"qwen/qwen3.8-max-0902":{"id":"qwen/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"Qwen3.8 Max 0902 is an updated snapshot of Qwen3.8 Max from Alibaba's Qwen team. It is a 2.4-trillion-parameter mixture-of-experts model that accepts text, image, and video input and returns text,...","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"qwen/qwen3.5-flash-02-23":{"id":"qwen/qwen3.5-flash-02-23","name":"Qwen: Qwen3.5-Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-25","last_updated":"2026-02-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.065,"output":0.26}},"inference-net/schematron-v2-small":{"id":"inference-net/schematron-v2-small","name":"Inference.net: Schematron V2 Small","description":"Schematron V2 Small is a 3B-parameter HTML-to-JSON extraction model from Inference.net. It prioritizes extraction quality for complex schemas and long pages. Extraction instructions must be supplied through a JSON schema...","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":0.05,"output":0.23,"cache_read":0.05},"canonical_model_id":"inference-net/schematron-v2-small"},"inference-net/schematron-v2-turbo":{"id":"inference-net/schematron-v2-turbo","name":"Inference.net: Schematron V2 Turbo","description":"Schematron V2 Turbo is a 3B-parameter HTML-to-JSON extraction model from Inference.net. It prioritizes throughput for high-volume extraction workloads. Extraction instructions must be supplied through a JSON schema in response_format rather...","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.03,"output":0.15,"cache_read":0.03},"canonical_model_id":"inference-net/schematron-v2-turbo"},"~x-ai/grok-latest":{"id":"~x-ai/grok-latest","name":"xAI: Grok Latest","description":"This model always redirects to the latest Grok model from xAI.","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5}},"~unbiased/pareto-latest":{"id":"~unbiased/pareto-latest","name":"Pareto Latest","description":"This model always redirects to the latest model in the Pareto family.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.8,"output":3.2,"cache_read":0.03}},"thedrummer/cydonia-24b-v4.1":{"id":"thedrummer/cydonia-24b-v4.1","name":"TheDrummer: Cydonia 24B V4.1","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04-30","release_date":"2025-09-27","last_updated":"2025-09-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.3,"output":0.5,"cache_read":0.15},"canonical_model_id":"thedrummer/cydonia-24b-v4.1"},"thedrummer/unslopnemo-12b":{"id":"thedrummer/unslopnemo-12b","name":"TheDrummer: UnslopNemo 12B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04-30","release_date":"2024-11-08","last_updated":"2024-11-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":819200},"cost":{"input":0.4,"output":0.4},"canonical_model_id":"thedrummer/unslopnemo-12b"},"thedrummer/skyfall-36b-v2":{"id":"thedrummer/skyfall-36b-v2","name":"TheDrummer: Skyfall 36B V2","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2025-03-10","last_updated":"2025-03-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":29491},"cost":{"input":0.55,"output":0.8,"cache_read":0.25},"canonical_model_id":"thedrummer/skyfall-36b-v2"},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 Instruct 0905","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.6,"output":2.5},"canonical_model_id":"moonshotai/kimi-k2-0905"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.65,"output":3.41,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI. It is suited for complex coding, knowledge work, and long-horizon agentic workflows, and is particularly strong at...","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2":{"id":"moonshotai/kimi-k2","name":"Kimi K2 0711","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-12-31","release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.57,"output":2.3},"canonical_model_id":"moonshotai/kimi-k2"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"MoonshotAI: Kimi K2.7 Code is a coding-focused model in Moonshot AI's Kimi K2 family, built to complete end-to-end programming tasks reliably over long contexts. It uses a native multimodal mixture-of-experts...","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.6712,"output":3.35,"cache_read":0.18},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.6,"output":3},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.6,"output":2.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"writer/palmyra-x5":{"id":"writer/palmyra-x5","name":"Writer: Palmyra X5","description":"General-purpose chat model for instruction following, writing, and analysis","family":"palmyra","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1040000,"output":8192},"cost":{"input":0.6,"output":6},"canonical_model_id":"writer/palmyra-x5"},"xingchen-agi/xing4.0-29b-a4b:free":{"id":"xingchen-agi/xing4.0-29b-a4b:free","name":"XingChen: Xing4.0 29B A4B (free)","description":"Xing4.0-29B-A4B is an open-weight mixture-of-experts language model from China Telecom's AI team (TeleAI) and the successor to the TeleChat series. It has 29B total parameters with 4B active per token...","family":"agi","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0,"output":0}},"poolside/laguna-s-2.1:free":{"id":"poolside/laguna-s-2.1:free","name":"Poolside: Laguna S 2.1 (free)","description":"Laguna S 2.1 is the latest coding agent model from [Poolside](). Laguna S 2.1 is a 118B total parameter model with 8B active parameters, scoring 70.2% on Terminal-Bench 2.1 and...","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-s-2.1"},"poolside/laguna-xs-2.1:free":{"id":"poolside/laguna-xs-2.1:free","name":"Poolside: Laguna XS 2.1 (free)","description":"Laguna XS 2.1 is the latest coding agent model in the 33B-A3B category from [Poolside](https://poolside.ai/) and a step forward from their Laguna XS.2 model (released in April 2026). It combines...","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-02","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-xs-2.1"},"poolside/laguna-xs-2.1":{"id":"poolside/laguna-xs-2.1","name":"Poolside: Laguna XS 2.1 (retires Oct 31)","description":"Laguna XS 2.1 is the latest coding agent model in the 33B-A3B category from [Poolside](https://poolside.ai/) and a step forward from their Laguna XS.2 model (released in April 2026). It combines...","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-02","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.1,"output":0.2,"cache_read":0.05},"canonical_model_id":"poolside/laguna-xs-2.1"},"poolside/laguna-s-2.1":{"id":"poolside/laguna-s-2.1","name":"Poolside: Laguna S 2.1 (retires Oct 31)","description":"Laguna S 2.1 is the latest coding agent model from [Poolside](). Laguna S 2.1 is a 118B total parameter model with 8B active parameters, scoring 70.2% on Terminal-Bench 2.1 and...","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.1,"output":0.2,"cache_read":0.01},"canonical_model_id":"poolside/laguna-s-2.1"},"fireworks/ember-1":{"id":"fireworks/ember-1","name":"Fireworks: Ember-1","description":"Ember-1 is a specialized reasoning model from Fireworks Research, built on [Kimi K3](https://openrouter.ai/moonshotai/kimi-k3). It is designed to make every token go further: it produces shorter reasoning traces, using roughly 40%...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"fireworks/ember-1"},"bytedance/ui-tars-1.5-7b":{"id":"bytedance/ui-tars-1.5-7b","name":"ByteDance: UI-TARS 7B ","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":2048},"cost":{"input":0.1,"output":0.2,"cache_read":0.1}},"sao10k/l3-lunaris-8b":{"id":"sao10k/l3-lunaris-8b","name":"Sao10K: Llama 3 8B Lunaris","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-13","last_updated":"2024-08-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":7372},"cost":{"input":0.04,"output":0.05},"canonical_model_id":"sao10k/l3-lunaris-8b"},"sao10k/l3.1-euryale-70b":{"id":"sao10k/l3.1-euryale-70b","name":"Sao10K: Llama 3.1 Euryale 70B v2.2","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-28","last_updated":"2024-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.85,"output":0.85},"canonical_model_id":"sao10k/l3.1-euryale-70b"},"dots-studio/dots-3-note-preview:free":{"id":"dots-studio/dots-3-note-preview:free","name":"Dots Studio: Dots3-Note Preview (free)","description":"Dots3-Note Preview is an open-weight mixture-of-experts model from Dots Studio, with 16B active parameters out of 280B total. It is the lightest model in the Dots 3 family and is...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":512000,"output":460800},"cost":{"input":0,"output":0}},"~moonshotai/kimi-latest":{"id":"~moonshotai/kimi-latest","name":"MoonshotAI: Kimi Latest","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.238,"output":16,"cache_read":0.228}},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Inkling is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 41B active parameters out of 975B total. It is designed for general-purpose reasoning, coding, agentic and tool-use systems,...","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":471859},"cost":{"input":0.95,"output":4.05,"cache_read":0.16},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/inkling-small:free":{"id":"thinkingmachines/inkling-small:free","name":"Thinking Machines: Inkling Small (free)","description":"Inkling Small is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 12B active parameters out of 276B total. It is positioned as the smaller, more efficient member of...","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Inkling Small is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 12B active parameters out of 276B total. It is positioned as the smaller, more efficient member of...","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":262144},"cost":{"input":0.45,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"kilo-auto/balanced":{"id":"kilo-auto/balanced","name":"Auto Balanced","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"1970-01-01","last_updated":"1970-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.325,"output":1.95,"reasoning":0,"cache_read":0.0325,"cache_write":0.40625}},"kilo-auto/small":{"id":"kilo-auto/small","name":"Auto Small","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"1970-01-01","last_updated":"1970-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.05,"output":0.4,"reasoning":0,"cache_read":0.005}},"kilo-auto/frontier":{"id":"kilo-auto/frontier","name":"Auto Frontier","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"1970-01-01","last_updated":"1970-01-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"reasoning":0,"cache_read":0.5,"cache_write":6.25}},"kilo-auto/free":{"id":"kilo-auto/free","name":"Auto Free","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"1970-01-01","last_updated":"1970-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32768},"cost":{"input":0,"output":0,"reasoning":0,"cache_read":0,"cache_write":0}},"kilo-auto/efficient":{"id":"kilo-auto/efficient","name":"Auto Efficient","description":"Routes each request to the cheapest model that gets the job done, based on continuously benchmarked accuracy and cost.","family":"auto","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"1970-01-01","last_updated":"1970-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.325,"output":1.95,"reasoning":0,"cache_read":0.0325,"cache_write":0.40625}},"reflection/beam-free":{"id":"reflection/beam-free","name":"Reflection: Beam 1.0 (free)","description":"Your prompts and completions may be retained and used to train or improve the provider's services. Beam is Reflection's reasoning model for coding and agentic workflows, combining strong coding capabilities with fast, low-latency inference, served through Vercel AI Gateway at no cost. It supports function calling, structured outputs, and streaming, and takes text input only. Zero Data Retention is not available for this model.","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"reasoning":0,"cache_read":0}},"reflection/beam-1.0-501b-a23b":{"id":"reflection/beam-1.0-501b-a23b","name":"Reflection: Beam 1.0 501B A23B","description":"Beam 1.0 is Reflection's open-weight reasoning model for coding and long-horizon agentic workflows. It uses a sparse mixture-of-experts architecture with 501B total and 23B active parameters per token, and a...","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":0.25,"output":1},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"stepfun/step-3.7-flash":{"id":"stepfun/step-3.7-flash","name":"Step 3.7 Flash","description":"Step 3.7 Flash is StepFun's latest high-efficiency multimodal Mixture-of-Experts model. It pairs a 196B-parameter language backbone with a vision encoder for native image and video understanding, activating roughly 11B parameters...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":230400},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"stepfun/step-5-preview":{"id":"stepfun/step-5-preview","name":"Step 5 Preview","description":"Step 5 Preview is StepFun's flagship model for agentic work, built on a sparse Mixture-of-Experts architecture (27B active / 600B total parameters). It performs strongly in software engineering and professional...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":64000},"cost":{"input":1,"output":2.7,"cache_read":0.1},"canonical_model_id":"stepfun/step-5-preview"},"stepfun/step-3.5-flash":{"id":"stepfun/step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"stepfun/step-3.5-flash"},"perceptron/perceptron-mk1":{"id":"perceptron/perceptron-mk1","name":"Perceptron: Perceptron Mk1","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":0.15,"output":1.5},"canonical_model_id":"perceptron/perceptron-mk1"},"perceptron/perceptron-mk1.5":{"id":"perceptron/perceptron-mk1.5","name":"Perceptron: Perceptron Mk1.5","description":"Perceptron Mk1.5 is Perceptron's embodied reasoning model for physical agents. It accepts text, image, video, and audio input, and answers with text plus optional structured annotations: points, boxes, polygons, tracks,...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":36864,"output":8192},"cost":{"input":0.15,"output":1.5}},"openrouter/pareto-code":{"id":"openrouter/pareto-code","name":"Pareto Code Router","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-05-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":65536},"cost":{"input":0,"output":0}},"openrouter/auto":{"id":"openrouter/auto","name":"Auto Router","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-15","last_updated":"2026-03-15","modalities":{"input":["audio","image","pdf","text","video"],"output":["image","text"]},"open_weights":false,"limit":{"context":2000000,"output":32768},"cost":{"input":0,"output":0}},"openrouter/bodybuilder":{"id":"openrouter/bodybuilder","name":"Body Builder (beta)","description":"Preview model for early access evaluation, prototyping, and compatibility testing","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-03-15","last_updated":"2026-03-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"status":"beta","cost":{"input":0,"output":0}},"openrouter/free":{"id":"openrouter/free","name":"OpenRouter Free Models Router","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-01","last_updated":"2026-02-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32768},"cost":{"input":0,"output":0}},"mistralai/mistral-small-3.1-24b-instruct":{"id":"mistralai/mistral-small-3.1-24b-instruct","name":"Mistral Small 3.1 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":102400},"cost":{"input":0.351,"output":0.555},"canonical_model_id":"mistral/mistral-small-3.1-24b-instruct"},"mistralai/mistral-medium-3":{"id":"mistralai/mistral-medium-3","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/mistral-medium-3"},"mistralai/mistral-nemo":{"id":"mistralai/mistral-nemo","name":"Mistral Nemo","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.15,"output":0.15,"cache_read":0.015},"canonical_model_id":"mistral/mistral-nemo"},"mistralai/mistral-large":{"id":"mistralai/mistral-large","name":"Mistral Large","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11-30","release_date":"2024-02-26","last_updated":"2024-02-26","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":102400},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mistral-large"},"mistralai/mistral-small-3.2-24b-instruct":{"id":"mistralai/mistral-small-3.2-24b-instruct","name":"Mistral Small 3.2 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.1,"output":0.3,"cache_read":0.01},"canonical_model_id":"mistral/mistral-small-3.2-24b-instruct"},"mistralai/mistral-saba":{"id":"mistralai/mistral-saba","name":"Saba","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-02-17","last_updated":"2025-02-17","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":26214},"cost":{"input":0.2,"output":0.6,"cache_read":0.02},"canonical_model_id":"mistral/mistral-saba"},"mistralai/ministral-8b-2512":{"id":"mistralai/ministral-8b-2512","name":"Ministral 3 8B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.15,"output":0.15,"cache_read":0.015},"canonical_model_id":"mistral/ministral-8b-2512"},"mistralai/devstral-2512":{"id":"mistralai/devstral-2512","name":"Devstral 2","description":"Devstral 2 is a state-of-the-art open-source model by Mistral AI specializing in agentic coding. It is a 123B-parameter dense transformer model supporting a 256K context window. Devstral 2 supports exploring...","family":"devstral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/devstral-2512"},"mistralai/mistral-medium-3.1":{"id":"mistralai/mistral-medium-3.1","name":"Mistral Medium 3.1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-30","release_date":"2025-08-12","last_updated":"2025-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/mistral-medium-3.1"},"mistralai/ministral-3b-2512":{"id":"mistralai/ministral-3b-2512","name":"Ministral 3 3B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":104857},"cost":{"input":0.1,"output":0.1,"cache_read":0.01},"canonical_model_id":"mistral/ministral-3b-2512"},"mistralai/mixtral-8x22b-instruct":{"id":"mistralai/mixtral-8x22b-instruct","name":"Mixtral 8x22B Instruct","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-01-31","release_date":"2024-04-17","last_updated":"2024-04-17","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":52428},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mixtral-8x22b-instruct"},"mistralai/mistral-medium-3-5":{"id":"mistralai/mistral-medium-3-5","name":"Mistral Medium 3.5","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":209715},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"mistral/mistral-medium-3-5"},"mistralai/mistral-large-2512":{"id":"mistralai/mistral-large-2512","name":"Mistral Large 3","description":"Mistral Large 3 2512 is Mistral’s most capable model to date, featuring a sparse mixture-of-experts architecture with 41B active parameters (675B total), and released under the Apache 2.0 license.","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"mistral/mistral-large-2512"},"mistralai/mistral-small-2603":{"id":"mistralai/mistral-small-2603","name":"Mistral Small 4","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"mistral/mistral-small-2603"},"mistralai/mistral-large-4-0":{"id":"mistralai/mistral-large-4-0","name":"Mistral: Mistral Large 4 (50% off)","description":"Mistral Large 4 is a frontier multimodal (text and image input) model from Mistral AI built for reasoning, coding, and agentic workloads. It offers a 512K-token context window with up...","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":262144},"cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"mistralai/voxtral-small-24b-2507":{"id":"mistralai/voxtral-small-24b-2507","name":"Voxtral Small 24B 2507","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":26214},"cost":{"input":0.1,"output":0.3,"cache_read":0.01},"canonical_model_id":"mistral/voxtral-small-24b-2507"},"mistralai/ministral-14b-2512":{"id":"mistralai/ministral-14b-2512","name":"Ministral 3 14B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.2,"output":0.2,"cache_read":0.02},"canonical_model_id":"mistral/ministral-14b-2512"},"mistralai/mistral-large-2407":{"id":"mistralai/mistral-large-2407","name":"Mistral Large 2407","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-03-31","release_date":"2024-07-24","last_updated":"2024-07-24","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mistral-large-2407"},"mistralai/codestral-2508":{"id":"mistralai/codestral-2508","name":"Codestral 25.08","description":"Mistral coding model for code completion, generation, and developer workflows","family":"codestral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":204800},"cost":{"input":0.3,"output":0.9,"cache_read":0.03},"canonical_model_id":"mistral/codestral-2508"},"mistralai/mistral-small-24b-instruct-2501":{"id":"mistralai/mistral-small-24b-instruct-2501","name":"Mistral Small 3","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-01-30","last_updated":"2025-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.05,"output":0.08},"canonical_model_id":"mistral/mistral-small-24b-instruct-2501"},"relace/relace-apply-3":{"id":"relace/relace-apply-3","name":"Relace: Relace Apply 3","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2025-09-26","last_updated":"2025-09-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.85,"output":1.25},"canonical_model_id":"relace/relace-apply-3"},"relace/relace-search":{"id":"relace/relace-search","name":"Relace: Relace Search","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":1,"output":3},"canonical_model_id":"relace/relace-search"},"minimax/minimax-m2-her":{"id":"minimax/minimax-m2-her","name":"MiniMax-M2 Her","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-01-23","last_updated":"2026-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":2048},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2-Her"},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax-M2","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-m1":{"id":"minimax/minimax-m1","name":"MiniMax: MiniMax M1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":40000},"cost":{"input":0.4,"output":2.2}},"minimax/minimax-m3.1-flash:free":{"id":"minimax/minimax-m3.1-flash:free","name":"MiniMax: MiniMax M3.1 Flash (free)","description":"MiniMax-M3.1-Flash is a multimodal model from MiniMax. Compared with M3, it delivers significant improvements across multiple benchmarks, particularly in complex agentic tasks, coding, and cybersecurity. It is built for complex,...","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0,"output":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax-M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-01":{"id":"minimax/minimax-01","name":"MiniMax: MiniMax-01","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-01-15","last_updated":"2025-01-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000192,"output":900172},"cost":{"input":0.2,"output":1.1}},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":32768},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"prism-ml/ternary-bonsai-2-27b":{"id":"prism-ml/ternary-bonsai-2-27b","name":"PrismML: Ternary Bonsai 2 27B","description":"Bonsai 2 27B is a 27B-parameter reasoning model from PrismML derived from Qwen3.8-27B. It supports coding, mathematics, tool calling, and image understanding with a 262K-token context window. Ternary compression shrinks...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.075,"output":0.5,"cache_read":0.0375}},"~google/gemini-pro-latest":{"id":"~google/gemini-pro-latest","name":"Google: Gemini Pro Latest","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["audio","pdf","image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":2,"output":10,"reasoning":10,"cache_read":0.1,"cache_write":0.1875}},"~google/gemini-flash-latest":{"id":"~google/gemini-flash-latest","name":"Google: Gemini Flash Latest","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667}},"~anthropic/claude-opus-latest":{"id":"~anthropic/claude-opus-latest","name":"Anthropic: Claude Opus Latest","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5}},"~anthropic/claude-haiku-latest":{"id":"~anthropic/claude-haiku-latest","name":"Anthropic: Claude Haiku Latest","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125}},"~anthropic/claude-fable-latest":{"id":"~anthropic/claude-fable-latest","name":"Anthropic: Claude Fable Latest ($$$$)","description":"This model always redirects to the latest model in the Claude Fable family.","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5}},"~anthropic/claude-sonnet-latest":{"id":"~anthropic/claude-sonnet-latest","name":"Anthropic: Claude Sonnet Latest","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5}},"elorian/glimpse":{"id":"elorian/glimpse","name":"Elorian: Glimpse","description":"Glimpse is a multimodal visual reasoning model from Elorian, built for visual understanding and complex visual reasoning in the physical world. It reasons directly over images and video rather than...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":256000},"cost":{"input":0.6,"output":3,"cache_read":0.06}},"liquid/lfm-2.5-2.6b:free":{"id":"liquid/lfm-2.5-2.6b:free","name":"LiquidAI: LFM2.5-2.6B (free)","description":"LFM2.5-2.6B is a compact reasoning model from Liquid AI. It is suited for agent workflows, data extraction, RAG, and long-context processing. Liquid advises against using it for agentic coding or...","family":"liquid","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":8192},"cost":{"input":0,"output":0}},"liquid/lfm-2.5-24b-a2b":{"id":"liquid/lfm-2.5-24b-a2b","name":"LiquidAI: LFM2.5-24B-A2B","description":"LFM2.5-24B-A2B is a mixture-of-experts reasoning model from Liquid AI with 24B total and 2.45B active parameters. It always reasons before answering and is built for agents that call tools. Liquid...","family":"liquid","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.05,"output":0.7,"cache_read":0.025}},"nousresearch/hermes-3-llama-3.1-405b":{"id":"nousresearch/hermes-3-llama-3.1-405b","name":"Nous: Hermes 3 405B Instruct (retires Nov 2)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-16","last_updated":"2024-08-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":1,"output":1},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-405b"},"nousresearch/hermes-4-405b":{"id":"nousresearch/hermes-4-405b","name":"Nous: Hermes 4 405B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"hermes","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":1,"output":3},"canonical_model_id":"nousresearch/hermes-4-405b"},"nousresearch/hermes-3-llama-3.1-70b":{"id":"nousresearch/hermes-3-llama-3.1-70b","name":"Nous: Hermes 3 70B Instruct (retires Nov 2)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-18","last_updated":"2024-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.7,"output":0.7},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-70b"},"~openai/gpt-luna-latest":{"id":"~openai/gpt-luna-latest","name":"OpenAI: GPT Luna Latest","description":"This model always redirects to the latest model in the OpenAI GPT Luna family.","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125}},"~openai/gpt-sol-latest":{"id":"~openai/gpt-sol-latest","name":"OpenAI: GPT Sol Latest","description":"This model always redirects to the latest model in the OpenAI GPT Sol family.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5}},"~openai/gpt-astra-latest":{"id":"~openai/gpt-astra-latest","name":"OpenAI: GPT Astra Latest ($$$$)","description":"This model always redirects to the latest model in the OpenAI GPT Astra family.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5}},"~openai/gpt-mini-latest":{"id":"~openai/gpt-mini-latest","name":"OpenAI: GPT Mini Latest","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075}},"~openai/gpt-terra-latest":{"id":"~openai/gpt-terra-latest","name":"OpenAI: GPT Terra Latest","description":"This model always redirects to the latest model in the OpenAI GPT Terra family.","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5}},"upstage/solar-pro-3":{"id":"upstage/solar-pro-3","name":"Upstage: Solar Pro 3","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07","release_date":"2026-01-26","last_updated":"2026-03-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":117964},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"upstage/solar-pro3"},"upstage/solar-pro4":{"id":"upstage/solar-pro4","name":"Upstage: Solar Pro 4","description":"Solar Pro 4 is a large language model from Upstage. It is suited for agentic workflows, office productivity, document-intensive work, and coding.","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"upstage/solar-pro4"},"upstage/solar-mini4":{"id":"upstage/solar-mini4","name":"Upstage: Solar Mini 4","description":"Solar Mini 4 is Upstage's compact, cost-efficient language model, a 35B-parameter mixture-of-experts with 3B active parameters and a 524K context window. It is built for agentic use cases where response...","family":"solar-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"upstage/solar-mini4"},"ibm-granite/granite-4.0-h-micro":{"id":"ibm-granite/granite-4.0-h-micro","name":"IBM: Granite 4.0 Micro","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"granite","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-10-20","last_updated":"2025-10-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":117900},"cost":{"input":0.017,"output":0.112}},"ibm-granite/granite-4.2-8b":{"id":"ibm-granite/granite-4.2-8b","name":"IBM: Granite 4.2 8B","description":"Granite 4.2 8B is a dense reasoning model from IBM. It is suited for mathematics, code generation, multilingual dialogue, and agentic workflows that need multi-step reasoning. It supports full, low-effort,...","family":"granite","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-25","last_updated":"2026-08-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.06,"output":0.25,"cache_read":0.015},"canonical_model_id":"ibm/granite-4.2-8b"},"cohere/command-a":{"id":"cohere/command-a","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a"},"cohere/command-r-08-2024":{"id":"cohere/command-r-08-2024","name":"Command R","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"cohere/command-r-08-2024"},"cohere/command-a-plus":{"id":"cohere/command-a-plus","name":"Cohere: Command A+","description":"Command A+ is Cohere's flagship model for enterprise agentic workflows. It accepts text and image inputs with a 192K context window, supports native tool calling with strict tool schemas, structured...","family":"command-a","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04-01","release_date":"2026-05-20","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":192000,"output":64000},"cost":{"input":0.3,"output":1.5,"cache_read":0.15},"canonical_model_id":"cohere/command-a-plus-05-2026"},"cohere/command-r7b-12-2024":{"id":"cohere/command-r7b-12-2024","name":"Command R7B","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-r","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-12-13","last_updated":"2024-12-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.0375,"output":0.15},"canonical_model_id":"cohere/command-r7b-12-2024"},"cohere/command-r-plus-08-2024":{"id":"cohere/command-r-plus-08-2024","name":"Command R+","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-r-plus-08-2024"},"cohere/north-mini-code:free":{"id":"cohere/north-mini-code:free","name":"Cohere: North Mini Code (free)","description":"North Mini Code is Cohere's first agentic coding model and the debut of its North family. A sparse mixture-of-experts model with 30B total parameters and 3B active, it is optimized...","family":"north","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09-23","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0,"output":0},"canonical_model_id":"cohere/north-mini-code-1-0"},"perplexity/sonar-deep-research":{"id":"perplexity/sonar-deep-research","name":"Perplexity: Sonar Deep Research","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar-deep-research","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-07","last_updated":"2025-03-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":115200},"cost":{"input":2,"output":8,"reasoning":3}},"perplexity/sonar-pro":{"id":"perplexity/sonar-pro","name":"Perplexity: Sonar Pro","description":"Advanced Sonar search model for deeper research and cited synthesis","family":"sonar-pro","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-07","last_updated":"2025-03-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8000},"cost":{"input":3,"output":15}},"perplexity/sonar":{"id":"perplexity/sonar","name":"Perplexity: Sonar","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-01-27","last_updated":"2025-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":127072,"output":114364},"cost":{"input":1,"output":1}},"perplexity/sonar-reasoning-pro":{"id":"perplexity/sonar-reasoning-pro","name":"Perplexity: Sonar Reasoning Pro","description":"Web-grounded reasoning model for multi-step research and cited answers","family":"sonar-reasoning","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-07","last_updated":"2025-03-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":115200},"cost":{"input":2,"output":8}},"perplexity/sonar-pro-search":{"id":"perplexity/sonar-pro-search","name":"Perplexity: Sonar Pro Search","description":"Advanced Sonar search model for deeper research and cited synthesis","family":"sonar-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-10-30","last_updated":"2025-10-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8000},"cost":{"input":3,"output":15}},"xiaomi/mimo-v2.5-pro":{"id":"xiaomi/mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.435,"output":0.87,"cache_read":0.004,"tiers":[{"input":2,"output":6,"cache_read":0.4,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.4}},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.6-pro-ultraspeed":{"id":"xiaomi/mimo-v2.6-pro-ultraspeed","name":"MiMo-V2.6-Pro-UltraSpeed","description":"MiMo-V2.6-Pro-UltraSpeed is the fast speed edition of Xiaomi's flagship foundation model, MiMo-V2.6-Pro. Built from the same 1T MiMo-V2.6-Pro checkpoint, it matches the original model in quality while delivering roughly 10x...","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.35,"output":8.7,"cache_read":0.036},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.003,"tiers":[{"input":0.8,"output":4,"cache_read":0.16,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.8,"output":4,"cache_read":0.16}},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2.6-flash":{"id":"xiaomi/mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo-V2.6-Flash is an open-source foundation model developed by Xiaomi. Built on a Mixture-of-Experts architecture with 309B total parameters and 15B activated per token, it employs a hybrid attention mechanism for...","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.03,"output":0.08,"cache_read":0.003},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"MiMo-V2.6-Pro is the flagship foundation model developed by Xiaomi. Built at a scale of over 1T parameters, it is designed to push the ceiling of capability for the most demanding...","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.43,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"unbiased/pareto-26.9":{"id":"unbiased/pareto-26.9","name":"Pareto 26.9","description":"Pareto is a multimodal composite model built for research, coding, and agentic workflows, while delivering frontier-level performance across a broad range of general-purpose tasks.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.25}},"unbiased/pareto-26.10":{"id":"unbiased/pareto-26.10","name":"Pareto 26.10","description":"Pareto is a multimodal composite model built for research, coding, and agentic workflows, while delivering frontier-level performance across a broad range of general-purpose tasks.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.8,"output":3.2,"cache_read":0.03}},"moondream/moondream3-preview":{"id":"moondream/moondream3-preview","name":"Moondream: Moondream 3 Preview","description":"Moondream 3 Preview is a vision-language model from Moondream with 9B total parameters and 2B active (mixture-of-experts). It handles visual question answering, captioning, OCR, and visual reasoning over images, with...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":4096},"cost":{"input":0.3,"output":2.5}},"moondream/moondream3.1-9b-a2b":{"id":"moondream/moondream3.1-9b-a2b","name":"Moondream: Moondream 3.1","description":"Moondream 3.1 is a vision-language model from Moondream, the successor to Moondream 3 Preview, with 9B total parameters and 2B active (mixture-of-experts). It handles visual question answering, captioning, OCR, and...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":4096},"cost":{"input":0.3,"output":2.5}},"x-ai/grok-4.20-multi-agent":{"id":"x-ai/grok-4.20-multi-agent","name":"SpaceXAI: Grok 4.20 Multi-Agent","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":1800000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2}},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":900000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"x-ai/grok-4.7":{"id":"x-ai/grok-4.7","name":"Grok 4.7","description":"Grok 4.7 is SpaceXAI's smartest model with frontier performance on coding, knowledge work, and STEM.","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"x-ai/grok-build-0.1":{"id":"x-ai/grok-build-0.1","name":"Grok Build 0.1","description":"Grok coding model for agentic engineering, edits, and codebase workflows","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":230400},"cost":{"input":1,"output":2,"cache_read":0.2},"canonical_model_id":"xai/grok-build-0.1"},"x-ai/grok-4.5":{"id":"x-ai/grok-4.5","name":"Grok 4.5","description":"Grok 4.5 is SpaceXAI's smartest model with frontier performance on coding, knowledge work, and STEM.","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.3},"canonical_model_id":"xai/grok-4.5"},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"Grok 4.6 is SpaceXAI's smartest model with frontier performance on coding, knowledge work, and STEM.","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"x-ai/grok-4.20":{"id":"x-ai/grok-4.20","name":"SpaceXAI: Grok 4.20","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":1800000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2}},"google/gemini-3.1-flash-image":{"id":"google/gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Gemini 3.1 Flash Image, a.k.a. \"Nano Banana 2,\" is Google’s latest state of the art image generation and editing model, delivering Pro-level visual quality at Flash speed. It combines advanced...","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["image","text","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image"},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.04,"output":0.12},"canonical_model_id":"google/gemma-3-12b-it"},"google/gemini-3-pro-image-preview":{"id":"google/gemini-3-pro-image-preview","type":"image","name":"Nano Banana Pro Preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["image","text","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":1,"output":6,"reasoning":6,"cache_read":0.1,"cache_write":0.1875},"canonical_model_id":"google/gemini-3-pro-image-preview"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.041667},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-3.1-flash-image-preview":{"id":"google/gemini-3.1-flash-image-preview","type":"image","name":"Nano Banana 2 Preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":58982},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image-preview"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.08,"output":0.16},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"status":"deprecated","cost":{"input":0.1,"output":0.4,"reasoning":0.4,"cache_read":0.01,"cache_write":0.083333},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":1.25,"output":10,"reasoning":10,"cache_read":0.125,"cache_write":0.375},"canonical_model_id":"google/gemini-2.5-pro"},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.125,"output":0.75,"reasoning":0.75,"cache_read":0.0125,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"Gemini 3.7 Flash is a multimodal model from Google for fast agentic workflows, coding, and complex multi-step reasoning. It is designed for tasks that require responsive performance and reliable multi-step...","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemma-2-27b-it":{"id":"google/gemma-2-27b-it","name":"Gemma 2 27B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-07-13","last_updated":"2024-07-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":2048},"cost":{"input":0.65,"output":0.65},"canonical_model_id":"google/gemma-2-27b-it"},"google/lyria-3-clip-preview":{"id":"google/lyria-3-clip-preview","name":"Lyria 3 Clip Preview","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"lyria","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-25","last_updated":"2026-03-25","modalities":{"input":["text","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/lyria-3-clip-preview"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1,"output":6,"reasoning":6,"cache_read":0.1,"cache_write":0.1875},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Gemini 3.8 Flash is Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows.","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-2.5-pro-preview":{"id":"google/gemini-2.5-pro-preview","name":"Gemini 2.5 Pro Preview 06-05","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01-31","release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["pdf","image","text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"reasoning":10,"cache_read":0.125,"cache_write":0.375},"canonical_model_id":"google/gemini-2.5-pro-preview"},"google/gemini-3-pro-image":{"id":"google/gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro is Google’s most advanced image-generation and editing model, built on Gemini 3 Pro. It extends the original Nano Banana with significantly improved multimodal reasoning, real-world grounding, and...","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["image","text","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3-pro-image"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":4.5,"reasoning":4.5,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-3.1-flash-lite-image":{"id":"google/gemini-3.1-flash-lite-image","type":"image","name":"Nano Banana 2 Lite","description":"Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) is Google's fastest, most cost-efficient Gemini image model, built for high-velocity developer pipelines and rapid-fire visual exploration. It delivers text-to-image generation...","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["image","text","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":58982},"cost":{"input":0.25,"output":1.5},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"google/gemini-nano-banana-2.1":{"id":"google/gemini-nano-banana-2.1","type":"image","name":"Nano Banana 2.1","description":"Nano Banana 2.1 (Gemini Nano Banana 2.1) is Google's image generation and editing model on the Flash tier, succeeding Nano Banana 2 and Nano Banana Pro. It improves product recontextualization,...","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["image","text","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":58982},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"google/gemini-nano-banana-2.1"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Gemini 4 Argon is Google's flagship reasoning model, built for complex software engineering, deep mathematical and scientific problem solving, and autonomous multi-step agents. Google trained it to keep context across...","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":2,"output":10,"reasoning":10,"cache_read":0.1,"cache_write":0.1875},"canonical_model_id":"google/gemini-4-argon"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.042,"output":0.22,"cache_read":0.021},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemma-3-4b-it":{"id":"google/gemma-3-4b-it","name":"Gemma 3 4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.04,"output":0.05,"cache_read":0.02},"canonical_model_id":"google/gemma-3-4b-it"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Gemini 3.5 Flash Lite is a high-efficiency model from Google with upgraded agentic capabilities. It is suited for subagents that execute focused tasks within complex, multi-agent workflows.","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.15,"output":1.25,"reasoning":1.25,"cache_read":0.015,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.1,"output":0.34,"cache_read":0.1},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"status":"deprecated","cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Gemini 3.6 Flash is a high-efficiency model from Google for coding, agentic workflows, and web and app development. It is designed to produce polished outputs with fewer unnecessary edits and...","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.375,"output":1.875,"reasoning":1.875,"cache_read":0.0375,"cache_write":0.020833},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.1-flash-lite-preview":{"id":"google/gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.125,"output":0.75,"reasoning":0.75,"cache_read":0.0125,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"google/gemini-2.5-flash-image":{"id":"google/gemini-2.5-flash-image","type":"image","name":"Nano Banana","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["image","text","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.083333},"canonical_model_id":"google/gemini-2.5-flash-image"},"google/lyria-3-pro-preview":{"id":"google/lyria-3-pro-preview","name":"Lyria 3 Pro Preview","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"lyria","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-25","last_updated":"2026-03-25","modalities":{"input":["text","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/lyria-3-pro-preview"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 preview","description":"Tencent: Hy4 preview is a mixture-of-experts model from Tencent, with 49B active parameters out of 770B total. It is designed for coding agents, complex tool-use workflows, and productivity tasks that...","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy-mt2-1.8b":{"id":"tencent/hy-mt2-1.8b","name":"Hy-MT2-1.8B","description":"Hy-MT2-1.8B is a compact 1.8B-parameter translation model from Tencent. It supports 33 language pairs and five Chinese dialect and minority-language pairs, with workflows for structured, delimiter-based, contextual, glossary-based, and style-guided...","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.044,"output":0.177},"canonical_model_id":"tencent/hy-mt2-1.8b"},"tencent/hy-mt2-7b":{"id":"tencent/hy-mt2-7b","name":"Hy-MT2-7B","description":"Hy-MT2-7B is a 7B-parameter translation model from Tencent. It supports 33 language pairs and five Chinese dialect and minority-language pairs, with workflows for structured, delimiter-based, contextual, glossary-based, and style-guided translation.","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.074,"output":0.295},"canonical_model_id":"tencent/hy-mt2-7b"},"tencent/hy-mt2-30b-a3b":{"id":"tencent/hy-mt2-30b-a3b","name":"Hy-MT2-30B-A3B","description":"Hy-MT2-30B-A3B is Tencent's flagship translation model in the Hy-MT2 family. It supports 33 language pairs and five Chinese dialect and minority-language pairs, with workflows for structured, delimiter-based, contextual, glossary-based, and...","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.074,"output":0.295},"canonical_model_id":"tencent/hy-mt2-30b-a3b"},"tencent/hy3-preview":{"id":"tencent/hy3-preview","name":"Hy3 preview","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.18,"output":0.6,"cache_read":0.06},"canonical_model_id":"tencent/hy3-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3","description":"Hy3 is a 295B-parameter Mixture-of-Experts model from Tencent (21B active, 192 experts with top-8 routing) built for reasoning, agentic workflows, and real-world production use. It supports a configurable reasoning effort:...","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":128000},"cost":{"input":0.132,"output":0.528,"cache_read":0.033},"canonical_model_id":"tencent/hy3"},"tencent/hunyuan-a13b-instruct":{"id":"tencent/hunyuan-a13b-instruct","name":"Tencent: Hunyuan A13B Instruct","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-08","last_updated":"2025-07-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":117964},"cost":{"input":0.14,"output":0.57}},"sakana/fugu-max":{"id":"sakana/fugu-max","name":"Fugu Max","description":"Fugu Max is the cost-performance model in Sakana AI's Fugu family. Rather than a single monolithic model, Fugu is a learned multi-agent orchestration system: a language model trained to route...","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"sakana/fugu-max"},"sakana/sakana-namazu":{"id":"sakana/sakana-namazu","name":"Sakana Namazu","description":"Sakana Namazu is a Japanese-specialized reasoning model from Sakana AI, based on Kimi K2.6 with additional training for Japanese language and business contexts. It is suited for Japanese instruction following,...","family":"sakana-namazu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.95,"output":4,"cache_read":0.15},"canonical_model_id":"sakana/sakana-namazu"},"sakana/fugu-ultra-v2":{"id":"sakana/fugu-ultra-v2","name":"Fugu Ultra v2","description":"Fugu Ultra v2 is the higher-performance model in Sakana AI's Fugu family. Rather than a single monolithic model, Fugu is a learned multi-agent orchestration system: a language model trained to...","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-08-28","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"sakana/fugu-ultra-v2"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Fugu Ultra is the higher-performance model in Sakana AI's Fugu family. Rather than a single monolithic model, Fugu is a learned multi-agent orchestration system: a language model trained to route...","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"sakana/fugu-ultra"},"cognitivecomputations/dolphin-mistral-24b-venice-edition":{"id":"cognitivecomputations/dolphin-mistral-24b-venice-edition","name":"Venice: Uncensored","description":"Venice Uncensored Dolphin Mistral 24B Venice Edition is a fine-tuned variant of Mistral-Small-24B-Instruct-2501, developed by dphn.ai in collaboration with Venice.ai. This model is designed as an “uncensored” instruct-tuned LLM, preserving...","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04-30","release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.2,"output":0.9},"canonical_model_id":"cognitivecomputations/dolphin-mistral-24b-venice-edition"},"rekaai/reka-edge":{"id":"rekaai/reka-edge","name":"Reka Edge","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"reka","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":14745},"cost":{"input":0.1,"output":0.1},"canonical_model_id":"rekaai/reka-edge"},"rekaai/reka-flash-3":{"id":"rekaai/reka-flash-3","name":"Reka Flash 3","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"reka","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01-31","release_date":"2025-03-10","last_updated":"2025-03-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":58982},"cost":{"input":0.1,"output":0.2},"canonical_model_id":"rekaai/reka-flash-3"},"openai/o3-mini-high":{"id":"openai/o3-mini-high","name":"OpenAI: o3 Mini High","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-02-12","last_updated":"2025-02-12","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55}},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4":{"id":"openai/gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":8191,"output":4096},"cost":{"input":30,"output":60},"canonical_model_id":"openai/gpt-4"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-3.5-turbo-instruct":{"id":"openai/gpt-3.5-turbo-instruct","name":"OpenAI: GPT-3.5 Turbo Instruct","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2023-09-28","last_updated":"2023-09-28","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":4095,"output":3685},"cost":{"input":1.5,"output":2}},"openai/gpt-5.6-luna-pro":{"id":"openai/gpt-5.6-luna-pro","name":"GPT-5.6 Luna","description":"GPT-5.6 Luna Pro is the same underlying model as [GPT-5.6 Luna](https://openrouter.ai/openai/gpt-5.6-luna), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-audio":{"id":"openai/gpt-audio","type":"chat","name":"GPT-Audio","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-01","release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text","audio","pdf"],"output":["text","audio"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10},"canonical_model_id":"openai/gpt-audio"},"openai/gpt-5.2-chat":{"id":"openai/gpt-5.2-chat","name":"OpenAI: GPT-5.2 Chat","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-chat-latest"},"openai/gpt-6-luna-pro":{"id":"openai/gpt-6-luna-pro","name":"OpenAI: GPT-6 Luna Pro","description":"GPT-6 Luna Pro is the same underlying model as [GPT-6 Luna](https://openrouter.ai/openai/gpt-6-luna), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125}},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"GPT-6 Luna is the fast, cost-efficient model in OpenAI's GPT-6 series, positioned below GPT-6 Sol. It is suited for high-volume and latency-sensitive workloads such as chat, classification, and lightweight agentic...","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/o3-pro":{"id":"openai/o3-pro","name":"o3-pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","pdf","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":20,"output":80},"canonical_model_id":"openai/o3-pro"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"GPT-5.6 Terra is a balanced model in OpenAI's GPT-5.6 series, positioned between the flagship Sol tier and the cost-efficient Luna tier. It is suited for everyday coding, reasoning, and agentic...","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5-turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-chat-latest":{"id":"openai/gpt-chat-latest","name":"OpenAI: GPT Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-05-05","last_updated":"2026-05-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5}},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-4o-2024-08-06":{"id":"openai/gpt-4o-2024-08-06","name":"GPT-4o (2024-08-06)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-08-06"},"openai/gpt-3.5-turbo-0613":{"id":"openai/gpt-3.5-turbo-0613","name":"OpenAI: GPT-3.5 Turbo (older v0613)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":4095,"output":3685},"cost":{"input":1,"output":2}},"openai/gpt-6-astra-pro":{"id":"openai/gpt-6-astra-pro","name":"OpenAI: GPT-6 Astra Pro ($$$$)","description":"GPT-6 Astra Pro is the same underlying model as [GPT-6 Astra](https://openrouter.ai/openai/gpt-6-astra), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5}},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"GPT-5.6 Luna is a fast, cost-efficient model in OpenAI's GPT-5.6 series. It is suited for high-volume, latency-sensitive tasks such as chat, classification, and lightweight agentic workflows, providing capable reasoning for...","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-6-sol-pro":{"id":"openai/gpt-6-sol-pro","name":"OpenAI: GPT-6 Sol Pro","description":"GPT-6 Sol Pro is the same underlying model as [GPT-6 Sol](https://openrouter.ai/openai/gpt-6-sol), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5}},"openai/gpt-oss-safeguard-20b":{"id":"openai/gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.018,"output":0.09,"cache_read":0.009},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/o3":{"id":"openai/o3","name":"o3","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/gpt-3.5-turbo-16k":{"id":"openai/gpt-3.5-turbo-16k","name":"OpenAI: GPT-3.5 Turbo 16k","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2023-08-28","last_updated":"2023-08-28","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":3,"output":4}},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.03},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/gpt-5.6-terra-pro":{"id":"openai/gpt-5.6-terra-pro","name":"GPT-5.6 Terra","description":"GPT-5.6 Terra Pro is the same underlying model as [GPT-5.6 Terra](https://openrouter.ai/openai/gpt-5.6-terra), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5.1-codex"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's flagship model for demanding end-to-end work. It is suited for advanced analysis, software engineering, deep research, scientific work, and document creation, with particular strengths in long-horizon...","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5-image":{"id":"openai/gpt-5-image","type":"image","name":"GPT-5 Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-10-01","release_date":"2025-10-14","last_updated":"2025-10-14","modalities":{"input":["image","text","pdf"],"output":["image","text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":10,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-5-image"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/gpt-4o-2024-05-13":{"id":"openai/gpt-4o-2024-05-13","name":"GPT-4o (2024-05-13)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-05-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":5,"output":15},"canonical_model_id":"openai/gpt-4o-2024-05-13"},"openai/o4-mini-high":{"id":"openai/o4-mini-high","name":"OpenAI: o4 Mini High","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275}},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"GPT-5.6 Sol is the flagship model in OpenAI's GPT-5.6 series. It is suited for complex reasoning, coding, and agentic workflows, and is particularly strong at command-line and multi-step coding tasks...","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.6-sol-pro":{"id":"openai/gpt-5.6-sol-pro","name":"GPT-5.6 Sol","description":"GPT-5.6 Sol Pro is the same underlying model as [GPT-5.6 Sol](https://openrouter.ai/openai/gpt-5.6-sol), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"GPT-6.1 Sol is an upgrade to GPT-6 Sol from OpenAI, positioned below the flagship GPT-6 Astra in the GPT-6 series. It is suited for agentic coding, computer use, document-heavy professional...","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5-image-mini":{"id":"openai/gpt-5-image-mini","type":"image","name":"GPT-5 Image Mini","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-10-16","last_updated":"2025-10-16","modalities":{"input":["pdf","image","text"],"output":["image","text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":2.5,"output":2,"cache_read":0.25},"canonical_model_id":"openai/gpt-5-image-mini"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.03,"output":0.15,"cache_read":0.015},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"GPT-6 Sol is the cost-efficient high-end model in OpenAI's GPT-6 series, positioned below the flagship GPT-6 Astra and above the fast GPT-6 Luna tier. It is suited for demanding professional...","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1-codex-max":{"id":"openai/gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex-max"},"openai/gpt-4o-mini-2024-07-18":{"id":"openai/gpt-4o-mini-2024-07-18","name":"OpenAI: GPT-4o-mini (2024-07-18)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/o1-pro":{"id":"openai/o1-pro","name":"o1-pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2025-03-19","last_updated":"2025-03-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":150,"output":600},"canonical_model_id":"openai/o1-pro"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-6.1-sol-pro":{"id":"openai/gpt-6.1-sol-pro","name":"OpenAI: GPT-6.1 Sol Pro","description":"GPT-6.1 Sol Pro is the same underlying model as [GPT-6.1 Sol](https://openrouter.ai/openai/gpt-6.1-sol), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. **Cost note:** pro mode spends far more...","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5}},"openai/gpt-4o-2024-11-20":{"id":"openai/gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"openai/gpt-5.4-image-2":{"id":"openai/gpt-5.4-image-2","type":"image","name":"GPT-5.4 Image 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":false,"structured_output":true,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["image","text","pdf"],"output":["image","text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":8,"output":15,"cache_read":2},"canonical_model_id":"openai/gpt-5.4-image-2"},"openai/gpt-audio-mini":{"id":"openai/gpt-audio-mini","type":"chat","name":"GPT-Audio Mini","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-01","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","audio","pdf"],"output":["text","audio"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.6,"output":2.4},"canonical_model_id":"openai/gpt-audio-mini"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"morph/morph-v3-fast":{"id":"morph/morph-v3-fast","name":"Morph: Morph V3 Fast","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-07","last_updated":"2025-07-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.8,"output":1.2},"canonical_model_id":"morph/morph-v3-fast"},"morph/morph-v3-large":{"id":"morph/morph-v3-large","name":"Morph: Morph V3 Large","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-07","last_updated":"2025-07-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.9,"output":1.9},"canonical_model_id":"morph/morph-v3-large"},"stealth/claude-opus-4.8":{"id":"stealth/claude-opus-4.8","name":"Stealth: Claude Opus 4.8 (20% off)","description":"Your prompts and completions may be retained and used to train or improve the provider's services. This third-party-served variant of Claude Opus 4.8 is offered at 20% lower cost than standard Claude Opus 4.8 pricing and is not served by Anthropic or Kilo Code.","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"reasoning":0,"cache_read":0.4,"cache_write":5}},"stealth/claude-opus-4.7":{"id":"stealth/claude-opus-4.7","name":"Stealth: Claude Opus 4.7 (20% off)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"reasoning":0,"cache_read":0.4,"cache_write":5}},"stealth/qwen3.6-plus":{"id":"stealth/qwen3.6-plus","name":"Stealth: Qwen3.6 Plus (50% off)","description":"Your prompts and completions may be retained and used to train or improve the provider's services. This third-party-served variant of Qwen3.6 Plus is offered at 50% lower cost than standard Qwen3.6 Plus pricing and is not served by Alibaba or Kilo Code. Note: a surcharge applies to long-context workloads exceeding 256K input tokens.","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":0,"cache_read":0.025,"cache_write":0.3125}},"stealth/claude-opus-4.6":{"id":"stealth/claude-opus-4.6","name":"Stealth: Claude Opus 4.6 (20% off)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"reasoning":0,"cache_read":0.4,"cache_write":5}},"stealth/glyph-cluster":{"id":"stealth/glyph-cluster","name":"Stealth: Glyph Cluster (free)","description":"Your prompts and completions may be retained and used to train or improve the provider's services. Glyph Cluster is a reasoning model for coding and long-context analysis, served through Vercel AI Gateway during its stealth period at no cost. It supports function calling and streaming, takes text input only, and does not support structured outputs. Zero Data Retention is not available for this model.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0,"output":0,"reasoning":0,"cache_read":0}},"stealth/claude-sonnet-4.6":{"id":"stealth/claude-sonnet-4.6","name":"Stealth: Claude Sonnet 4.6 (20% off)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2.4,"output":12,"reasoning":0,"cache_read":0.24,"cache_write":3}},"arcee-ai/trinity-large-thinking":{"id":"arcee-ai/trinity-large-thinking","name":"Trinity Large Thinking","description":"Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use","family":"trinity","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":80000},"cost":{"input":0.25,"output":0.8,"cache_read":0.06},"canonical_model_id":"arcee-ai/trinity-large-thinking"}}},"baseten":{"id":"baseten","env":["BASETEN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.baseten.co/v1","name":"Baseten","doc":"https://docs.baseten.co/inference/model-apis/overview","models":{"nvidia/Nemotron-120B-A12B":{"id":"nvidia/Nemotron-120B-A12B","name":"Nemotron Super","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":202800},"cost":{"input":0.3,"output":0.75,"cache_read":0.06},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B":{"id":"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B","name":"Nemotron Ultra","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":202800},"cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek V3.1","description":"Legacy model retained for compatibility with older integrations","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-08-25","last_updated":"2025-08-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":131000},"status":"deprecated","cost":{"input":0.5,"output":1.5}},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.13,"output":0.26,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-V4.1-Flash":{"id":"deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-ai/DeepSeek-V4.1-Flash-1017":{"id":"deepseek-ai/DeepSeek-V4.1-Flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-ai/DeepSeek-V4.1-Flash-1017-Fast":{"id":"deepseek-ai/DeepSeek-V4.1-Flash-1017-Fast","name":"DeepSeek V4.1 Flash 1017 Fast","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0.6,"output":2.4},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-ai/DeepSeek-V4.1-Flash-Fast":{"id":"deepseek-ai/DeepSeek-V4.1-Flash-Fast","name":"DeepSeek V4.1 Flash Fast","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.6,"output":2.4},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":1.74,"output":3.48,"cache_read":0.145},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":1.32,"output":3.96},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.95,"output":4,"cache_read":0.16}},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":3,"output":15},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.5":{"id":"moonshotai/Kimi-K2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2026-01-30","last_updated":"2026-02-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.6,"output":3,"cache_read":0.12}},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax-M2.5","description":"Legacy model retained for compatibility with older integrations","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2026-01","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204000,"output":204000},"status":"deprecated","cost":{"input":0.3,"output":1.2}},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":1,"output":4.05},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.5,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"zai-org/GLM-4.7":{"id":"zai-org/GLM-4.7","name":"GLM 4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":200000},"cost":{"input":0.6,"output":2.2,"cache_read":0.12}},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"GLM 5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":202800},"cost":{"input":1.3,"output":4.3,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM 5.3","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5":{"id":"zai-org/GLM-5","name":"GLM 5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":202800},"cost":{"input":0.95,"output":3.15,"cache_read":0.2}},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":1.4,"output":4.4,"cache_read":0.3},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.2-Fast":{"id":"zai-org/GLM-5.2-Fast","name":"GLM 5.2 Fast","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":2.1,"output":6.6,"cache_read":0.21},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.3-Fast":{"id":"zai-org/GLM-5.3-Fast","name":"GLM 5.3 Fast","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":2.1,"output":6.6},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM 5.3 Flash","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5},"canonical_model_id":"zhipuai/glm-5.3-flash"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"OpenAI GPT 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128072,"output":128072},"cost":{"input":0.1,"output":0.5}}}},"salad-cloud":{"id":"salad-cloud","env":["SALAD_CLOUD_API_KEY"],"npm":"@saladtechnologies-oss/ai-sdk-provider","name":"SaladCloud AI Gateway","doc":"https://docs.salad.com/ai-gateway/explanation/overview","models":{"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Qwen MoE for agentic tasks, complex reasoning, code generation, and instruction following","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":262144},"cost":{"input":0.09,"output":0.6},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"}}},"moark":{"id":"moark","env":["MOARK_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://moark.com/v1","name":"Moark","doc":"https://moark.com/docs/openapi/v1#tag/%E6%96%87%E6%9C%AC%E7%94%9F%E6%88%90","models":{"GLM-4.7":{"id":"GLM-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.42,"output":2.2}},"Kimi-K2.6":{"id":"Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.98,"output":4},"canonical_model_id":"moonshotai/kimi-k2.6"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.3},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15},"canonical_model_id":"moonshotai/kimi-k3"},"GLM-5.3":{"id":"GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.2,"output":4.18},"canonical_model_id":"zhipuai/glm-5.3"},"Qwen3.6-Flash":{"id":"Qwen3.6-Flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.18,"output":1.03},"canonical_model_id":"alibaba/qwen3.6-flash"},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.29,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.7"},"DeepSeek-V4-Flash":{"id":"DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.5,"output":0.5},"canonical_model_id":"deepseek/deepseek-v4-flash"},"DeepSeek-V4-Pro":{"id":"DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.8,"output":3.6},"canonical_model_id":"deepseek/deepseek-v4-pro"},"MiniMax-M2.1":{"id":"MiniMax-M2.1","name":"MiniMax-M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.29,"output":1.2}}}},"pendra":{"id":"pendra","env":["PENDRA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.pendra.ai/api/v1","name":"Pendra","doc":"https://pendra.ai/docs/integrations/opencode","models":{"gpt-oss:120b":{"id":"gpt-oss:120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-oss-120b"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"glm-4.7-flash":{"id":"glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"llama3.3:70b":{"id":"llama3.3:70b","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"qwen3.6:27b":{"id":"qwen3.6:27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.6-27b"},"qwen3-coder:30b":{"id":"qwen3-coder:30b","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"}}},"agnes":{"id":"agnes","env":["AGNES_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://apihub.agnes-ai.com/v1","name":"Agnes AI","doc":"https://agnes-ai.com/doc","models":{"agnes-2.5-pro-alpha":{"id":"agnes-2.5-pro-alpha","name":"Agnes 2.5 Pro Alpha","description":"Paid reasoning model for advanced coding, scientific reasoning, long-context analysis, agentic workflows, and multimodal understanding.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0.45,"output":0.9,"cache_read":0.0038}},"agnes-2.5-flash":{"id":"agnes-2.5-flash","name":"Agnes 2.5 Flash","description":"Upgraded model with improved coding, agent workflows, tool calling, and multimodal understanding.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07","last_updated":"2026-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":512000,"output":65536},"cost":{"input":0,"output":0}},"agnes-3.0-flash":{"id":"agnes-3.0-flash","name":"Agnes 3.0 Flash","description":"A new-generation Agnes AI text model with stronger Agnes Code task execution, tool orchestration, and trustworthy delivery.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09","last_updated":"2026-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":512000,"output":65536},"cost":{"input":0,"output":0}},"agnes-2.0-flash":{"id":"agnes-2.0-flash","name":"Agnes 2.0 Flash","description":"Fast and efficient model for agent workflows, tool calling, coding, and image understanding.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-25","last_updated":"2026-05-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":512000,"output":65536},"status":"deprecated","cost":{"input":0,"output":0}}}},"kosmik":{"id":"kosmik","env":["KOSMIK_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.koscompute.com/v1","name":"Kosmik Compute","doc":"https://api.koscompute.com/docs/","models":{"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.35,"output":2.2,"cache_read":0.09},"canonical_model_id":"alibaba/qwen3.8-27b"}}},"sap-ai-core":{"id":"sap-ai-core","env":["AICORE_SERVICE_KEY"],"npm":"@jerome-benoit/sap-ai-provider-v2","name":"SAP AI Core","doc":"https://help.sap.com/docs/sap-ai-core","models":{"anthropic--claude-4-sonnet":{"id":"anthropic--claude-4-sonnet","name":"anthropic--claude-4-sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"sonar-deep-research":{"id":"sonar-deep-research","name":"sonar-deep-research","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar-deep-research","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-02-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":2,"output":8,"reasoning":3}},"anthropic--claude-4.5-haiku":{"id":"anthropic--claude-4.5-haiku","name":"anthropic--claude-4.5-haiku","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic--claude-4.6-sonnet":{"id":"anthropic--claude-4.6-sonnet","name":"anthropic--claude-4.6-sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-08","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gpt-5.5":{"id":"gpt-5.5","name":"gpt-5.5","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"gpt-4.1-nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.08,"output":0.26},"canonical_model_id":"openai/gpt-4.1-nano"},"anthropic--claude-4.5-sonnet":{"id":"anthropic--claude-4.5-sonnet","name":"anthropic--claude-4.5-sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"gpt-4.1-mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"gpt-5.2":{"id":"gpt-5.2","name":"gpt-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":9.44,"cache_read":0.12},"canonical_model_id":"openai/gpt-5.2"},"anthropic--claude-3.7-sonnet":{"id":"anthropic--claude-3.7-sonnet","name":"anthropic--claude-3.7-sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2024-10-31","release_date":"2025-02-24","last_updated":"2025-02-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-3-7-sonnet-20250219"},"amazon--titan-embed-text":{"id":"amazon--titan-embed-text","name":"amazon--titan-embed-text","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-04-30","last_updated":"2024-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536},"cost":{"input":0.14,"output":0}},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"canonical_model_id":"openai/gpt-6-luna"},"anthropic--claude-4.5-opus":{"id":"anthropic--claude-4.5-opus","name":"anthropic--claude-4.5-opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"gpt-5":{"id":"gpt-5","name":"gpt-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"sonar-pro":{"id":"sonar-pro","name":"sonar-pro","description":"Advanced Sonar search model for deeper research and cited synthesis","family":"sonar-pro","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":3,"output":15},"canonical_model_id":"perplexity/sonar-pro"},"text-embedding-3-large":{"id":"text-embedding-3-large","name":"text-embedding-3-large","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8191,"output":3072},"cost":{"input":0.09,"output":0}},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"gemini-2.5-flash-lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"gpt-5.6-terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.6-terra"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"gemini-2.5-pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-03-25","last_updated":"2025-06-05","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"google/gemini-2.5-pro"},"gpt-4.1":{"id":"gpt-4.1","name":"gpt-4.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.32},"canonical_model_id":"openai/gpt-4.1"},"amazon--nova-lite":{"id":"amazon--nova-lite","name":"amazon--nova-lite","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.3,"output":2.37}},"mistralai--mistral-small":{"id":"mistralai--mistral-small","name":"mistralai--mistral-small","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.07,"output":0.28},"canonical_model_id":"mistral/mistral-small-2603"},"gpt-5-mini":{"id":"gpt-5-mini","name":"gpt-5-mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"gemini-3.1-flash-lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"anthropic--claude-3-sonnet":{"id":"anthropic--claude-3-sonnet","name":"anthropic--claude-3-sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-08-31","release_date":"2024-03-04","last_updated":"2024-03-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":4096},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"gpt-5.6-luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":1,"output":6,"cache_read":0.1,"tiers":[{"input":2,"output":9,"cache_read":0.2,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":2,"output":9,"cache_read":0.2}},"canonical_model_id":"openai/gpt-5.6-luna"},"gemini-embedding-2":{"id":"gemini-embedding-2","type":"embedding","name":"Gemini Embedding 2","description":"Multimodal embedding model mapping text, images, video, audio, and PDFs into a unified embedding space","family":"gemini","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-11","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":3072},"canonical_model_id":"google/gemini-embedding-2"},"anthropic--claude-3-opus":{"id":"anthropic--claude-3-opus","name":"anthropic--claude-3-opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-08-31","release_date":"2024-02-29","last_updated":"2024-02-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":4096},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-3.8-flash"},"gpt-5.4":{"id":"gpt-5.4","name":"gpt-5.4","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"nvidia--llama-3.2-nv-embedqa-1b":{"id":"nvidia--llama-3.2-nv-embedqa-1b","name":"nvidia--llama-3.2-nv-embedqa-1b","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":4096},"cost":{"input":0.07,"output":0}},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"gemini-3.5-flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"gemini-embedding":{"id":"gemini-embedding","type":"embedding","name":"Gemini Embedding 001","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-05","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2048,"output":1},"canonical_model_id":"google/gemini-embedding-001"},"amazon--nova-micro":{"id":"amazon--nova-micro","name":"amazon--nova-micro","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.03,"output":0.1}},"sonar":{"id":"sonar","name":"sonar","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":1,"output":1},"canonical_model_id":"perplexity/sonar"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"canonical_model_id":"openai/gpt-6-astra"},"mistralai--mistral-medium":{"id":"mistralai--mistral-medium","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"canonical_model_id":"mistral/mistral-medium-2604"},"nvidia--llama-nemotron-embed-1b":{"id":"nvidia--llama-nemotron-embed-1b","type":"embedding","name":"Llama Nemotron Embed 1B v2","description":"Multilingual text embedding model for retrieval and question answering pipelines","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-16","last_updated":"2025-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":2048},"canonical_model_id":"nvidia/llama-nemotron-embed-1b-v2"},"anthropic--claude-4-opus":{"id":"anthropic--claude-4-opus","name":"anthropic--claude-4-opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-20250514"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"gpt-5.6-sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.6-sol"},"cohere--command-a-reasoning":{"id":"cohere--command-a-reasoning","name":"cohere--command-a-reasoning","description":"Cohere reasoning model for multilingual enterprise agents, tools, and complex workflows","family":"command-a","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high"]},{"type":"budget_tokens","min":1}],"tool_call":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"cost":{"input":0.63,"output":5.05},"canonical_model_id":"cohere/command-a-reasoning-08-2025"},"anthropic--claude-3.5-sonnet":{"id":"anthropic--claude-3.5-sonnet","name":"anthropic--claude-3.5-sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04-30","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-3-5-sonnet-20241022"},"anthropic--claude-4.8-opus":{"id":"anthropic--claude-4.8-opus","name":"anthropic--claude-4.8-opus","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"sap-abap-1":{"id":"sap-abap-1","name":"sap-abap-1","description":"SAP-hosted model for ABAP code generation and enterprise development tasks","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-11-26","last_updated":"2025-11-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":4096},"cost":{"input":0.48,"output":1.7}},"anthropic--claude-3-haiku":{"id":"anthropic--claude-3-haiku","name":"anthropic--claude-3-haiku","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-08-31","release_date":"2024-03-13","last_updated":"2024-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":4096},"cost":{"input":0.25,"output":1.25,"cache_read":0.03,"cache_write":0.3},"canonical_model_id":"anthropic/claude-3-haiku-20240307"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"canonical_model_id":"openai/gpt-6.1-sol"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-3.5-flash-lite"},"anthropic--claude-4.6-opus":{"id":"anthropic--claude-4.6-opus","name":"anthropic--claude-4.6-opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic--claude-4.7-opus":{"id":"anthropic--claude-4.7-opus","name":"anthropic--claude-4.7-opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"text-embedding-3-small":{"id":"text-embedding-3-small","name":"text-embedding-3-small","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8191,"output":1536},"cost":{"input":0.02,"output":0}},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"gemini-2.5-flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-04-17","last_updated":"2025-06-05","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-3.6-flash"},"gpt-5-nano":{"id":"gpt-5-nano","name":"gpt-5-nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"mistralai--mistral-medium-instruct":{"id":"mistralai--mistral-medium-instruct","name":"mistralai--mistral-medium-instruct","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.36,"output":1.22},"canonical_model_id":"mistral/mistral-medium-2505"},"amazon--nova-pro":{"id":"amazon--nova-pro","name":"amazon--nova-pro","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":8192},"cost":{"input":0.56,"output":2.13}}}},"coralbricks":{"id":"coralbricks","env":["CORAL_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.coralbricks.ai/v1","name":"CoralBricks","doc":"https://www.coralbricks.ai/docs","models":{"deepseek-v4.1-flash-fast":{"id":"deepseek-v4.1-flash-fast","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.01,"output":1.2,"cache_read":0,"cache_write":0.09},"canonical_model_id":"deepseek/deepseek-v4.1-flash"}}},"wallaby":{"id":"wallaby","env":["WALLABY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.wallabytoken.com/v1","name":"Wallaby","doc":"https://wallabytoken.com/docs","models":{"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":2.7,"output":13.5,"cache_read":0.27},"canonical_model_id":"moonshotai/kimi-k3"}}},"kenari":{"id":"kenari","env":["KENARI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://kenari.id/v1","name":"Kenari","doc":"https://kenari.id/docs","models":{"nano-banana-2":{"id":"nano-banana-2","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","video","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemini-3.1-flash-image"},"gpt-5-6-sol":{"id":"gpt-5-6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.6-sol"},"kimi-k2-6":{"id":"kimi-k2-6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"gemini-3-1-pro":{"id":"gemini-3-1-pro","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemini-3.1-pro-preview"},"step-3-7-flash":{"id":"step-3-7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0,"output":0},"canonical_model_id":"stepfun/step-3.7-flash"},"whisper-large-v3-turbo":{"id":"whisper-large-v3-turbo","type":"transcription","name":"Whisper Large v3 Turbo","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-10-01","last_updated":"2024-10-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":448},"cost":{"input":0,"output":0},"canonical_model_id":"openai/whisper-large-v3-turbo"},"grok-4-5":{"id":"grok-4-5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":0,"output":0},"canonical_model_id":"xai/grok-4.5"},"mistral-medium-3-5:free":{"id":"mistral-medium-3-5:free","name":"Mistral Medium 3.5 (Free)","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"mistral/mistral-medium-2604"},"hy3:free":{"id":"hy3:free","name":"Hy3 (Free)","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"tencent/hy3"},"gemini-2-5-flash-lite":{"id":"gemini-2-5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemini-2.5-flash-lite"},"nano-banana-2-lite":{"id":"nano-banana-2-lite","type":"image","name":"Nano Banana 2 Lite","description":"Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-pro"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"anthropic/claude-fable-5"},"mimo-v2-5-pro":{"id":"mimo-v2-5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"gpt-5-6-terra":{"id":"gpt-5-6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.6-terra"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"moonshotai/kimi-k3"},"kimi-k2-6:free":{"id":"kimi-k2-6:free","name":"Kimi K2.6 (Free)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"deepseek-v4-1-flash":{"id":"deepseek-v4-1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen3-8-max":{"id":"qwen3-8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.8-max"},"glm-4-7-flash:free":{"id":"glm-4-7-flash:free","name":"GLM-4.7-Flash (Free)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"mimo-v2-5":{"id":"mimo-v2-5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"nemotron-3-ultra-550b-a55b":{"id":"nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"minimax-m2-7":{"id":"minimax-m2-7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"minimax/MiniMax-M2.7"},"gemini-3-1-flash-tts":{"id":"gemini-3-1-flash-tts","type":"speech","name":"Gemini 3.1 Flash TTS Preview","description":"Low-latency speech generation with steerable prompts and expressive audio tags","family":"gemini-flash","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-15","last_updated":"2026-04-15","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":16384},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemini-3.1-flash-tts-preview"},"kimi-k2-7-code:free":{"id":"kimi-k2-7-code:free","name":"Kimi K2.7 Code (Free)","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"glm-5-1":{"id":"glm-5-1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.1"},"hy3":{"id":"hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"tencent/hy3"},"grok-imagine-image-2-0":{"id":"grok-imagine-image-2-0","type":"image","name":"Grok Imagine Image 2.0","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-08-07","last_updated":"2026-08-07","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":64000,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"xai/grok-imagine-image-2.0"},"gpt-5-4-mini":{"id":"gpt-5-4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.4-mini"},"gpt-image-2":{"id":"gpt-image-2","type":"image","name":"GPT-Image-2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":272000,"output":16384},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-image-2"},"gemini-3-1-flash-lite":{"id":"gemini-3-1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemini-3.1-flash-lite"},"kimi-k2-7-code":{"id":"kimi-k2-7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"nano-banana-pro":{"id":"nano-banana-pro","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemini-3-pro-image"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"anthropic/claude-opus-4-8"},"grok-4-6":{"id":"grok-4-6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":0,"output":0},"canonical_model_id":"xai/grok-4.6"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0,"output":0}},"gpt-5-5":{"id":"gpt-5-5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.5"},"gpt-5-6-luna":{"id":"gpt-5-6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.6-luna"},"mimo-v2-5:free":{"id":"mimo-v2-5:free","name":"MiMo-V2.5 (Free)","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"grok-build-0-1":{"id":"grok-build-0-1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0,"output":0},"canonical_model_id":"xai/grok-build-0.1"},"glm-5-3":{"id":"glm-5-3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3"},"gemini-3-6-flash":{"id":"gemini-3-6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemini-3.6-flash"},"grok-imagine-image":{"id":"grok-imagine-image","type":"image","name":"Grok Imagine Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-28","last_updated":"2026-01-28","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":16000,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"xai/grok-imagine-image"},"glm-5-3-flash":{"id":"glm-5-3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"anthropic/claude-opus-4-7"},"nemotron-3-nano-30b-a3b":{"id":"nemotron-3-nano-30b-a3b","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"anthropic/claude-opus-5"},"glm-5-2":{"id":"glm-5-2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.2"},"nemotron-3-super-120b-a12b":{"id":"nemotron-3-super-120b-a12b","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"gemini-3-5-flash":{"id":"gemini-3-5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemini-3.5-flash"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0,"output":0},"canonical_model_id":"minimax/MiniMax-M3"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-oss-120b"},"mistral-large:free":{"id":"mistral-large:free","name":"Mistral Large (Free)","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"mistral/mistral-large-latest"},"gemini-2-5-flash":{"id":"gemini-2-5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemini-2.5-flash"},"gemma-4-31b-it":{"id":"gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-31b-it"},"step-3-7-flash:free":{"id":"step-3-7-flash:free","name":"Step 3.7 Flash (Free)","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0,"output":0},"canonical_model_id":"stepfun/step-3.7-flash"},"nemotron-3-super-120b-a12b:free":{"id":"nemotron-3-super-120b-a12b:free","name":"Nemotron 3 Super 120B A12B (Free)","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"anthropic/claude-sonnet-5"},"deepseek-v4-flash:free":{"id":"deepseek-v4-flash:free","name":"DeepSeek V4 Flash (Free)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"minimax-m2-7-highspeed":{"id":"minimax-m2-7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"qwen3-7-plus":{"id":"qwen3-7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.7-plus"},"gemini-3-7-flash":{"id":"gemini-3-7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemini-3.7-flash"}}},"lilac":{"id":"lilac","env":["LILAC_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.getlilac.com/v1","name":"Lilac","doc":"https://docs.getlilac.com/inference/models","models":{"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.7,"output":3.5,"cache_read":0.2},"canonical_model_id":"moonshotai/kimi-k2.6"},"minimaxai/minimax-m3":{"id":"minimaxai/minimax-m3","name":"MiniMax M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax-m3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.28,"output":1.1,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M3"},"zai-org/glm-5.2":{"id":"zai-org/glm-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":524288},"cost":{"input":0.9,"output":3,"cache_read":0.27},"canonical_model_id":"zhipuai/glm-5.2"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262100,"output":262100},"cost":{"input":0.11,"output":0.35},"canonical_model_id":"google/gemma-4-31b-it"}}},"zai-coding-plan":{"id":"zai-coding-plan","env":["ZHIPU_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.z.ai/api/coding/paas/v4","name":"Z.AI Coding Plan","doc":"https://docs.z.ai/devpack/overview","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5.3-highspeed":{"id":"glm-5.3-highspeed","name":"GLM-5.3 Highspeed","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7"},"glm-5-turbo":{"id":"glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5-turbo"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5.2-highspeed":{"id":"glm-5.2-highspeed","name":"GLM-5.2 Highspeed","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"}}},"gmicloud":{"id":"gmicloud","env":["GMICLOUD_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.gmi-serving.com/v1","name":"GMI Cloud","doc":"https://docs.gmicloud.ai/inference-engine/api-reference/llm-api-reference","models":{"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":409600,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":4.5,"output":22.5,"cache_read":0.45},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":127999}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":409600,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":409600,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"deepseek-ai/DeepSeek-V4-Flash":{"id":"deepseek-ai/DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048575,"output":384000},"cost":{"input":0.112,"output":0.224,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.392,"output":2.784,"cache_read":0.116},"canonical_model_id":"deepseek/deepseek-v4-pro"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.855,"output":3.6,"cache_read":0.144},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k2.7-code-highspeed":{"id":"moonshotai/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.6,"output":2.4,"cache_read":0.12,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.24,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.24}},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMaxAI/MiniMax-M2.7":{"id":"MiniMaxAI/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"Qwen/Qwen3.8-Flash":{"id":"Qwen/Qwen3.8-Flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048575,"output":131072},"cost":{"input":0.16,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"Qwen/Qwen3.8-Max":{"id":"Qwen/Qwen3.8-Max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"Qwen/Qwen3.7-Max":{"id":"Qwen/Qwen3.7-Max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.7-max"},"zai-org/GLM-5-FP8":{"id":"zai-org/GLM-5-FP8","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.6,"output":1.92,"cache_read":0.12},"canonical_model_id":"zhipuai/glm-5"},"zai-org/GLM-5.1-FP8":{"id":"zai-org/GLM-5.1-FP8","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.98,"output":3.08,"cache_read":0.182},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-5.2-FP8":{"id":"zai-org/GLM-5.2-FP8","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.979,"output":3.08,"cache_read":0.182},"canonical_model_id":"zhipuai/glm-5.2"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"}}},"minimax-coding-plan":{"id":"minimax-coding-plan","env":["MINIMAX_API_KEY"],"npm":"@ai-sdk/anthropic","api":"https://api.minimax.io/anthropic/v1","name":"MiniMax Token Plan (minimax.io)","doc":"https://platform.minimax.io/docs/token-plan/intro","models":{"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":524288},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.7-highspeed":{"id":"MiniMax-M2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M3.1-Flash":{"id":"MiniMax-M3.1-Flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"MiniMax-M2":{"id":"MiniMax-M2","name":"MiniMax-M2","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M3.1-Flash-Preview":{"id":"MiniMax-M3.1-Flash-Preview","name":"MiniMax-M3.1-Flash-Preview","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-27","last_updated":"2026-09-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":512000},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash-Preview"},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.1":{"id":"MiniMax-M2.1","name":"MiniMax-M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.5-highspeed":{"id":"MiniMax-M2.5-highspeed","name":"MiniMax-M2.5-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}}}},"pioneer":{"id":"pioneer","env":["PIONEER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.fastino.ai/v1","name":"Pioneer","doc":"https://agent.pioneer.ai/llms.txt","models":{"gpt-4o-mini":{"id":"gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075,"cache_write":0.15},"canonical_model_id":"openai/gpt-4o-mini"},"gemini-3-flash":{"id":"gemini-3-flash","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65000},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.083333},"canonical_model_id":"google/gemini-3-flash-preview"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":5},"canonical_model_id":"openai/gpt-5.5"},"claude-opus-4-1":{"id":"claude-opus-4-1","name":"Claude Opus 4.1 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"mistral-medium-3.5":{"id":"mistral-medium-3.5","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":131072},"cost":{"input":1.5,"output":7.5,"cache_read":1.5,"cache_write":1.5},"canonical_model_id":"mistral/mistral-medium-2604"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.05,"cache_write":0.1},"canonical_model_id":"openai/gpt-4.1-nano"},"ministral-3b":{"id":"ministral-3b","name":"Ministral 3B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-03","release_date":"2024-10-16","last_updated":"2024-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.1,"output":0.1,"cache_read":0.1,"cache_write":0.1},"canonical_model_id":"mistral/ministral-3b"},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.2,"cache_write":0.4},"canonical_model_id":"openai/gpt-4.1-mini"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5"},"devstral-small-2":{"id":"devstral-small-2","name":"Devstral Small 2","description":"Compact multimodal coding model for repository exploration, file editing, and software agents","family":"devstral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":131072},"cost":{"input":0.1,"output":0.3,"cache_read":0.1,"cache_write":0.1},"canonical_model_id":"mistral/devstral-small-2"},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.325,"output":1.95,"cache_read":0.065,"cache_write":0.40625},"canonical_model_id":"alibaba/qwen3.6-plus"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"openai/gpt-5.6-terra"},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1875,"output":1.125,"cache_read":0.0375,"cache_write":0.234375},"canonical_model_id":"alibaba/qwen3.6-flash"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":1,"cache_write":2},"canonical_model_id":"openai/gpt-4.1"},"devstral-2":{"id":"devstral-2","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":131072},"cost":{"input":0.4,"output":2,"cache_read":0.4,"cache_write":0.4},"canonical_model_id":"mistral/devstral-2512"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025,"cache_write":0.25},"canonical_model_id":"openai/gpt-5-mini"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65000},"cost":{"input":0.25,"output":1.5,"cache_read":0.03,"cache_write":0.25},"canonical_model_id":"google/gemini-3.1-flash-lite"},"gemini-3.1-pro":{"id":"gemini-3.1-pro","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3.1-pro-preview"},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.32,"output":1.28,"cache_read":0.064,"cache_write":0.4},"canonical_model_id":"alibaba/qwen3.7-plus"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":1,"output":6,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"openai/gpt-5.6-luna"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.4"},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25,"cache_write":2.5},"canonical_model_id":"openai/gpt-4o"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.5-flash"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"magistral-medium":{"id":"magistral-medium","name":"Magistral Medium (latest)","description":"Mistral reasoning model for transparent analysis, math, and complex decisions","family":"magistral-medium","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-06","release_date":"2025-03-17","last_updated":"2025-03-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"cost":{"input":2,"output":5,"cache_read":2,"cache_write":2},"canonical_model_id":"mistral/magistral-medium-latest"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.5,"cache_write":2},"canonical_model_id":"xai/grok-4.5"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"qwen3.6-max-preview":{"id":"qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":240000,"output":64000},"cost":{"input":1.04,"output":6.24,"cache_read":0.208,"cache_write":1.3},"canonical_model_id":"alibaba/qwen3.6-max-preview"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02,"cache_write":0.2},"canonical_model_id":"openai/gpt-5.4-nano"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65000},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.3},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":131072},"cost":{"input":1.25,"output":10,"cache_read":0.125,"cache_write":1.25},"canonical_model_id":"openai/gpt-5.1"},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":64000},"cost":{"input":1.25,"output":3.75,"cache_read":0.25,"cache_write":1.5625},"canonical_model_id":"alibaba/qwen3.7-max"},"claude-3-7-sonnet-latest":{"id":"claude-3-7-sonnet-latest","name":"Claude Sonnet 3.7","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-10-31","release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-3-7-sonnet-20250219"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075,"cache_write":0.75},"canonical_model_id":"openai/gpt-5.4-mini"},"mistral-large-3":{"id":"mistral-large-3","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":131072},"cost":{"input":0.5,"output":1.5,"cache_read":0.5,"cache_write":0.5},"canonical_model_id":"mistral/mistral-large-2512"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":1.5,"output":7.5,"cache_read":0.15,"cache_write":1.5},"canonical_model_id":"google/gemini-3.6-flash"},"mistral-medium":{"id":"mistral-medium","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"cost":{"input":0.4,"output":2,"cache_read":0.4,"cache_write":0.4},"canonical_model_id":"mistral/mistral-medium-2505"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"ministral-14b":{"id":"ministral-14b","name":"Ministral 14B","description":"Compact multimodal Mistral model for local assistants, edge agents, and efficient tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":131072},"cost":{"input":0.2,"output":0.2,"cache_read":0.2,"cache_write":0.2},"canonical_model_id":"mistral/ministral-14b"},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005,"cache_write":0.05},"canonical_model_id":"openai/gpt-5-nano"},"claude-opus-5-fast":{"id":"claude-opus-5-fast","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175,"cache_write":1.75},"canonical_model_id":"openai/gpt-5.3-codex"},"pioneer/auto":{"id":"pioneer/auto","name":"Pioneer Auto","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2024-01-01","last_updated":"2025-06-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":4096}},"nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-NVFP4":{"id":"nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-NVFP4","name":"Nemotron 3.5 Super VL 120B A12B","description":"NVIDIA vision-language reasoning model for text, image, and video understanding and agentic tool use","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":4096},"cost":{"input":0.08,"output":0.45,"cache_read":0.04,"cache_write":0.08},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16":{"id":"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.05,"output":0.2,"cache_read":0.05,"cache_write":0.05},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16":{"id":"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65000},"cost":{"input":0.5,"output":2.5,"cache_read":0.15,"cache_write":0.5},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8":{"id":"nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"cost":{"input":0.09,"output":0.45,"cache_read":0.09,"cache_write":0.09},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16":{"id":"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16","name":"Nemotron 3.5 Lightning 30B A3B","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.5,"output":0.5,"cache_read":0.5,"cache_write":0.5},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"LiquidAI/LFM2-24B-A2B":{"id":"LiquidAI/LFM2-24B-A2B","name":"LFM2 24B A2B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"liquid","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-01-31","last_updated":"2026-02-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.03,"output":0.12,"cache_read":0.03,"cache_write":0.03}},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15,"cache_write":1.25},"canonical_model_id":"meta/muse-spark-1.1"},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek-V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":131072},"cost":{"input":0.56,"output":1.68,"cache_read":0.56,"cache_write":0.56},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek-ai/DeepSeek-V3":{"id":"deepseek-ai/DeepSeek-V3","name":"DeepSeek-V3","description":"Open DeepSeek MoE chat model for coding, math, and general reasoning","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":8192},"cost":{"input":0.27,"output":1.12,"cache_read":0.135,"cache_write":0.27},"canonical_model_id":"deepseek/deepseek-v3"},"deepseek-ai/DeepSeek-V4-Flash":{"id":"deepseek-ai/DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.0197,"cache_write":0.1},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.003625,"cache_write":0.435},"canonical_model_id":"deepseek/deepseek-v4-pro"},"meta-llama/Llama-3.3-70B-Instruct":{"id":"meta-llama/Llama-3.3-70B-Instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":0.9,"output":0.9,"cache_read":0.9,"cache_write":0.9},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/Llama-3.2-1B-Instruct":{"id":"meta-llama/Llama-3.2-1B-Instruct","name":"Llama 3.2 1B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-31","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":60000},"cost":{"input":0.1,"output":0.201,"cache_read":0.1,"cache_write":0.1}},"meta-llama/Llama-3.2-3B":{"id":"meta-llama/Llama-3.2-3B","name":"Llama-3.2-3B","description":"Small open Llama base model for lightweight text generation and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.1,"output":0.1,"cache_read":0.1,"cache_write":0.1},"canonical_model_id":"meta/llama-3.2-3b"},"meta-llama/Llama-3.1-8B-Instruct":{"id":"meta-llama/Llama-3.1-8B-Instruct","name":"Llama 3.1 8B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2023-12-31","release_date":"2024-06-30","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.2,"output":0.2,"cache_read":0.2,"cache_write":0.2}},"meta-llama/Llama-3.2-1B":{"id":"meta-llama/Llama-3.2-1B","name":"Llama-3.2-1B","description":"Compact open Llama base model for lightweight and on-device use","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.1,"output":0.1,"cache_read":0.1,"cache_write":0.1},"canonical_model_id":"meta/llama-3.2-1b"},"meta-llama/Llama-3.2-3B-Instruct":{"id":"meta-llama/Llama-3.2-3B-Instruct","name":"Llama 3.2 3B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-31","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":80000},"cost":{"input":0.1,"output":0.335,"cache_read":0.1,"cache_write":0.1}},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":131072},"cost":{"input":0.95,"output":4,"cache_read":0.34,"cache_write":0.95},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/Kimi-K3-Fast":{"id":"moonshotai/Kimi-K3-Fast","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":4.5,"output":22.5,"cache_read":0.45,"cache_write":4.5},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"cost":{"input":0.95,"output":4,"cache_read":0.19,"cache_write":0.95},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3},"canonical_model_id":"moonshotai/kimi-k3"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.3},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMaxAI/MiniMax-M2.7":{"id":"MiniMaxAI/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131000},"cost":{"input":0.279,"output":1.2,"cache_read":0.279,"cache_write":0.279},"canonical_model_id":"minimax/MiniMax-M2.7"},"poolside/laguna-s-2.1":{"id":"poolside/laguna-s-2.1","name":"Laguna S 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.01,"cache_write":0.1},"canonical_model_id":"poolside/laguna-s-2.1"},"XiaomiMiMo/MiMo-V2.5-Pro":{"id":"XiaomiMiMo/MiMo-V2.5-Pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":131000},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036,"cache_write":0.435},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"XiaomiMiMo/MiMo-V2.5":{"id":"XiaomiMiMo/MiMo-V2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028,"cache_write":0.14},"canonical_model_id":"xiaomi/mimo-v2.5"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.5,"output":1.2,"cache_read":0.1,"cache_write":0.5},"canonical_model_id":"thinkingmachines/inkling-small"},"fastino/gliner2.5-multi-v1":{"id":"fastino/gliner2.5-multi-v1","name":"GLiNER 2.5 Multi","description":"Multilingual boundary NER and span extraction; non-trainable encoder.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-24","last_updated":"2026-08-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}},"fastino/gliner2-multi-v1":{"id":"fastino/gliner2-multi-v1","name":"GLiNER2 Multi","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-11-30","last_updated":"2025-11-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}},"fastino/gliner2-large-v1":{"id":"fastino/gliner2-large-v1","name":"GLiNER2 Large","description":"Flagship model for demanding analysis, coding, and production agent workflows","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-06-30","last_updated":"2025-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}},"fastino/gliguard-LLMGuardrails-300M":{"id":"fastino/gliguard-LLMGuardrails-300M","name":"GLiGuard LLM Guardrails 300M","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}},"fastino/GLiNER-2.5-Decide":{"id":"fastino/GLiNER-2.5-Decide","name":"GLiNER-2.5-Decide","description":"Classification-tuned GLiNER2 large. Entities, relations, and JSON extraction still work.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-24","last_updated":"2026-09-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}},"fastino/gliner2-multi-large-v1":{"id":"fastino/gliner2-multi-large-v1","name":"GLiNER2 Multi Large","description":"Flagship model for demanding analysis, coding, and production agent workflows","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-11-30","last_updated":"2025-11-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}},"fastino/gliner2-base-v1":{"id":"fastino/gliner2-base-v1","name":"GLiNER2 Base","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-06-30","last_updated":"2025-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}},"fastino/gliner2-privacy-filter-PII-multi":{"id":"fastino/gliner2-privacy-filter-PII-multi","name":"GLiNER2 Privacy Filter PII (Multi)","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}},"fastino/gliner2.5-base-v1":{"id":"fastino/gliner2.5-base-v1","name":"GLiNER 2.5 Base","description":"English boundary NER and structured extraction with LoRA fine-tuning.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-24","last_updated":"2026-08-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":0.03,"output":0,"cache_read":0.03,"cache_write":0.03}},"fastino/gliguard-PII-multi":{"id":"fastino/gliguard-PII-multi","name":"GLiNER2-Guardrails-PII-Multi","description":"A 300M-parameter multilingual model that runs LLM safety moderation and PII detection in a single forward pass.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}},"mistralai/Ministral-8B-Instruct-2410":{"id":"mistralai/Ministral-8B-Instruct-2410","name":"Ministral 8B Instruct","description":"Efficient open Mistral edge model for on-device chat and function calling","family":"ministral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-10-16","last_updated":"2024-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15},"canonical_model_id":"mistral/ministral-8b-instruct-2410"},"mistralai/Magistral-Small-2506":{"id":"mistralai/Magistral-Small-2506","name":"Magistral Small","description":"Open Mistral reasoning model for transparent step-by-step problem solving","family":"magistral","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":64000},"cost":{"input":0.5,"output":1.5,"cache_read":0.5,"cache_write":0.5},"canonical_model_id":"mistral/magistral-small-2506"},"mistralai/Mistral-Nemo-Instruct-2407":{"id":"mistralai/Mistral-Nemo-Instruct-2407","name":"Mistral Nemo","description":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.02,"output":0.03,"cache_read":0.02,"cache_write":0.02},"canonical_model_id":"mistral/mistral-nemo"},"mistralai/Mistral-Small-4-119B-2603":{"id":"mistralai/Mistral-Small-4-119B-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015,"cache_write":0.15},"canonical_model_id":"mistral/mistral-small-2603"},"mistralai/Pixtral-12B-2409":{"id":"mistralai/Pixtral-12B-2409","name":"Pixtral 12B","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2024-09-11","last_updated":"2024-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15},"canonical_model_id":"mistral/pixtral-12b"},"mistralai/Mistral-7B-Instruct-v0.3":{"id":"mistralai/Mistral-7B-Instruct-v0.3","name":"Mistral 7B Instruct v0.3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2023-04-30","last_updated":"2023-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.2,"output":0.2,"cache_read":0.2,"cache_write":0.2}},"mistralai/Codestral-22B-v0.1":{"id":"mistralai/Codestral-22B-v0.1","name":"Codestral-22B-v0.1","description":"Open Mistral code model for fill-in-the-middle and 80+ programming languages","family":"codestral","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-05-29","last_updated":"2024-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.3,"output":0.9,"cache_read":0.3,"cache_write":0.3},"canonical_model_id":"mistral/codestral-22b-v0.1"},"Qwen/Qwen3.6-35B-A3B":{"id":"Qwen/Qwen3.6-35B-A3B","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.14,"output":1,"cache_read":0.028,"cache_write":0.175},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen/Qwen3-4B-Instruct-2507":{"id":"Qwen/Qwen3-4B-Instruct-2507","name":"Qwen3 4B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.2,"output":0.2,"cache_read":0.2,"cache_write":0.2}},"Qwen/Qwen3-4B-Base":{"id":"Qwen/Qwen3-4B-Base","name":"Qwen3 4B Base","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-03-31","last_updated":"2025-03-31","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}},"Qwen/Qwen3-1.7B-Base":{"id":"Qwen/Qwen3-1.7B-Base","name":"Qwen3 1.7B Base","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-03-31","last_updated":"2025-03-31","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.1,"output":0.1,"cache_read":0.1,"cache_write":0.1}},"Qwen/Qwen3.6-27B":{"id":"Qwen/Qwen3.6-27B","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.6,"output":0.6,"cache_read":0.6,"cache_write":0.6},"canonical_model_id":"alibaba/qwen3.6-27b"},"Qwen/Qwen3.5-9B":{"id":"Qwen/Qwen3.5-9B","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":0.3,"cache_read":0.3,"cache_write":0.3},"canonical_model_id":"alibaba/qwen3.5-9b"},"Qwen/Qwen3-32B":{"id":"Qwen/Qwen3-32B","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.9,"output":0.9,"cache_read":0.9,"cache_write":0.9},"canonical_model_id":"alibaba/qwen3-32b"},"Qwen/Qwen3-8B":{"id":"Qwen/Qwen3-8B","name":"Qwen3 8B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-03-31","release_date":"2025-03-31","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":40960},"cost":{"input":0.2,"output":0.2,"cache_read":0.2,"cache_write":0.2}},"Qwen/Qwen2.5-Coder-0.5B":{"id":"Qwen/Qwen2.5-Coder-0.5B","name":"Qwen2.5-Coder-0.5B","description":"Tiny open Qwen code model for lightweight completion and on-device coding","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.1,"output":0.1,"cache_read":0.1,"cache_write":0.1},"canonical_model_id":"alibaba/qwen2.5-coder-0.5b"},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B-A22B Instruct 2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":1.2,"output":1.2,"cache_read":1.2,"cache_write":1.2},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202000,"output":131072},"cost":{"input":0.98,"output":3.08,"cache_read":0.182,"cache_write":0.98},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1040000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":1.4},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.2-Fast":{"id":"zai-org/GLM-5.2-Fast","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":2.1,"output":6.6,"cache_read":0.21,"cache_write":2.1},"canonical_model_id":"zhipuai/glm-5.2"},"google/gemma-4-E4B-it":{"id":"google/gemma-4-E4B-it","name":"Gemma 4 E4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.2,"output":0.2,"cache_read":0.2,"cache_write":0.2},"canonical_model_id":"google/gemma-4-E4B-it"},"google/gemma-3-4b-pt":{"id":"google/gemma-3-4b-pt","name":"Gemma 3 4B (Pretrained)","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-02-28","last_updated":"2025-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}},"google/gemma-4-12B-it":{"id":"google/gemma-4-12B-it","name":"Gemma 4 12B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-05-31","last_updated":"2026-05-31","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.25,"output":0.25,"cache_read":0.25,"cache_write":0.25}},"google/gemma-4-E2B-it":{"id":"google/gemma-4-E2B-it","name":"Gemma 4 E2B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.1,"output":0.1,"cache_read":0.1,"cache_write":0.1},"canonical_model_id":"google/gemma-4-E2B-it"},"google/diffusiongemma-26B-A4B-it":{"id":"google/diffusiongemma-26B-A4B-it","name":"DiffusionGemma 26B-A4B IT","description":"Gemini model for general assistance, reasoning, and multimodal workflows","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-05-31","last_updated":"2026-05-31","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.5,"output":0.5,"cache_read":0.5,"cache_write":0.5}},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.5,"output":0.5,"cache_read":0.5,"cache_write":0.5},"canonical_model_id":"google/gemma-4-31b-it"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":5},"canonical_model_id":"sakana/fugu-ultra"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.07,"output":0.3,"cache_read":0.035,"cache_write":0.07},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.15,"output":0.6,"cache_read":0.015,"cache_write":0.15},"canonical_model_id":"openai/gpt-oss-120b"},"HuggingFaceTB/SmolLM3-3B-Base":{"id":"HuggingFaceTB/SmolLM3-3B-Base","name":"SmolLM3 3B Base","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-06-30","last_updated":"2025-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}}}},"inception":{"id":"inception","env":["INCEPTION_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.inceptionlabs.ai/v1/","name":"Inception","doc":"https://docs.inceptionlabs.ai/get-started/models","models":{"mercury-2":{"id":"mercury-2","name":"Mercury 2","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":50000},"cost":{"input":0.25,"output":0.75,"cache_read":0.025}},"mercury-2.5":{"id":"mercury-2.5","name":"Mercury 2.5","description":"Mercury 2.5 is the fastest reasoning LLM, and the latest diffusion LLM (dLLM) from Inception","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-11-01","release_date":"2026-09-08","last_updated":"2026-09-10","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":260000,"output":65536},"cost":{"input":0.04,"output":0.15,"cache_read":0.004}},"mercury-3":{"id":"mercury-3","name":"Mercury 3","description":"Inception's diffusion reasoning model for chat, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-16","last_updated":"2026-10-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1,"output":0.35,"cache_read":0.01}},"mercury-edit-2":{"id":"mercury-edit-2","name":"Mercury Edit 2","description":"Code editing dLLM for autocomplete (FIM) and next-edit suggestions","family":"mercury","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-30","last_updated":"2026-03-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":8192},"cost":{"input":0.25,"output":0.75,"cache_read":0.025}}}},"aimlapi":{"id":"aimlapi","env":["AIMLAPI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.aimlapi.com/v1","name":"AI/ML API","doc":"https://docs.aimlapi.com","models":{"undi95/remm-slerp-l2-13b":{"id":"undi95/remm-slerp-l2-13b","name":"ReMM SLERP 13B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-06-30","release_date":"2023-07-22","last_updated":"2023-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":6144,"output":4096},"cost":{"input":0.6877,"output":1.03155},"canonical_model_id":"undi95/remm-slerp-l2-13b"},"aion-labs/aion-rp-llama-3.1-8b":{"id":"aion-labs/aion-rp-llama-3.1-8b","name":"Aion-RP 1.0 (8B)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12-31","release_date":"2025-02-04","last_updated":"2025-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":1.10032,"output":2.20064},"canonical_model_id":"aion-labs/aion-rp-llama-3.1-8b"},"inception/mercury-2":{"id":"inception/mercury-2","name":"Mercury 2","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":50000},"cost":{"input":0.34385,"output":1.03155,"cache_read":0.034385},"canonical_model_id":"inception/mercury-2"},"inception/mercury-2.5":{"id":"inception/mercury-2.5","name":"Mercury 2.5","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":260000,"output":65536},"cost":{"input":0.27508,"output":1.03155,"cache_read":0.027508},"canonical_model_id":"inception/mercury-2.5"},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude 4.8 Opus","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":6.877,"output":34.385,"cache_read":0.6877},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":13,"output":65,"cache_read":1.3},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude 4.7 Opus","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":6.877,"output":34.385,"cache_read":0.6877},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.2,"output":26,"cache_read":0.26},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude 4.5 Opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":6.877,"output":34.385,"cache_read":0.6877},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.6,"output":13,"cache_read":0.26},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.13,"output":0.65,"cache_read":0.013},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":6.5,"output":32.5,"cache_read":0.65},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude 4.5 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":4.1262,"output":20.631,"cache_read":0.41262},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude 4.5 Haiku","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.3754,"output":6.877,"cache_read":0.13754},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.6,"output":13,"cache_read":0.26},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude 4.6 Sonnet","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":4.1262,"output":20.631,"cache_read":0.41262},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":13,"output":65,"cache_read":0.325},"canonical_model_id":"anthropic/claude-fable-5-1"},"amazon/nova-lite-v1":{"id":"amazon/nova-lite-v1","name":"Nova Lite 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":5120},"cost":{"input":0.082524,"output":0.330096},"canonical_model_id":"amazon/nova-lite-v1"},"amazon/nova-pro-v1":{"id":"amazon/nova-pro-v1","name":"Nova Pro 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":5120},"cost":{"input":1.10032,"output":4.40128},"canonical_model_id":"amazon/nova-pro-v1"},"amazon/nova-micro-v1":{"id":"amazon/nova-micro-v1","name":"Nova Micro 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":5120},"cost":{"input":0.048139,"output":0.192556},"canonical_model_id":"amazon/nova-micro-v1"},"moonshot/kimi-k3":{"id":"moonshot/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3.9,"output":19.5,"cache_read":0.39},"canonical_model_id":"moonshotai/kimi-k3"},"anthracite-org/magnum-v4-72b":{"id":"anthracite-org/magnum-v4-72b","name":"Magnum v4 72B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":16384},"cost":{"input":4.1262,"output":6.877},"canonical_model_id":"anthracite-org/magnum-v4-72b"},"mancer/weaver":{"id":"mancer/weaver","name":"Weaver (alpha)","description":"General-purpose chat model for instruction following, writing, and analysis","family":"alpha","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-06-30","release_date":"2023-08-02","last_updated":"2023-08-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":2000},"cost":{"input":1.03155,"output":1.3754},"canonical_model_id":"mancer/weaver"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.17,"output":3.51,"cache_read":0.039},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":2.393196,"output":4.786392,"cache_read":0.199433},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.572,"output":1.716,"cache_read":0.0182},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.39,"output":1.56,"cache_read":0.0078},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-chat":{"id":"deepseek/deepseek-chat","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.39,"output":1.56,"cache_read":0.0078},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-chat-v3-0324":{"id":"deepseek/deepseek-chat-v3-0324","name":"DeepSeek V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07-31","release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":124000},"cost":{"input":0.398866,"output":1.567956,"cache_read":0.192556},"canonical_model_id":"deepseek/deepseek-chat-v3-0324"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.39,"output":1.56,"cache_read":0.0078},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"gryphe/mythomax-l2-13b":{"id":"gryphe/mythomax-l2-13b","name":"MythoMax 13B","description":"Llama 2 merge combining MythoLogic and Huginn for creative writing and roleplay","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2023-08-10","last_updated":"2023-08-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":2.578875,"output":2.578875,"cache_read":2.578875},"canonical_model_id":"gryphe/mythomax-l2-13b"},"z-ai/glm-4.6v":{"id":"z-ai/glm-4.6v","name":"GLM 4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.41262,"output":1.23786,"cache_read":0.075647},"canonical_model_id":"zhipuai/glm-4.6v"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.195,"output":0.65,"cache_read":0.039},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-4.5v":{"id":"z-ai/glm-4.5v","name":"GLM 4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":0.82524,"output":2.47572,"cache_read":0.151294},"canonical_model_id":"zhipuai/glm-4.5v"},"z-ai/glm-4.7-flash":{"id":"z-ai/glm-4.7-flash","name":"GLM 4.7 Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"cost":{"input":0.171925,"output":0.6877,"cache_read":0.013754},"canonical_model_id":"zhipuai/glm-4.7-flash"},"z-ai/glm-5v-turbo":{"id":"z-ai/glm-5v-turbo","name":"GLM 5V Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.56,"output":5.2,"cache_read":0.312},"canonical_model_id":"zhipuai/glm-5v-turbo"},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM 5 Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":1.56,"output":5.2,"cache_read":0.312},"canonical_model_id":"zhipuai/glm-5-turbo"},"zhipu/glm-5.3":{"id":"zhipu/glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":131072},"cost":{"input":1.82,"output":5.72,"cache_read":0.338},"canonical_model_id":"zhipuai/glm-5.3"},"zhipu/glm-5.2":{"id":"zhipu/glm-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.82,"output":5.72,"cache_read":0.338},"canonical_model_id":"zhipuai/glm-5.2"},"meta/muse-glimmer-30b":{"id":"meta/muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.48139,"output":2.0631,"cache_read":0.055016},"canonical_model_id":"meta/muse-glimmer-30b"},"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.71925,"output":5.84545,"cache_read":0.20631},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.71925,"output":5.84545,"cache_read":0.20631},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.71925,"output":5.84545,"cache_read":0.20631},"canonical_model_id":"meta/muse-spark-1.2"},"bytedance-seed/seed-1.6-flash":{"id":"bytedance-seed/seed-1.6-flash","name":"Seed 1.6 Flash","description":"Low-latency ByteDance Seed model for high-throughput chat, extraction, and lightweight tool use","family":"seed","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.103155,"output":0.41262},"canonical_model_id":"bytedance-seed/seed-1-6-flash"},"bytedance-seed/seed-2.0-mini":{"id":"bytedance-seed/seed-2.0-mini","name":"Seed-2.0-Mini","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.13754,"output":0.55016},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"bytedance-seed/seed-2.0-code":{"id":"bytedance-seed/seed-2.0-code","name":"Seed-2.0-Code","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.6877,"output":4.1262},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"bytedance-seed/seed-2.0-lite":{"id":"bytedance-seed/seed-2.0-lite","name":"Seed-2.0-Lite","description":"Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.34385,"output":2.7508},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"meta-llama/Llama-3.3-70B-Instruct-Turbo":{"id":"meta-llama/Llama-3.3-70B-Instruct-Turbo","name":"Llama 3.3 70B Instruct Turbo","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":127000},"cost":{"input":1.352,"output":1.352},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"inference-net/schematron-v2-small":{"id":"inference-net/schematron-v2-small","name":"Schematron V2 Small","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":0.06877,"output":0.316342},"canonical_model_id":"inference-net/schematron-v2-small"},"inference-net/schematron-v2-turbo":{"id":"inference-net/schematron-v2-turbo","name":"Schematron V2 Turbo","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.041262,"output":0.20631},"canonical_model_id":"inference-net/schematron-v2-turbo"},"thedrummer/cydonia-24b-v4.1":{"id":"thedrummer/cydonia-24b-v4.1","name":"Cydonia 24B V4.1","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04-30","release_date":"2025-09-27","last_updated":"2025-09-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.41262,"output":0.6877,"cache_read":0.20631},"canonical_model_id":"thedrummer/cydonia-24b-v4.1"},"thedrummer/unslopnemo-12b":{"id":"thedrummer/unslopnemo-12b","name":"UnslopNemo 12B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04-30","release_date":"2024-11-08","last_updated":"2024-11-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.55016,"output":0.55016},"canonical_model_id":"thedrummer/unslopnemo-12b"},"thedrummer/skyfall-36b-v2":{"id":"thedrummer/skyfall-36b-v2","name":"Skyfall 36B V2","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2025-03-10","last_updated":"2025-03-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.75647,"output":1.10032,"cache_read":0.34385},"canonical_model_id":"thedrummer/skyfall-36b-v2"},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 0905","description":"September 2025 Kimi K2 instruction update with a 256K context and improved agentic coding","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":1.3754,"output":4.1262},"canonical_model_id":"moonshotai/kimi-k2-0905"},"moonshotai/kimi-k2":{"id":"moonshotai/kimi-k2","name":"Kimi K2 0711","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-12-31","release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.783978,"output":3.16342},"canonical_model_id":"moonshotai/kimi-k2"},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.82524,"output":3.4385,"cache_read":0.20631},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"sao10k/l3.3-euryale-70b":{"id":"sao10k/l3.3-euryale-70b","name":"Llama 3.3 Euryale 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-12-18","last_updated":"2024-12-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.89401,"output":1.03155},"canonical_model_id":"sao10k/l3.3-euryale-70b"},"sao10k/l3-lunaris-8b":{"id":"sao10k/l3-lunaris-8b","name":"Llama 3 8B Lunaris","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-13","last_updated":"2024-08-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":7372},"cost":{"input":0.06877,"output":0.06877},"canonical_model_id":"sao10k/l3-lunaris-8b"},"sao10k/l3.1-euryale-70b":{"id":"sao10k/l3.1-euryale-70b","name":"Llama 3.1 Euryale 70B v2.2","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-28","last_updated":"2024-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":1.99488,"output":1.99488},"canonical_model_id":"sao10k/l3.1-euryale-70b"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.3754,"output":5.57037,"cache_read":0.233818},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":262144},"cost":{"input":0.797732,"output":1.980576,"cache_read":0.159546},"canonical_model_id":"thinkingmachines/inkling-small"},"stepfun/step-5-preview":{"id":"stepfun/step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":64000},"cost":{"input":1.3754,"output":3.71358,"cache_read":0.06877},"canonical_model_id":"stepfun/step-5-preview"},"mistralai/mistral-small-3.1-24b-instruct":{"id":"mistralai/mistral-small-3.1-24b-instruct","name":"Small 3.1 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.482765,"output":0.763347},"canonical_model_id":"mistral/mistral-small-3.1-24b-instruct"},"mistralai/mistral-medium-3":{"id":"mistralai/mistral-medium-3","name":"Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":0.55016,"output":2.7508,"cache_read":0.055016},"canonical_model_id":"mistral/mistral-medium-3"},"mistralai/mistral-nemo":{"id":"mistralai/mistral-nemo","name":"Mistral NeMo","description":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.20631,"output":0.233818,"cache_read":0.20631},"canonical_model_id":"mistral/mistral-nemo"},"mistralai/mistral-large":{"id":"mistralai/mistral-large","name":"Large","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11-30","release_date":"2024-02-26","last_updated":"2024-02-26","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":102400},"cost":{"input":2.7508,"output":8.2524,"cache_read":0.27508},"canonical_model_id":"mistral/mistral-large"},"mistralai/mistral-small-3.2-24b-instruct":{"id":"mistralai/mistral-small-3.2-24b-instruct","name":"Small 3.2 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.103155,"output":0.27508},"canonical_model_id":"mistral/mistral-small-3.2-24b-instruct"},"mistralai/mistral-saba":{"id":"mistralai/mistral-saba","name":"Saba","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-02-17","last_updated":"2025-02-17","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":26214},"cost":{"input":0.27508,"output":0.82524,"cache_read":0.027508},"canonical_model_id":"mistral/mistral-saba"},"mistralai/ministral-8b-2512":{"id":"mistralai/ministral-8b-2512","name":"Ministral 3 8B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.41262,"output":0.41262,"cache_read":0.020631},"canonical_model_id":"mistral/ministral-8b-2512"},"mistralai/devstral-2512":{"id":"mistralai/devstral-2512","name":"Devstral 2 2512","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.605176,"output":3.02588,"cache_read":0.060518},"canonical_model_id":"mistral/devstral-2512"},"mistralai/mistral-medium-3.1":{"id":"mistralai/mistral-medium-3.1","name":"Medium 3.1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-30","release_date":"2025-08-12","last_updated":"2025-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":0.55016,"output":2.7508,"cache_read":0.055016},"canonical_model_id":"mistral/mistral-medium-3.1"},"mistralai/ministral-3b-2512":{"id":"mistralai/ministral-3b-2512","name":"Ministral 3 3B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":104857},"cost":{"input":0.20631,"output":0.20631,"cache_read":0.013754},"canonical_model_id":"mistral/ministral-3b-2512"},"mistralai/mixtral-8x22b-instruct":{"id":"mistralai/mixtral-8x22b-instruct","name":"Mixtral 8x22B Instruct","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-01-31","release_date":"2024-04-17","last_updated":"2024-04-17","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":52428},"cost":{"input":2.7508,"output":8.2524,"cache_read":0.27508},"canonical_model_id":"mistral/mixtral-8x22b-instruct"},"mistralai/mistral-medium-3-5":{"id":"mistralai/mistral-medium-3-5","name":"Medium 3.5","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":209715},"cost":{"input":2.0631,"output":10.3155},"canonical_model_id":"mistral/mistral-medium-3-5"},"mistralai/mistral-small-2603":{"id":"mistralai/mistral-small-2603","name":"Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.257888,"output":1.03155,"cache_read":0.020631},"canonical_model_id":"mistral/mistral-small-2603"},"mistralai/voxtral-small-24b-2507":{"id":"mistralai/voxtral-small-24b-2507","name":"Voxtral Small 24B 2507","description":"Open audio-language model for speech transcription, audio understanding, and voice-driven tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32768},"cost":{"input":0.13754,"output":0.41262,"cache_read":0.013754},"canonical_model_id":"mistral/voxtral-small-24b-2507"},"mistralai/ministral-14b-2512":{"id":"mistralai/ministral-14b-2512","name":"Ministral 3 14B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.48139,"output":0.48139,"cache_read":0.027508},"canonical_model_id":"mistral/ministral-14b-2512"},"mistralai/mistral-large-2407":{"id":"mistralai/mistral-large-2407","name":"Large 2407","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-03-31","release_date":"2024-07-24","last_updated":"2024-07-24","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":2.7508,"output":8.2524,"cache_read":0.27508},"canonical_model_id":"mistral/mistral-large-2407"},"mistralai/codestral-2508":{"id":"mistralai/codestral-2508","name":"Codestral 2508","description":"Mistral code model for low-latency completion, fill-in-the-middle, and code generation","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":8192},"cost":{"input":0.41262,"output":1.23786,"cache_read":0.041262},"canonical_model_id":"mistral/codestral-2508"},"mistralai/mistral-small-24b-instruct-2501":{"id":"mistralai/mistral-small-24b-instruct-2501","name":"Small 3","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-01-30","last_updated":"2025-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.06877,"output":0.110032},"canonical_model_id":"mistral/mistral-small-24b-instruct-2501"},"relace/relace-apply-3":{"id":"relace/relace-apply-3","name":"Apply 3","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2025-09-26","last_updated":"2025-09-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":1.16909,"output":1.71925},"canonical_model_id":"relace/relace-apply-3"},"relace/relace-search":{"id":"relace/relace-search","name":"Search","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":1.3754,"output":4.1262},"canonical_model_id":"relace/relace-search"},"nousresearch/hermes-3-llama-3.1-405b":{"id":"nousresearch/hermes-3-llama-3.1-405b","name":"Hermes 3 405B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-16","last_updated":"2024-08-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":1.3754,"output":1.3754},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-405b"},"nousresearch/hermes-3-llama-3.1-70b":{"id":"nousresearch/hermes-3-llama-3.1-70b","name":"Hermes 3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-18","last_updated":"2024-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.96278,"output":0.96278},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-70b"},"upstage/solar-pro-3":{"id":"upstage/solar-pro-3","name":"Solar Pro 3","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2026-01-26","last_updated":"2026-03-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.20631,"output":0.82524,"cache_read":0.020631},"canonical_model_id":"upstage/solar-pro3"},"upstage/solar-pro4":{"id":"upstage/solar-pro4","name":"Solar Pro 4","description":"Upstage's flagship model, specialized for agentic use","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.041262,"output":0.165048,"cache_read":0.008252},"canonical_model_id":"upstage/solar-pro4"},"ibm-granite/granite-4.0-h-micro":{"id":"ibm-granite/granite-4.0-h-micro","name":"Granite 4.0 Micro","description":"Compact open-weight hybrid Granite model for lightweight enterprise chat and tool calling","family":"granite","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-02","last_updated":"2025-10-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.023382,"output":0.154045},"canonical_model_id":"ibm/granite-4-h-micro"},"cohere/command-a":{"id":"cohere/command-a","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":3.4385,"output":13.754},"canonical_model_id":"cohere/command-a"},"cohere/command-r-08-2024":{"id":"cohere/command-r-08-2024","name":"Command R (08-2024)","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.20631,"output":0.82524},"canonical_model_id":"cohere/command-r-08-2024"},"cohere/command-r7b-12-2024":{"id":"cohere/command-r7b-12-2024","name":"Command R7B (12-2024)","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-12-13","last_updated":"2024-12-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.051578,"output":0.20631},"canonical_model_id":"cohere/command-r7b-12-2024"},"cohere/command-r-plus-08-2024":{"id":"cohere/command-r-plus-08-2024","name":"Command R+ (08-2024)","description":"Cohere's RAG workhorse for long-context enterprise search and tool use","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":3.4385,"output":13.754},"canonical_model_id":"cohere/command-r-plus-08-2024"},"perplexity/sonar-pro":{"id":"perplexity/sonar-pro","name":"Sonar Pro","description":"Deeper Sonar search model with broader retrieval and stronger synthesis","family":"sonar-pro","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":3.9,"output":19.5},"canonical_model_id":"perplexity/sonar-pro"},"perplexity/sonar":{"id":"perplexity/sonar","name":"Sonar","description":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","family":"sonar","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":100000},"cost":{"input":1.3,"output":1.3},"canonical_model_id":"perplexity/sonar"},"xiaomi/mimo-v2.5-pro":{"id":"xiaomi/mimo-v2.5-pro","name":"MiMo V2.5 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.598299,"output":1.196598,"cache_read":0.004951},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.192556,"output":0.385112,"cache_read":0.003851},"canonical_model_id":"xiaomi/mimo-v2.5"},"unbiased/pareto":{"id":"unbiased/pareto","name":"Pareto","description":"Blended multimodal model that runs multiple language models in parallel and returns a synthesized answer","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":3.4385,"output":10.3155,"cache_read":0.34385},"canonical_model_id":"unbiased/pareto"},"x-ai/grok-4-fast-non-reasoning":{"id":"x-ai/grok-4-fast-non-reasoning","name":"Grok 4 Fast Non-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":1999000},"cost":{"input":1.625,"output":3.25,"cache_read":0.26},"canonical_model_id":"xai/grok-4-fast-non-reasoning"},"x-ai/grok-4-20-0309-non-reasoning":{"id":"x-ai/grok-4-20-0309-non-reasoning","name":"Grok 4.20 Non-Reasoning","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":1999000},"cost":{"input":1.625,"output":3.25,"cache_read":0.26},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":96000},"cost":{"input":0.06877,"output":0.20631,"cache_read":0.055016},"canonical_model_id":"google/gemma-3-12b-it"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":117964},"cost":{"input":0.20631,"output":0.632684,"cache_read":0.103155},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048756,"output":65536},"cost":{"input":2.7508,"output":16.5048,"cache_read":0.27508},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.975,"output":4.875,"cache_read":0.0975},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemma-2-27b-it":{"id":"google/gemma-2-27b-it","name":"Gemma 2 27B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-07-13","last_updated":"2024-07-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":2048},"cost":{"input":0.89401,"output":0.89401},"canonical_model_id":"google/gemma-2-27b-it"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":2.6,"output":15.6,"cache_read":0.65},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.975,"output":4.875,"cache_read":0.0975},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"cost":{"input":0.20631,"output":0.82524,"cache_read":0.20631},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-flash-latest":{"id":"google/gemini-flash-latest","name":"Gemini Flash Latest","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2.0631,"output":12.3786,"cache_read":0.20631},"canonical_model_id":"google/gemini-flash-latest"},"google/gemma-3-4b-it":{"id":"google/gemma-3-4b-it","name":"Gemma 3 4B","description":"Open multimodal Gemma instruction model for efficient text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":32768},"cost":{"input":0.06877,"output":0.13754,"cache_read":0.055016},"canonical_model_id":"google/gemma-3-4b-it"},"google/gemini-3-8-flash":{"id":"google/gemini-3-8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.975,"output":4.875,"cache_read":0.0975},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B Instruct","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":1.361646,"output":2.049346,"cache_read":1.361646},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.95,"output":9.75,"cache_read":0.195},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.1-flash-lite-preview":{"id":"google/gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.34385,"output":2.0631,"cache_read":0.034385},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":1.147084,"output":3.439875,"cache_read":0.057767},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy-mt2-1.8b":{"id":"tencent/hy-mt2-1.8b","name":"Hy-MT2-1.8B","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.060518,"output":0.243446},"canonical_model_id":"tencent/hy-mt2-1.8b"},"tencent/hy-mt2-7b":{"id":"tencent/hy-mt2-7b","name":"Hy-MT2-7B","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.10178,"output":0.405743},"canonical_model_id":"tencent/hy-mt2-7b"},"tencent/hy-mt2-30b-a3b":{"id":"tencent/hy-mt2-30b-a3b","name":"Hy-MT2-30B-A3B","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.10178,"output":0.405743},"canonical_model_id":"tencent/hy-mt2-30b-a3b"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.27508,"output":1.10032,"cache_read":0.06877},"canonical_model_id":"tencent/hy3"},"sakana/fugu-max":{"id":"sakana/fugu-max","name":"Fugu Max","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.7508,"output":8.2524,"cache_read":0.34385},"canonical_model_id":"sakana/fugu-max"},"sakana/fugu-ultra-v2":{"id":"sakana/fugu-ultra-v2","name":"Fugu Ultra v2","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-08-28","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":6.877,"output":41.262,"cache_read":0.6877},"canonical_model_id":"sakana/fugu-ultra-v2"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":6.877,"output":41.262,"cache_read":0.6877},"canonical_model_id":"sakana/fugu-ultra"},"cognitivecomputations/dolphin-mistral-24b-venice-edition":{"id":"cognitivecomputations/dolphin-mistral-24b-venice-edition","name":"Venice Uncensored","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04-30","release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.27508,"output":1.23786},"canonical_model_id":"cognitivecomputations/dolphin-mistral-24b-venice-edition"},"rekaai/reka-edge":{"id":"rekaai/reka-edge","name":"Reka Edge","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"reka","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":0.13754,"output":0.13754},"canonical_model_id":"rekaai/reka-edge"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o Mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.195,"output":0.78,"cache_read":0.0975},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4":{"id":"openai/gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":8192},"cost":{"input":39,"output":78,"cache_read":39},"canonical_model_id":"openai/gpt-4"},"openai/gpt-5.6-luna-pro":{"id":"openai/gpt-5.6-luna-pro","name":"GPT-5.6 Luna Pro","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.27508,"output":1.65048,"cache_read":0.027508},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5-4":{"id":"openai/gpt-5-4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":3.25,"output":19.5,"cache_read":0.325},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 Nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.13,"output":0.52,"cache_read":0.0325},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 Mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.52,"output":2.08,"cache_read":0.13},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5-1":{"id":"openai/gpt-5-1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.625,"output":13,"cache_read":0.1625},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":41.262,"output":247.572},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":20.631,"output":165.048},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.13,"output":0.65,"cache_read":0.013},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.625,"output":13,"cache_read":0.1625},"canonical_model_id":"openai/gpt-5"},"openai/o3-pro":{"id":"openai/o3-pro","name":"o3 Pro","description":"High-effort o3 tier for difficult technical reasoning and careful answers","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":27.508,"output":110.032},"canonical_model_id":"openai/o3-pro"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.6,"output":15.6,"cache_read":0.26},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3 Mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.43,"output":5.72,"cache_read":0.715},"canonical_model_id":"openai/o3-mini"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":4096},"cost":{"input":0.65,"output":1.95},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2.6,"output":10.4,"cache_read":0.65},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.325,"output":2.6,"cache_read":0.0325},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-4o-2024-08-06":{"id":"openai/gpt-4o-2024-08-06","name":"GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":3.25,"output":13,"cache_read":1.625},"canonical_model_id":"openai/gpt-4o-2024-08-06"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.26,"output":1.56,"cache_read":0.026},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-oss-safeguard-20b":{"id":"openai/gpt-oss-safeguard-20b","name":"gpt-oss-safeguard-20b","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.103155,"output":0.41262,"cache_read":0.051578},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"openai/gpt-5-2":{"id":"openai/gpt-5-2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":2.275,"output":18.2,"cache_read":0.2275},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":3.25,"output":13,"cache_read":1.625},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":117964},"cost":{"input":0.103155,"output":0.41262,"cache_read":0.051578},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-5-5":{"id":"openai/gpt-5-5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":6.5,"output":39,"cache_read":0.65},"canonical_model_id":"openai/gpt-5.5"},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":100000},"cost":{"input":19.5,"output":78,"cache_read":9.75},"canonical_model_id":"openai/o1"},"openai/gpt-5.6-terra-pro":{"id":"openai/gpt-5.6-terra-pro","name":"GPT-5.6 Terra Pro","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.7508,"output":16.5048,"cache_read":0.27508},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2 Pro","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":28.8834,"output":231.0672},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":13,"output":65,"cache_read":1.3},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":13,"output":39},"canonical_model_id":"openai/gpt-4-turbo"},"openai/gpt-4o-2024-05-13":{"id":"openai/gpt-4o-2024-05-13","name":"GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-05-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":6.5,"output":19.5},"canonical_model_id":"openai/gpt-4o-2024-05-13"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5.2,"output":26,"cache_read":0.52},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.6-sol-pro":{"id":"openai/gpt-5.6-sol-pro","name":"GPT-5.6 Sol Pro","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5.5016,"output":27.508,"cache_read":0.55016},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.6,"output":13,"cache_read":0.13},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.27508,"output":1.71925,"cache_read":0.027508},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":117964},"cost":{"input":0.48139,"output":1.30663,"cache_read":0.48139},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.6,"output":13,"cache_read":0.26},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1-codex-max":{"id":"openai/gpt-5.1-codex-max","name":"GPT-5.1-Codex-Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":3.4385,"output":27.508,"cache_read":0.34385},"canonical_model_id":"openai/gpt-5.1-codex-max"},"openai/o1-pro":{"id":"openai/o1-pro","name":"o1-pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2025-03-19","last_updated":"2025-03-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":206.31,"output":825.24},"canonical_model_id":"openai/o1-pro"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":2.0631,"output":12.3786,"cache_read":0.20631},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":41.262,"output":247.572},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.065,"output":0.52,"cache_read":0.0065},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-4o-2024-11-20":{"id":"openai/gpt-4o-2024-11-20","name":"GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":3.25,"output":13,"cache_read":1.625},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"openai/gpt-5.4-image-2":{"id":"openai/gpt-5.4-image-2","type":"image","name":"GPT-5.4 Image 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":true,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["image","text","pdf"],"output":["image","text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":11.0032,"output":20.631,"cache_read":2.7508},"canonical_model_id":"openai/gpt-5.4-image-2"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":4.8139,"output":38.5112,"cache_read":0.48139},"canonical_model_id":"openai/gpt-5.3-codex"},"morph/morph-v3-fast":{"id":"morph/morph-v3-fast","name":"V3 Fast","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-07","last_updated":"2025-07-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":81920,"output":38000},"cost":{"input":1.10032,"output":1.65048},"canonical_model_id":"morph/morph-v3-fast"},"morph/morph-v3-large":{"id":"morph/morph-v3-large","name":"V3 Large","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-07","last_updated":"2025-07-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":1.23786,"output":2.61326},"canonical_model_id":"morph/morph-v3-large"},"alibaba/qwen-plus":{"id":"alibaba/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":16384},"cost":{"input":0.52,"output":1.56,"cache_read":0.104},"canonical_model_id":"alibaba/qwen-plus"},"alibaba/qwen-max":{"id":"alibaba/qwen-max","name":"Qwen Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":8192},"cost":{"input":2.08,"output":8.32,"cache_read":0.416},"canonical_model_id":"alibaba/qwen-max"},"alibaba/qwen3.8-flash":{"id":"alibaba/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.208,"output":0.611,"cache_read":0.0208},"canonical_model_id":"alibaba/qwen3.8-flash"},"alibaba/qwen3.8-max-prime":{"id":"alibaba/qwen3.8-max-prime","name":"Qwen3.8 Max Prime","description":"High-throughput edition of Qwen3.8 Max for coding, professional work, multimodal understanding, and long-running agent workflows","family":"qwen3.8-max","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":5.5016,"output":16.5048,"cache_read":0.6877},"canonical_model_id":"alibaba/qwen3.8-max-prime"},"alibaba/qwen3.8-2.4t-a95b":{"id":"alibaba/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":52429},"cost":{"input":3.4385,"output":8.59625,"cache_read":0.6877},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"alibaba/qwen3-next-80b-a3b-instruct":{"id":"alibaba/qwen3-next-80b-a3b-instruct","name":"Qwen3 Next 80B A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":129024,"output":16384},"cost":{"input":0.20631,"output":1.65048},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"alibaba/qwen3-coder-480b-a35b-instruct":{"id":"alibaba/qwen3-coder-480b-a35b-instruct","name":"Qwen3 Coder","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":65536},"cost":{"input":1.3754,"output":6.705075,"cache_read":1.3754},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"alibaba/qwen3.8-27b":{"id":"alibaba/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.61893,"output":4.40128},"canonical_model_id":"alibaba/qwen3.8-27b"},"alibaba/qwen3.8-max":{"id":"alibaba/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.6,"output":7.8,"cache_read":0.325},"canonical_model_id":"alibaba/qwen3.8-max"}}},"anthropic":{"id":"anthropic","env":["ANTHROPIC_API_KEY"],"npm":"@ai-sdk/anthropic","name":"Anthropic","doc":"https://docs.anthropic.com/en/docs/about-claude/models","models":{"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-07","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5}},"claude-opus-4-5-20251101":{"id":"claude-opus-4-5-20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-04","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"claude-sonnet-4-5-20250929":{"id":"claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-14","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25}},"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":8,"output":40,"cache_read":0.4,"cache_write":10},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-29","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5}},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"}}},"bee":{"id":"bee","env":["BEE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.bee.heossi.com/bee","name":"Bee by HEOSSI","doc":"https://bee.heossi.com/docs/sdks","models":{"bee-brood":{"id":"bee-brood","name":"Bee Brood 1.0","description":"Production multimodal reasoning for image and video analysis, tool-enabled workflows, and multi-step planning at controlled cost.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-10","last_updated":"2026-09-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":16384},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.72}},"bee-buzz":{"id":"bee-buzz","name":"Bee Buzz 1.0","description":"Mid-tier production model for tool-using agents, multi-step workflows, and builder/integration work. Bigger context than Comb, deeper planning, faster than Hive on routine tasks.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-10","last_updated":"2026-09-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":24576},"cost":{"input":2.5,"output":10,"cache_read":0.25,"cache_write":3}},"bee-swarm":{"id":"bee-swarm","name":"Bee Swarm 1.0","description":"Premium reasoning, multi-model cloud execution, large-document analysis, and high-value technical work. Includes the evolution engine and stronger eval gates by default.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":false,"structured_output":false,"release_date":"2026-07-10","last_updated":"2026-09-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":25,"output":150,"cache_read":2.5,"cache_write":30}},"bee-cell":{"id":"bee-cell","name":"Bee Cell 1.0","description":"Fast hosted AI for chat, coding help, document question answering, image understanding, and lightweight technical reasoning.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-10","last_updated":"2026-09-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.36}},"bee-hive":{"id":"bee-hive","name":"Bee Hive 1.0","description":"Scalable Bee serving with quantum-reasoning routing and multi-base domain-adapter routing for shared knowledge workflows, document intelligence, and lower-cost cloud inference. Domain adapters activate per customer domain after evaluation.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-10","last_updated":"2026-09-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":5,"output":20,"cache_read":0.5,"cache_write":6}},"bee-comb":{"id":"bee-comb","name":"Bee Comb 2.0","description":"Stronger local/workstation AI for coding, architecture, security reasoning, larger RAG, and domain-specific workflows. The default workhorse for engineers and dev teams.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-10","last_updated":"2026-09-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":16384},"cost":{"input":1.25,"output":5,"cache_read":0.125,"cache_write":1.5}}}},"longcat":{"id":"longcat","env":["LONGCAT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.longcat.chat/openai","name":"LongCat","doc":"https://longcat.chat/platform/docs/","models":{"LongCat-2.0":{"id":"LongCat-2.0","name":"LongCat-2.0","description":"Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.75,"output":2.95,"cache_read":0.015},"canonical_model_id":"meituan/longcat-2.0"}}},"ambient":{"id":"ambient","env":["AMBIENT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.ambient.xyz/v1","name":"Ambient","doc":"https://ambient.xyz","models":{"ambient/large":{"id":"ambient/large","name":"Ambient Large","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":202752},"cost":{"input":0.6,"output":2,"cache_read":0.15,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"ambient/small":{"id":"ambient/small","name":"Ambient Small","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":101072,"output":101072},"cost":{"input":0.001,"output":0.001,"cache_read":0,"cache_write":0}},"unsloth/qwen3.8-27b-gguf":{"id":"unsloth/qwen3.8-27b-gguf","name":"qwen3.8-27b-4q","description":"Tool-capable chat model for instruction following and agentic application workflows","family":"unsloth","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":101072,"output":101072},"cost":{"input":0.001,"output":0.001,"cache_read":0,"cache_write":0}},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"status":"deprecated","cost":{"input":0.08,"output":0.18,"cache_read":0.016,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.028,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":202752},"cost":{"input":0.6,"output":2,"cache_read":0.15,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"qwen/qwen3.6-35b-a3b":{"id":"qwen/qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.15,"output":1,"cache_read":0.15,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3-30b-a3b-instruct-2507":{"id":"qwen/qwen3-30b-a3b-instruct-2507","name":"Qwen3 30B A3B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-09","last_updated":"2026-10-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.09,"output":0.3,"cache_read":0.09,"cache_write":0}},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0.32,"output":3.2,"cache_read":0.16,"cache_write":0},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0.32,"output":3.2,"cache_read":0.16,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-27b"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.95,"output":4,"cache_read":0.2,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.69,"output":3.49,"cache_read":0.14,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"stepfun/step-3.7-flash":{"id":"stepfun/step-3.7-flash","name":"Step 3.7 Flash","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.19,"output":1.14,"cache_read":0.03,"cache_write":0},"canonical_model_id":"stepfun/step-3.7-flash"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.23,"output":0.96,"cache_read":0.05,"cache_write":0},"canonical_model_id":"minimax/MiniMax-M3"},"zai-org/GLM-5.1-FP8":{"id":"zai-org/GLM-5.1-FP8","name":"GLM 5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-5.2-FP8":{"id":"zai-org/GLM-5.2-FP8","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":202752},"status":"deprecated","cost":{"input":1.2,"output":4.2,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo-V2.5","description":"MiMo omni model for text, image, video, audio, and agents","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.4,"output":2,"cache_read":0.08,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.5"}}},"qiniu-ai":{"id":"qiniu-ai","env":["QINIU_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.qnaigc.com/v1","name":"Qiniu","doc":"https://developer.qiniu.com/aitokenapi","models":{"qwen3-235b-a22b-thinking-2507":{"id":"qwen3-235b-a22b-thinking-2507","name":"Qwen3 235B A22B Thinking 2507","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-12","last_updated":"2025-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":4096}},"claude-3.5-haiku":{"id":"claude-3.5-haiku","name":"Claude 3.5 Haiku","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192}},"claude-4.5-haiku":{"id":"claude-4.5-haiku","name":"Claude 4.5 Haiku","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-10-16","last_updated":"2025-10-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000}},"qwen3-30b-a3b-thinking-2507":{"id":"qwen3-30b-a3b-thinking-2507","name":"Qwen3 30b A3b Thinking 2507","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-04","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":126000,"output":32000}},"doubao-1.5-thinking-pro":{"id":"doubao-1.5-thinking-pro","name":"Doubao 1.5 Thinking Pro","description":"Flagship model for demanding analysis, coding, and production agent workflows","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000}},"doubao-seed-1.6":{"id":"doubao-seed-1.6","name":"Doubao-Seed 1.6","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-15","last_updated":"2025-08-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000}},"gemini-3.0-pro-image-preview":{"id":"gemini-3.0-pro-image-preview","name":"Gemini 3.0 Pro Image Preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":8192}},"gemini-3.0-flash-preview":{"id":"gemini-3.0-flash-preview","name":"Gemini 3.0 Flash Preview","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000}},"claude-3.5-sonnet":{"id":"claude-3.5-sonnet","name":"Claude 3.5 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-09","last_updated":"2025-09-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8200}},"claude-4.0-sonnet":{"id":"claude-4.0-sonnet","name":"Claude 4.0 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000}},"deepseek-r1-0528":{"id":"deepseek-r1-0528","name":"DeepSeek-R1-0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000}},"doubao-seed-1.6-flash":{"id":"doubao-seed-1.6-flash","name":"Doubao-Seed 1.6 Flash","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-15","last_updated":"2025-08-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000}},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"Gemini 2.5 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":64000}},"qwen3-235b-a22b-instruct-2507":{"id":"qwen3-235b-a22b-instruct-2507","name":"Qwen3 235b A22B Instruct 2507","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-12","last_updated":"2025-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":64000}},"gemini-3.0-pro-preview":{"id":"gemini-3.0-pro-preview","name":"Gemini 3.0 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000}},"qwen3-235b-a22b":{"id":"qwen3-235b-a22b","name":"Qwen 3 235B A22B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000}},"deepseek-v3":{"id":"deepseek-v3","name":"DeepSeek-V3","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-08-13","last_updated":"2025-08-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000}},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536}},"kimi-k2":{"id":"kimi-k2","name":"Kimi K2","description":"Kimi model for long-context chat, coding, and agentic reasoning","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000}},"claude-4.5-opus":{"id":"claude-4.5-opus","name":"Claude 4.5 Opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":200000}},"qwen3-30b-a3b":{"id":"qwen3-30b-a3b","name":"Qwen3 30B A3B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":40000,"output":4096}},"deepseek-r1":{"id":"deepseek-r1","name":"DeepSeek-R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000}},"qwen3-vl-30b-a3b-thinking":{"id":"qwen3-vl-30b-a3b-thinking","name":"Qwen3-Vl 30b A3b Thinking","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-09","last_updated":"2026-02-09","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000}},"qwen-turbo":{"id":"qwen-turbo","name":"Qwen-Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":4096}},"glm-4.5-air":{"id":"glm-4.5-air","name":"GLM 4.5 Air","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":4096}},"qwen3-next-80b-a3b-instruct":{"id":"qwen3-next-80b-a3b-instruct","name":"Qwen3 Next 80B A3B Instruct","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-12","last_updated":"2025-09-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768}},"deepseek-v3-0324":{"id":"deepseek-v3-0324","name":"DeepSeek-V3-0324","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000}},"qwen3-next-80b-a3b-thinking":{"id":"qwen3-next-80b-a3b-thinking","name":"Qwen3 Next 80B A3B Thinking","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-12","last_updated":"2025-09-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768}},"deepseek-v3.1":{"id":"deepseek-v3.1","name":"DeepSeek-V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-19","last_updated":"2025-08-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000}},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"gpt-oss-20b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-06","last_updated":"2025-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096}},"doubao-seed-1.6-thinking":{"id":"doubao-seed-1.6-thinking","name":"Doubao-Seed 1.6 Thinking","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-15","last_updated":"2025-08-15","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000}},"gemini-2.0-flash":{"id":"gemini-2.0-flash","name":"Gemini 2.0 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":8192}},"qwen3-coder-480b-a35b-instruct":{"id":"qwen3-coder-480b-a35b-instruct","name":"Qwen3 Coder 480B A35B Instruct","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-14","last_updated":"2025-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":4096}},"qwen-vl-max-2025-01-25":{"id":"qwen-vl-max-2025-01-25","name":"Qwen VL-MAX-2025-01-25","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096}},"MiniMax-M1":{"id":"MiniMax-M1","name":"MiniMax M1","description":"MiniMax model for chat, coding, office work, and agentic tasks","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":80000}},"qwen3-30b-a3b-instruct-2507":{"id":"qwen3-30b-a3b-instruct-2507","name":"Qwen3 30b A3b Instruct 2507","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-04","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000}},"kling-v2-6":{"id":"kling-v2-6","name":"Kling-V2 6","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-01-13","last_updated":"2026-01-13","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":99999999,"output":99999999}},"qwen3-32b":{"id":"qwen3-32b","name":"Qwen3 32B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":40000,"output":4096}},"doubao-seed-2.0-mini":{"id":"doubao-seed-2.0-mini","name":"Doubao Seed 2.0 Mini","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000}},"mimo-v2-flash":{"id":"mimo-v2-flash","name":"Mimo-V2-Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12-01","release_date":"2025-12-16","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.1,"output":0.3,"cache_read":0.01},"canonical_model_id":"xiaomi/mimo-v2-flash"},"claude-4.5-sonnet":{"id":"claude-4.5-sonnet","name":"Claude 4.5 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000}},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5 397B A17B","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-22","last_updated":"2026-02-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000}},"claude-4.0-opus":{"id":"claude-4.0-opus","name":"Claude 4.0 Opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000}},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"gpt-oss-120b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-06","last_updated":"2025-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096}},"qwen3-max-preview":{"id":"qwen3-max-preview","name":"Qwen3 Max Preview","description":"Flagship model for demanding analysis, coding, and production agent workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-06","last_updated":"2025-09-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000}},"doubao-seed-2.0-code":{"id":"doubao-seed-2.0-code","name":"Doubao Seed 2.0 Code","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000}},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":64000}},"gemini-2.0-flash-lite":{"id":"gemini-2.0-flash-lite","name":"Gemini 2.0 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":8192}},"doubao-1.5-vision-pro":{"id":"doubao-1.5-vision-pro","name":"Doubao 1.5 Vision Pro","description":"Flagship model for demanding analysis, coding, and production agent workflows","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000}},"doubao-seed-2.0-lite":{"id":"doubao-seed-2.0-lite","name":"Doubao Seed 2.0 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000}},"doubao-seed-2.0-pro":{"id":"doubao-seed-2.0-pro","name":"Doubao Seed 2.0 Pro","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000}},"qwen2.5-vl-72b-instruct":{"id":"qwen2.5-vl-72b-instruct","name":"Qwen 2.5 VL 72B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192}},"glm-4.5":{"id":"glm-4.5","name":"GLM 4.5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":98304}},"qwen2.5-vl-7b-instruct":{"id":"qwen2.5-vl-7b-instruct","name":"Qwen 2.5 VL 7B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192}},"claude-3.7-sonnet":{"id":"claude-3.7-sonnet","name":"Claude 3.7 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000}},"qwen-max-2025-01-25":{"id":"qwen-max-2025-01-25","name":"Qwen2.5-Max-2025-01-25","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096}},"qwen3-max":{"id":"qwen3-max","name":"Qwen3 Max","description":"Flagship model for demanding analysis, coding, and production agent workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-24","last_updated":"2025-09-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536}},"gemini-2.5-flash-image":{"id":"gemini-2.5-flash-image","name":"Gemini 2.5 Flash Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-10-22","last_updated":"2025-10-22","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":32768,"output":8192}},"doubao-1.5-pro-32k":{"id":"doubao-1.5-pro-32k","name":"Doubao 1.5 Pro 32k","description":"Flagship model for demanding analysis, coding, and production agent workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":12000}},"claude-4.1-opus":{"id":"claude-4.1-opus","name":"Claude 4.1 Opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-06","last_updated":"2025-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000}},"meituan/longcat-flash-chat":{"id":"meituan/longcat-flash-chat","name":"Meituan/Longcat-Flash-Chat","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-11-05","last_updated":"2025-11-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072}},"meituan/longcat-flash-lite":{"id":"meituan/longcat-flash-lite","name":"Meituan/Longcat-Flash-Lite","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-06","last_updated":"2026-02-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":320000}},"deepseek/deepseek-v3.1-terminus-thinking":{"id":"deepseek/deepseek-v3.1-terminus-thinking","name":"DeepSeek/DeepSeek-V3.1-Terminus-Thinking","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-09-22","last_updated":"2025-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000}},"deepseek/deepseek-v3.2-251201":{"id":"deepseek/deepseek-v3.2-251201","name":"Deepseek/DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000}},"deepseek/deepseek-v3.2-exp":{"id":"deepseek/deepseek-v3.2-exp","name":"DeepSeek/DeepSeek-V3.2-Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000}},"deepseek/deepseek-v3.1-terminus":{"id":"deepseek/deepseek-v3.1-terminus","name":"DeepSeek/DeepSeek-V3.1-Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-22","last_updated":"2025-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000}},"deepseek/deepseek-math-v2":{"id":"deepseek/deepseek-math-v2","name":"Deepseek/Deepseek-Math-V2","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-04","last_updated":"2025-12-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":160000,"output":160000}},"deepseek/deepseek-v3.2-exp-thinking":{"id":"deepseek/deepseek-v3.2-exp-thinking","name":"DeepSeek/DeepSeek-V3.2-Exp-Thinking","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000}},"z-ai/autoglm-phone-9b":{"id":"z-ai/autoglm-phone-9b","name":"Z-Ai/Autoglm Phone 9b","description":"GLM vision model for visual reasoning, documents, and multimodal agents","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":12800,"output":4096}},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"Z-Ai/GLM 4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":200000}},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"Z-Ai/GLM 5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000}},"z-ai/glm-4.6":{"id":"z-ai/glm-4.6","name":"Z-AI/GLM 4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-10-11","last_updated":"2025-10-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":200000}},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 0905","description":"Kimi model for long-context chat, coding, and agentic reasoning","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-08","last_updated":"2025-09-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":100000}},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Moonshotai/Kimi-K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-01-28","last_updated":"2026-01-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000}},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-11-07","last_updated":"2025-11-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":100000}},"stepfun-ai/gelab-zero-4b-preview":{"id":"stepfun-ai/gelab-zero-4b-preview","name":"Stepfun-Ai/Gelab Zero 4b Preview","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":4096}},"stepfun/step-3.5-flash":{"id":"stepfun/step-3.5-flash","name":"Stepfun/Step-3.5 Flash","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-02","last_updated":"2026-02-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":64000,"output":4096}},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"Minimax/Minimax-M2","description":"MiniMax model for chat, coding, office work, and agentic tasks","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-10-28","last_updated":"2025-10-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000}},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"Minimax/Minimax-M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":204800,"output":128000}},"minimax/minimax-m2.5-highspeed":{"id":"minimax/minimax-m2.5-highspeed","name":"Minimax/Minimax-M2.5 Highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":204800,"output":128000}},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"Minimax/Minimax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":204800,"output":128000}},"xiaomi/mimo-v2-flash":{"id":"xiaomi/mimo-v2-flash","name":"Xiaomi/Mimo-V2-Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12-01","release_date":"2025-12-16","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.1,"output":0.3,"cache_read":0.01},"canonical_model_id":"xiaomi/mimo-v2-flash"},"x-ai/grok-code-fast-1":{"id":"x-ai/grok-code-fast-1","name":"x-AI/Grok-Code-Fast 1","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":10000}},"x-ai/grok-4.1-fast-non-reasoning":{"id":"x-ai/grok-4.1-fast-non-reasoning","name":"X-Ai/Grok 4.1 Fast Non Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-19","last_updated":"2025-12-19","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000}},"x-ai/grok-4-fast-non-reasoning":{"id":"x-ai/grok-4-fast-non-reasoning","name":"X-Ai/Grok-4-Fast-Non-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000}},"x-ai/grok-4.1-fast-reasoning":{"id":"x-ai/grok-4.1-fast-reasoning","name":"X-Ai/Grok 4.1 Fast Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-19","last_updated":"2025-12-19","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":20000000,"output":2000000}},"x-ai/grok-4-fast-reasoning":{"id":"x-ai/grok-4-fast-reasoning","name":"X-Ai/Grok-4-Fast-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000}},"x-ai/grok-4.1-fast":{"id":"x-ai/grok-4.1-fast","name":"x-AI/Grok-4.1-Fast","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000}},"x-ai/grok-4-fast":{"id":"x-ai/grok-4-fast","name":"x-AI/Grok-4-Fast","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-20","last_updated":"2025-09-20","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000}},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"OpenAI/GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000}},"openai/gpt-5":{"id":"openai/gpt-5","name":"OpenAI/GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-19","last_updated":"2025-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000}}}},"xai":{"id":"xai","env":["XAI_API_KEY"],"npm":"@ai-sdk/xai","name":"xAI","doc":"https://docs.x.ai/docs/models","models":{"grok-imagine-video":{"id":"grok-imagine-video","type":"video","name":"Grok Imagine Video","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-28","last_updated":"2026-01-28","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":1024,"output":0}},"grok-imagine-image-quality":{"id":"grok-imagine-image-quality","type":"image","name":"Grok Imagine Image Quality","description":"Higher-fidelity Grok Imagine image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-03","last_updated":"2026-04-03","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":16000,"output":0},"status":"deprecated","canonical_model_id":"xai/grok-imagine-image-quality"},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"xAI's Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"grok-4.20-multi-agent-0309":{"id":"grok-4.20-multi-agent-0309","name":"Grok 4.20 Multi-Agent","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"grok-imagine-video-1.5":{"id":"grok-imagine-video-1.5","type":"video","name":"Grok Imagine Video 1.5","description":"Video model for text-to-video, image-to-video, and reference-to-video generation with optional voice references","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-05-30","last_updated":"2026-05-30","modalities":{"input":["text","image","audio"],"output":["video"]},"open_weights":false,"limit":{"context":1024,"output":0},"canonical_model_id":"xai/grok-imagine-video-1.5"},"grok-4.20-0309-reasoning":{"id":"grok-4.20-0309-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"grok-build-0.1":{"id":"grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}}},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"grok-imagine-image":{"id":"grok-imagine-image","type":"image","name":"Grok Imagine Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-28","last_updated":"2026-01-28","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":16000,"output":0},"canonical_model_id":"xai/grok-imagine-image"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"grok-4.20-0309-non-reasoning":{"id":"grok-4.20-0309-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"grok-imagine-video-1.5-lite":{"id":"grok-imagine-video-1.5-lite","type":"video","name":"Grok Imagine Video 1.5 Lite","description":"Cost-efficient video model for text-to-video and image-to-video generation","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":1024,"output":0},"canonical_model_id":"xai/grok-imagine-video-1.5-lite"}}},"kimi-code-plan-global":{"id":"kimi-code-plan-global","env":["KIMI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.kimi.ai/coding/v1","name":"Kimi For Coding (kimi.ai)","doc":"https://www.kimi.ai/code/docs/en/kimi-code/models.html","models":{"k3":{"id":"k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k3"},"kimi-for-coding":{"id":"kimi-for-coding","name":"kimi-for-coding","description":"Kimi coding model with more efficient thinking and up to 1M context, available through Kimi Code","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.8-preview"},"k3-256k":{"id":"k3-256k","name":"Kimi K3-256K","description":"256K-context version of Kimi K3, reducing token consumption for shorter coding sessions","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k3"},"kimi-for-coding-highspeed":{"id":"kimi-for-coding-highspeed","name":"Kimi For Coding HighSpeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"}}},"anyapi":{"id":"anyapi","env":["ANYAPI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.anyapi.ai/v1","name":"AnyAPI","doc":"https://docs.anyapi.ai","models":{"anthropic/claude-haiku-4-5":{"id":"anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024,"max":127999}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-sonnet-4-5":{"id":"anthropic/claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"canonical_model_id":"anthropic/claude-opus-4-7"},"xai/grok-4.3":{"id":"xai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"canonical_model_id":"xai/grok-4.3"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-r1":{"id":"deepseek/deepseek-r1","name":"DeepSeek Reasoner","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"canonical_model_id":"deepseek/deepseek-reasoner"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-chat":{"id":"deepseek/deepseek-chat","name":"DeepSeek Chat","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"canonical_model_id":"deepseek/deepseek-chat"},"mistralai/devstral-2512":{"id":"mistralai/devstral-2512","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","canonical_model_id":"mistral/devstral-2512"},"mistralai/mistral-large-2512":{"id":"mistralai/mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"canonical_model_id":"mistral/mistral-large-2512"},"cohere/command-r-plus-08-2024":{"id":"cohere/command-r-plus-08-2024","name":"Command R+","description":"Cohere's RAG workhorse for long-context enterprise search and tool use","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"canonical_model_id":"cohere/command-r-plus-08-2024"},"perplexity/sonar-pro":{"id":"perplexity/sonar-pro","name":"Sonar Pro","description":"Deeper Sonar search model with broader retrieval and stronger synthesis","family":"sonar-pro","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"canonical_model_id":"perplexity/sonar-pro"},"perplexity/sonar-reasoning-pro":{"id":"perplexity/sonar-reasoning-pro","name":"Sonar Reasoning Pro","description":"Web-grounded Sonar for multi-step research questions that need cited reasoning","family":"sonar-reasoning","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"canonical_model_id":"perplexity/sonar-reasoning-pro"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-2.5-pro"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-3-pro-preview":{"id":"google/gemini-3-pro-preview","name":"Gemini 3 Pro Preview","description":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-3-pro-preview"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"canonical_model_id":"openai/o4-mini"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"canonical_model_id":"openai/gpt-5"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"canonical_model_id":"openai/o3-mini"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"canonical_model_id":"openai/gpt-5.4"},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"canonical_model_id":"openai/o3"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"canonical_model_id":"openai/gpt-5.1"}}},"frogbot":{"id":"frogbot","env":["FROGBOT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://app.frogbot.ai/api/v1","name":"FrogBot","doc":"https://docs.frogbot.ai","models":{"kimi-k2-6":{"id":"kimi-k2-6","name":"Kimi-K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"1970-01-01","last_updated":"1970-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.95,"output":4,"cache_read":0.16}},"grok-code-fast-1":{"id":"grok-code-fast-1","name":"Grok 4.1 Fast (Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2023-10","release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.2,"output":1.5,"cache_read":0.02}},"gemini-3-1-pro-preview":{"id":"gemini-3-1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-02-18","last_updated":"2026-02-18","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2,"output":12,"cache_read":0.2}},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25}},"gpt-5-4-nano":{"id":"gpt-5-4-nano","name":"GPT-5.4 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02}},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05}},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek v4 Pro","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":1.74,"output":3.48,"cache_read":0.14}},"gpt-5-3-codex":{"id":"gpt-5-3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"qwen-3-6-plus":{"id":"qwen-3-6-plus","name":"Qwen 3.6 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.5,"output":3,"cache_read":0.1}},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-03-20","last_updated":"2025-06-05","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.31}},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"minimax-m2-7":{"id":"minimax-m2-7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":192000,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.06}},"gpt-5-4-mini":{"id":"gpt-5-4-mini","name":"GPT-5.4 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075}},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25}},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"1970-01-01","last_updated":"1970-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.07,"output":0.2}},"gpt-5-5":{"id":"gpt-5-5","name":"GPT-5.5","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25}},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"minimax-m2-5":{"id":"minimax-m2-5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2025-01-15","last_updated":"2025-02-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":192000,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.03}},"zai-glm-5-1":{"id":"zai-glm-5-1","name":"Z.AI GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-01-20","last_updated":"2025-02-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":8192},"cost":{"input":1.4,"output":4.4,"cache_read":0.26}},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"1970-01-01","last_updated":"1970-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6}},"grok-4-1-fast-non-reasoning":{"id":"grok-4-1-fast-non-reasoning","name":"Grok 4.1 Fast (Non-Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":128000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-07-17","last_updated":"2025-07-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.075}},"grok-4-3":{"id":"grok-4-3","name":"Grok 4.3","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2}},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi-K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"1970-01-01","last_updated":"1970-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.6,"output":3,"cache_read":0.1}},"grok-4-1-fast-reasoning":{"id":"grok-4-1-fast-reasoning","name":"Grok 4.1 Fast (Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":128000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}}}},"opencode-go":{"id":"opencode-go","env":["OPENCODE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://opencode.ai/inference/go/openai/v1","name":"OpenCode Go","doc":"https://opencode.ai/v2/docs/console/go/","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"muse-spark-1.3-contributor":{"id":"muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for coding and agentic workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"hy4-preview":{"id":"hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":64000},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax-m2.7","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/go/anthropic/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375}},"longcat-2.0":{"id":"longcat-2.0","name":"LongCat-2.0","description":"Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"meituan/longcat-2.0"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro (New)","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.66,"output":1.98,"cache_read":0.022}},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo V2.5 Pro","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo-v2.5-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.435,"output":0.87,"cache_read":0.003625}},"cat-jam-free":{"id":"cat-jam-free","name":"Cat Jam Free","description":"A text-only reasoning model for coding and agentic workflows, with tool calling and structured JSON outputs.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"step-5-preview-free":{"id":"step-5-preview-free","name":"Step 5 Preview Free","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"stepfun/step-5-preview"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"longcat-2.5-preview-free":{"id":"longcat-2.5-preview-free","name":"LongCat 2.5 Preview Free","description":"Meituan's multimodal reasoning model for coding and agent tasks, with image understanding and a 1M-token context window","family":"longcat","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"meituan/longcat-2.5-preview"},"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo V2.5","description":"MiMo omni model for text, image, video, audio, and agents","family":"mimo-v2.5","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028}},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/go/anthropic/v1"},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"beam-free":{"id":"beam-free","name":"Beam Free","description":"Sparse mixture-of-experts reasoning model with 23B active parameters for coding and agentic workflows","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"hy3":{"id":"hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.14,"output":0.58,"cache_read":0.035},"canonical_model_id":"tencent/hy3"},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen3.7-plus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/go/anthropic/v1"},"cost":{"input":0.4,"output":1.6,"cache_read":0.04,"cache_write":0.5,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.12,"cache_write":1.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.12,"cache_write":1.5}}},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"muse-spark-1.2-contributor":{"id":"muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.1,"output":0.2,"cache_read":0.002},"canonical_model_id":"meta/muse-spark-1.2"},"gemini-4-argon-contributor":{"id":"gemini-4-argon-contributor","name":"Gemini 4 Argon Contributor","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/go/google/v1beta"},"cost":{"input":4,"output":20,"cache_read":0.2,"input_audio":4},"canonical_model_id":"google/gemini-4-argon"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":1,"output":2.7,"cache_read":0.1},"canonical_model_id":"stepfun/step-5-preview"},"space-bunny":{"id":"space-bunny","name":"Space Bunny","description":"Anonymous preview reasoning model for coding, agentic tasks, tool use, and multimodal input","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":524288,"output":524288},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.15,"output":0.6,"cache_read":0.03,"cache_write":0}},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax-m3","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-31","last_updated":"2026-05-31","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/go/anthropic/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"tiers":[{"input":0.6,"output":2.4,"cache_read":0.12,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":0.6,"output":2.4,"cache_read":0.12}}},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.435,"output":0.87,"cache_read":0.003625},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/go/anthropic/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter multimodal flagship for coding, professional work, and long-horizon agentic workflows","family":"qwen3.8-max","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/go/anthropic/v1"},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"minimax-m3.1-flash-free":{"id":"minimax-m3.1-flash-free","name":"MiniMax M3.1 Flash Free","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":512000},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"}}},"the-grid-ai":{"id":"the-grid-ai","env":["THEGRID_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.thegrid.ai/v1","name":"The Grid AI","doc":"https://thegrid.ai/docs","models":{"deepseek-pro-latest":{"id":"deepseek-pro-latest","name":"DeepSeek Pro Latest","description":"Lab latest market for DeepSeek Pro supply. You contract for the latest qualifying route, not a specific model.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":384000}},"gemini-pro-latest":{"id":"gemini-pro-latest","name":"Gemini Pro Latest","description":"Lab latest market for Gemini Pro frontier supply. You contract for the latest qualifying route, not a specific model.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":65536}},"gpt-sol-latest":{"id":"gpt-sol-latest","name":"GPT Sol Latest","description":"Lab latest market for GPT frontier supply. You contract for the latest qualifying route, not a specific model.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000}},"agent-max":{"id":"agent-max","name":"Agent Max","description":"Frontier models for autonomous research, deep multi-step tool chains, and complex long-horizon tasks. Any model that meets the contract spec can serve your request.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-05-04","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"status":"beta"},"claude-opus-latest":{"id":"claude-opus-latest","name":"Claude Opus Latest","description":"Lab latest market for Claude Opus frontier supply. You contract for the latest qualifying route, not a specific model.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-06","last_updated":"2026-07-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000}},"minimax-latest":{"id":"minimax-latest","name":"MiniMax Latest","description":"Lab latest market for MiniMax supply. You contract for the latest qualifying route, not a specific model.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":512000}},"agent-standard":{"id":"agent-standard","name":"Agent Standard","description":"Price-optimized models for fast tool calls, simple agent loops, high-throughput automation, and orchestration. Any model that meets the contract spec can serve your request.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-04","last_updated":"2026-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":120000,"output":65536},"status":"beta"},"text-max":{"id":"text-max","name":"Text Max","description":"Frontier models for deep reasoning, long context, and complex workflows. Any model that meets the contract spec can serve your request.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-02-26","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000}},"agent-prime":{"id":"agent-prime","name":"Agent Prime","description":"Reliable models for dependable agentic applications, multi-step tool use, and reasoning workflows. Any model that meets the contract spec can serve your request.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-04","last_updated":"2026-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":196608,"input":120000,"output":131072},"status":"beta"},"code-prime":{"id":"code-prime","name":"Code Prime","description":"Reliable models for everyday software tasks, code completion, review, and standard debugging. Any model that meets the contract spec can serve your request.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-04","last_updated":"2026-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":196608,"input":120000,"output":131072},"status":"beta"},"text-prime":{"id":"text-prime","name":"Text Prime","description":"Reliable models for everyday text generation, editing, and analysis across diverse workflows. Any model that meets the contract spec can serve your request.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-26","last_updated":"2026-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":196608,"input":120000,"output":131072}},"text-standard":{"id":"text-standard","name":"Text Standard","description":"Price-optimized models with low-latency, high-throughput and shorter maximum outputs. Any model that meets the contract spec can serve your request.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-26","last_updated":"2026-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":120000,"output":65536}},"code-standard":{"id":"code-standard","name":"Code Standard","description":"Price-optimized models for rapid autocomplete, linting, high-frequency suggestions, and batch edits. Any model that meets the contract spec can serve your request.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-04","last_updated":"2026-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":120000,"output":65536},"status":"beta"},"bytedance-pro-latest":{"id":"bytedance-pro-latest","name":"ByteDance Pro Latest","description":"Lab latest market for ByteDance Pro supply. You contract for the latest qualifying route, not a specific model.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":256000,"output":131072}},"glm-latest":{"id":"glm-latest","name":"GLM Latest","description":"Lab latest market for GLM supply. You contract for the latest qualifying route, not a specific model.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000}},"code-max":{"id":"code-max","name":"Code Max","description":"Frontier models for complex research, architectural decisions, debugging, and multi-file development. Any model that meets the contract spec can serve your request.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-05-04","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"status":"beta"}}},"nano-gpt":{"id":"nano-gpt","env":["NANO_GPT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://nano-gpt.com/api/v1","name":"NanoGPT","doc":"https://docs.nano-gpt.com","models":{"doubao-1.5-pro-256k":{"id":"doubao-1.5-pro-256k","name":"Doubao 1.5 Pro 256k","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":16384},"cost":{"input":0.799,"output":1.445,"cache_read":0.3995}},"doubao-seed-2-0-code-preview-260215":{"id":"doubao-seed-2-0-code-preview-260215","name":"Doubao Seed 2.0 Code Preview","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":128000},"cost":{"input":0.782,"output":3.893,"cache_read":0.391},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"command-a-reasoning-08-2025":{"id":"command-a-reasoning-08-2025","name":"Cohere Command A (08/2025)","description":"Cohere reasoning model for multilingual enterprise agents, tools, and complex workflows","family":"command-a","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":8192},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"cohere/command-a-reasoning-08-2025"},"auto-model":{"id":"auto-model","name":"Auto model","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2025-04-16","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":1000000},"cost":{"input":0,"output":0}},"Gemma-4-31B-GarnetV2":{"id":"Gemma-4-31B-GarnetV2","name":"Gemma 4 31B Garnet V2","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.306,"output":0.306,"cache_read":0.153}},"gemini-2.5-pro-preview-06-05":{"id":"gemini-2.5-pro-preview-06-05","name":"Gemini 2.5 Pro Preview 0605","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2.5,"output":10,"cache_read":0.25}},"celeris-1":{"id":"celeris-1","name":"Celeris 1","description":"Celeris 1 is a diffusion language model built for ultra-low-latency classification, extraction, judging, query rewriting, and other short structured responses.","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-07-25","last_updated":"2026-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"input":8192,"output":8192},"cost":{"input":2,"output":6,"cache_read":1}},"longcat-2.0:thinking":{"id":"longcat-2.0:thinking","name":"LongCat 2.0 Thinking","description":"Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":131072},"cost":{"input":0.75,"output":3,"cache_read":0.015},"canonical_model_id":"meituan/longcat-2.0"},"gemma-4-12b-it-station-keeper":{"id":"gemma-4-12b-it-station-keeper","name":"Gemma 4 12B StationKeeper","description":"Gemma 4 12B StationKeeper is an open-weight roleplay finetune with image understanding, tool calling, optional reasoning, and a 131,072-token context window.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-09","last_updated":"2026-09-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.05,"output":0.25,"cache_read":0.025}},"gemma-4-31b-it-novelist":{"id":"gemma-4-31b-it-novelist","name":"Novelist","description":"Novelist is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"gemma-4-31b-it-garnet":{"id":"gemma-4-31b-it-garnet","name":"Garnet","description":"Garnet is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"gemma-4-12b-it":{"id":"gemma-4-12b-it","name":"Gemma 4 12B Instruct","description":"Compact Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.05,"output":0.25,"cache_read":0.025},"canonical_model_id":"google/gemma-4-12b-it"},"glm-4-plus-0111":{"id":"glm-4-plus-0111","name":"GLM 4 Plus 0111","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":4096},"cost":{"input":9.996,"output":9.996,"cache_read":4.998}},"sarvam-105b":{"id":"sarvam-105b","name":"Sarvam 105B","description":"Flagship Indian-language reasoning model for enterprise multilingual applications","family":"sarvam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-18","last_updated":"2026-02-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":4096},"cost":{"input":0.054,"output":0.2124,"cache_read":0.0336},"canonical_model_id":"sarvam/sarvam-105b"},"venice-uncensored":{"id":"venice-uncensored","name":"Venice Uncensored","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"venice","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-10-01","last_updated":"2025-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.4,"output":1.8,"cache_read":0.4}},"gemini-2.0-pro-exp-02-05":{"id":"gemini-2.0-pro-exp-02-05","name":"Gemini 2.0 Pro 0205","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-02-05","last_updated":"2025-02-05","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":2097152,"input":2097152,"output":8192},"cost":{"input":1.989,"output":7.956,"cache_read":0.49725}},"glm-4.1v-thinking-flash":{"id":"glm-4.1v-thinking-flash","name":"GLM 4.1V Thinking Flash","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"glm-flash","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":64000,"input":64000,"output":8192},"cost":{"input":0.3,"output":0.3,"cache_read":0.15}},"longcat-2.0":{"id":"longcat-2.0","name":"LongCat 2.0","description":"Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window","family":"longcat","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":131072},"cost":{"input":0.75,"output":3,"cache_read":0.015},"canonical_model_id":"meituan/longcat-2.0"},"Gemma-4-31B-Queen":{"id":"Gemma-4-31B-Queen","name":"Gemma 4 31B Queen","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.306,"output":0.306,"cache_read":0.153}},"gemini-3-pro-image-preview":{"id":"gemini-3-pro-image-preview","type":"image","name":"Gemini 3 Pro Image","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"input":65536,"output":32768},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"google/gemini-3-pro-image-preview"},"mercury-2":{"id":"mercury-2","name":"Mercury 2","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":50000},"cost":{"input":0.25,"output":0.75,"cache_read":0.025},"canonical_model_id":"inception/mercury-2"},"gemma-4-26b-a4b-it-moonlight":{"id":"gemma-4-26b-a4b-it-moonlight","name":"Moonlight Dusk","description":"Moonlight Dusk is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"gemma-4-31b-it-scarlet":{"id":"gemma-4-31b-it-scarlet","name":"Gemma 4 31B Scarlet","description":"Scarlet is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-08","last_updated":"2026-10-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"glm-z1-airx":{"id":"glm-z1-airx","name":"GLM Z1 AirX","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"input":32000,"output":16384},"cost":{"input":0.7,"output":0.7,"cache_read":0.35}},"ernie-5.1":{"id":"ernie-5.1","name":"ERNIE 5.1","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"ernie","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-05-10","last_updated":"2026-05-10","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":119000,"input":119000,"output":64000},"cost":{"input":0.75,"output":3,"cache_read":0.75}},"Qwen3.5-27B-BlueStar-v3-Derestricted":{"id":"Qwen3.5-27B-BlueStar-v3-Derestricted","name":"Qwen3.5 27B BlueStar v3 Derestricted","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.306,"output":0.306,"cache_read":0.153}},"gemini-2.5-flash-preview-09-2025":{"id":"gemini-2.5-flash-preview-09-2025","name":"Gemini 2.5 Flash Preview (09/2025)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03}},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"Gemini 2.5 Flash Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"google/gemini-2.5-flash-lite"},"gemma-4-31b-it-isometry":{"id":"gemma-4-31b-it-isometry","name":"Isometry","description":"Isometry is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"phi-4-multimodal-instruct":{"id":"phi-4-multimodal-instruct","name":"Phi 4 Multimodal","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"phi","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.07,"output":0.11,"cache_read":0.035}},"mistral-code-agent-latest":{"id":"mistral-code-agent-latest","name":"Mistral Code Agent Latest","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"mistral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.4,"output":2,"cache_read":0.2}},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"cache_write":0.375},"canonical_model_id":"google/gemini-2.5-pro"},"longcat-2.5-preview":{"id":"longcat-2.5-preview","name":"LongCat 2.5 Preview","description":"Meituan's multimodal reasoning model for coding and agent tasks, with image understanding and a 1M-token context window","family":"longcat","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":131072},"cost":{"input":0.75,"output":2.95,"cache_read":0.015},"canonical_model_id":"meituan/longcat-2.5-preview"},"gemini-2.5-flash-nothinking":{"id":"gemini-2.5-flash-nothinking","name":"Gemini 2.5 Flash (No Thinking)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03}},"qwen3-vl-235b-a22b-instruct-original":{"id":"qwen3-vl-235b-a22b-instruct-original","name":"Qwen3 VL 235B A22B Instruct Original","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":0.3,"output":1.9,"cache_read":0.15}},"gemma-4-26b-a4b-it-opusdistill":{"id":"gemma-4-26b-a4b-it-opusdistill","name":"Opus Distill","description":"Opus Distill is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"glm-4-long":{"id":"glm-4-long","name":"GLM-4 Long","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-08-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":4095},"cost":{"input":0.2006,"output":0.2006,"cache_read":0.1003}},"Meta-Llama-3-1-8B-Instruct-FP8":{"id":"Meta-Llama-3-1-8B-Instruct-FP8","name":"Llama 3.1 8B (decentralized)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.02,"output":0.03,"cache_read":0.01}},"Gemma-4-31B-DarkIdol":{"id":"Gemma-4-31B-DarkIdol","name":"Gemma 4 31B DarkIdol","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.306,"output":0.306,"cache_read":0.153}},"doubao-seed-2-0-mini-260215":{"id":"doubao-seed-2-0-mini-260215","name":"Doubao Seed 2.0 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":32000},"cost":{"input":0.0493,"output":0.4845,"cache_read":0.02465}},"gemini-2.5-flash-lite-preview-09-2025-thinking":{"id":"gemini-2.5-flash-lite-preview-09-2025-thinking","name":"Gemini 2.5 Flash Lite Preview (09/2025) – Thinking","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01}},"asi1-mini":{"id":"asi1-mini","name":"ASI1 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-03-25","last_updated":"2025-03-25","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":1,"output":1,"cache_read":0.5}},"claw-medium":{"id":"claw-medium","name":"Claw Medium","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-11","last_updated":"2026-05-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2}},"auto-model-standard":{"id":"auto-model-standard","name":"Auto model (Standard)","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2025-04-16","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":1000000},"cost":{"input":9.996,"output":19.992,"cache_read":4.998}},"ernie-5.0-thinking-preview":{"id":"ernie-5.0-thinking-preview","name":"Ernie 5.0 Thinking Preview","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"ernie","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":1,"output":3.5,"cache_read":0.5}},"claw-high":{"id":"claw-high","name":"Claw High","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-11","last_updated":"2026-05-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2}},"agnes-3.0-flash":{"id":"agnes-3.0-flash","name":"Agnes 3.0 Flash","description":"Agnes 3.0 Flash is a low-cost model for coding, tool use, and multi-turn agent tasks. It supports text and image input, optional thinking, and a 512K-token context window.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-09","last_updated":"2026-09-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"input":524288,"output":65536},"cost":{"input":0.05,"output":0.15,"cache_read":0.005}},"qvq-max":{"id":"qvq-max","name":"Qwen: QvQ Max","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-03-28","last_updated":"2025-03-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":1.2,"output":4.8,"cache_read":0.6}},"gemma-4-26b-a4b-it-musica":{"id":"gemma-4-26b-a4b-it-musica","name":"Musica","description":"Musica is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"gemini-2.0-pro-reasoner":{"id":"gemini-2.0-pro-reasoner","name":"Gemini 2.0 Pro Reasoner","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-02-05","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":1.292,"output":4.998,"cache_read":0.323}},"glm-4.1v-thinking-flashx":{"id":"glm-4.1v-thinking-flashx","name":"GLM 4.1V Thinking FlashX","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"glm-flash","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":64000,"input":64000,"output":8192},"cost":{"input":0.3,"output":0.3,"cache_read":0.15}},"gemma-4-31b-it-fabled":{"id":"gemma-4-31b-it-fabled","name":"Fabled","description":"Fabled is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"deepseek-v3-0324":{"id":"deepseek-v3-0324","name":"DeepSeek Chat 0324","description":"March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.2,"output":0.77,"cache_read":0.135},"canonical_model_id":"deepseek/deepseek-v3-0324"},"gemma-4-26b-a4b-it-darksoul":{"id":"gemma-4-26b-a4b-it-darksoul","name":"Dark Soul","description":"Dark Soul is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"pokee-isaac":{"id":"pokee-isaac","name":"Pokee-Isaac 28B","description":"Pokee-Isaac is a 28B agentic model with a roughly 10-million-token context window, function calling, and OpenAI-compatible structured output. Pokee bills in $0.01 increments, rounding each non-zero request up to the next cent.","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-08-04","last_updated":"2026-08-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":10000000,"input":10000000,"output":60000},"cost":{"input":0.15,"output":1,"cache_read":0.075}},"doubao-seed-2-0-pro-260215":{"id":"doubao-seed-2-0-pro-260215","name":"Doubao Seed 2.0 Pro","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":128000},"cost":{"input":0.782,"output":3.876,"cache_read":0.391}},"Gemma-4-31B-MeroMero-v2":{"id":"Gemma-4-31B-MeroMero-v2","name":"Gemma 4 31B MeroMero v2","description":"Gemma 4 31B MeroMero v2 is a LoRA finetune for emotive dialogue, relationship scenes, creative writing, and multimodal roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-08-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"gemini-2.5-flash-preview-09-2025-thinking":{"id":"gemini-2.5-flash-preview-09-2025-thinking","name":"Gemini 2.5 Flash Preview (09/2025) – Thinking","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03}},"gemini-exp-1206":{"id":"gemini-exp-1206","name":"Gemini 2.0 Pro 1206","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":2097152,"input":2097152,"output":8192},"cost":{"input":1.258,"output":4.998,"cache_read":0.629}},"gemini-2.5-flash-lite-preview-09-2025":{"id":"gemini-2.5-flash-lite-preview-09-2025","name":"Gemini 2.5 Flash Lite Preview (09/2025)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01}},"universal-summarizer":{"id":"universal-summarizer","name":"Universal Summarizer","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-12-23","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":30,"output":30}},"gemma-4-31b-it-gemsicle":{"id":"gemma-4-31b-it-gemsicle","name":"Gemsicle","description":"Gemsicle is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"claw-low":{"id":"claw-low","name":"Claw Low","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-11","last_updated":"2026-05-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2,"cache_write":0.08333}},"deepseek-chat-cheaper":{"id":"deepseek-chat-cheaper","name":"DeepSeek V3/Chat Cheaper","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"deepseek","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.1,"output":0.425,"cache_read":0.05}},"GLM-4.6-Derestricted-v5":{"id":"GLM-4.6-Derestricted-v5","name":"GLM 4.6 Derestricted v5","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":8192},"cost":{"input":0.4,"output":1.5,"cache_read":0.2}},"holo3-35b-a3b:thinking":{"id":"holo3-35b-a3b:thinking","name":"Holo3-35B-A3B Thinking","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.25,"output":1.8,"cache_read":0.125}},"MiniMax-M1":{"id":"MiniMax-M1","name":"MiniMax M1","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"minimax","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-01-08","last_updated":"2025-06-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":131072},"cost":{"input":0.1394,"output":1.3328,"cache_read":0.0697}},"deepseek-chat":{"id":"deepseek-chat","name":"DeepSeek V3/Deepseek Chat","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.1,"output":0.425,"cache_read":0.05},"canonical_model_id":"deepseek/deepseek-chat"},"glm-4-air-0111":{"id":"glm-4-air-0111","name":"GLM 4 Air 0111","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-01-11","last_updated":"2025-01-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":4096},"cost":{"input":0.1394,"output":0.1394,"cache_read":0.0697}},"Gemma-4-26B-A4B-MeroMero":{"id":"Gemma-4-26B-A4B-MeroMero","name":"Gemma 4 26B A4B MeroMero","description":"Gemma 4 26B A4B MeroMero is an NVFP4 multimodal mixture-of-experts fine-tune for emotive dialogue, relationship scenes, creative writing, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"Gemma-4-26B-A4B-MeroMero:thinking":{"id":"Gemma-4-26B-A4B-MeroMero:thinking","name":"Gemma 4 26B A4B MeroMero Thinking","description":"Gemma 4 26B A4B MeroMero with thinking enabled for more deliberate emotive dialogue, relationship scenes, creative writing, and multimodal roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"fastgpt":{"id":"fastgpt","name":"Web Answer","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-12-23","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":7.5,"output":7.5}},"gemini-2.5-flash-preview-04-17:thinking":{"id":"gemini-2.5-flash-preview-04-17:thinking","name":"Gemini 2.5 Flash Preview Thinking","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"release_date":"2025-04-17","last_updated":"2025-04-17","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.15,"output":3.5,"cache_read":0.015}},"gemma-4-31b-it-gembrain":{"id":"gemma-4-31b-it-gembrain","name":"Gembrain","description":"Gembrain is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"hermes-low":{"id":"hermes-low","name":"Hermes Low","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"hermes","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-11","last_updated":"2026-05-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2,"cache_write":0.08333}},"gemini-2.5-flash-preview-05-20":{"id":"gemini-2.5-flash-preview-05-20","name":"Gemini 2.5 Flash 0520","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048000,"input":1048000,"output":65536},"cost":{"input":0.15,"output":0.6,"cache_read":0.015}},"auto-model-premium":{"id":"auto-model-premium","name":"Auto model (Premium)","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2025-04-16","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":1000000},"cost":{"input":9.996,"output":19.992,"cache_read":4.998}},"mercury-coder-small":{"id":"mercury-coder-small","name":"Mercury Coder Small","description":"Model by Inception AI. A diffusion large language model that runs incredibly quickly (500+ tokens/second) while matching Claude 3.5 Haiku and GPT-4o-mini. 1st in speed on Copilot arena, and matching 2nd in quality.","family":"mercury","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.25,"output":1,"cache_read":0.125}},"nano-gpt-help":{"id":"nano-gpt-help","name":"NanoGPT Help","description":"Text-only NanoGPT support assistant. Questions are processed by the Help inference provider; do not paste secrets or account credentials. Covers the website, models, API, pricing, memory, media generation, and support.","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-06-06","last_updated":"2026-06-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":6000,"input":6000,"output":512},"cost":{"input":0,"output":0}},"Gemma-4-31B-Cognitive-Unshackled":{"id":"Gemma-4-31B-Cognitive-Unshackled","name":"Gemma 4 31B Cognitive Unshackled","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.306,"output":0.306,"cache_read":0.153}},"kimi-k2-instruct-fast":{"id":"kimi-k2-instruct-fast","name":"Kimi K2 0711 Fast","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"kimi-k2","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-12-15","last_updated":"2025-07-15","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":16384},"cost":{"input":0.4,"output":1.8,"cache_read":0.2}},"gemini-2.5-pro-preview-03-25":{"id":"gemini-2.5-pro-preview-03-25","name":"Gemini 2.5 Pro Preview 0325","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"release_date":"2025-03-25","last_updated":"2025-03-25","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2.5,"output":10,"cache_read":0.25}},"doubao-seed-2-0-lite-260215":{"id":"doubao-seed-2-0-lite-260215","name":"Doubao Seed 2.0 Lite","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":32000},"cost":{"input":0.1462,"output":0.8738,"cache_read":0.0731}},"holo3-35b-a3b":{"id":"holo3-35b-a3b","name":"Holo3-35B-A3B","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.25,"output":1.8,"cache_read":0.125}},"phi-4-mini-instruct":{"id":"phi-4-mini-instruct","name":"Phi 4 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"phi","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.17,"output":0.68,"cache_read":0.085}},"gemma-4-26b-a4b-it-chimerax":{"id":"gemma-4-26b-a4b-it-chimerax","name":"Chimera X","description":"Chimera X is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"gemini-2.5-pro-preview-05-06":{"id":"gemini-2.5-pro-preview-05-06","name":"Gemini 2.5 Pro Preview 0506","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"release_date":"2025-05-06","last_updated":"2025-05-06","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2.5,"output":10,"cache_read":0.25}},"gemini-2.5-pro-exp-03-25":{"id":"gemini-2.5-pro-exp-03-25","name":"Gemini 2.5 Pro Experimental 0325","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"release_date":"2025-03-25","last_updated":"2025-03-25","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2.5,"output":10,"cache_read":0.25}},"Gemma-4-31B-MeroMero-v2:thinking":{"id":"Gemma-4-31B-MeroMero-v2:thinking","name":"Gemma 4 31B MeroMero v2 Thinking","description":"Gemma 4 31B MeroMero v2 with thinking enabled for more deliberate emotive dialogue, relationship scenes, creative writing, and multimodal roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-08-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"gemma-4-26b-a4b-it-luminous":{"id":"gemma-4-26b-a4b-it-luminous","name":"Luminous Mirror","description":"Luminous Mirror is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"Qwen3.5-27B-Queen-Derestricted":{"id":"Qwen3.5-27B-Queen-Derestricted","name":"Qwen3.5 27B Queen Derestricted","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.306,"output":0.306,"cache_read":0.153}},"mistral-code-latest":{"id":"mistral-code-latest","name":"Mistral Code Latest","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"mistral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.15}},"gemma-4-12b-it-semancer":{"id":"gemma-4-12b-it-semancer","name":"Gemma 4 12B Semancer","description":"Gemma 4 12B Semancer is an open-weight roleplay finetune with image understanding, tool calling, optional reasoning, and a 131,072-token context window.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-09","last_updated":"2026-09-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.05,"output":0.25,"cache_read":0.025}},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-2.5-flash"},"doubao-1.5-vision-pro-32k":{"id":"doubao-1.5-vision-pro-32k","name":"Doubao 1.5 Vision Pro 32k","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-11-20","last_updated":"2025-01-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"input":32000,"output":8192},"cost":{"input":0.459,"output":1.377,"cache_read":0.2295}},"hermes-medium":{"id":"hermes-medium","name":"Hermes Medium","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"hermes","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-11","last_updated":"2026-05-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2}},"gemini-2.5-flash-lite-preview-06-17":{"id":"gemini-2.5-flash-lite-preview-06-17","name":"Gemini 2.5 Flash Lite Preview","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.15,"output":0.6,"cache_read":0.015}},"gemini-2.5-flash-preview-05-20:thinking":{"id":"gemini-2.5-flash-preview-05-20:thinking","name":"Gemini 2.5 Flash 0520 Thinking","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048000,"input":1048000,"output":65536},"cost":{"input":0.15,"output":3.5,"cache_read":0.015}},"gemma-4-31b-it-darkidol":{"id":"gemma-4-31b-it-darkidol","name":"DarkIdol","description":"DarkIdol is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"gemma-4-26b-a4b-it-chimeraxheretic":{"id":"gemma-4-26b-a4b-it-chimeraxheretic","name":"Gemma 4 26B A4B Chimera X Heretic","description":"Chimera X Heretic is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-08","last_updated":"2026-10-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"auto-model-basic":{"id":"auto-model-basic","name":"Auto model (Basic)","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2025-04-16","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":1000000},"cost":{"input":9.996,"output":19.992,"cache_read":4.998}},"Gemma-4-31B-Claude-4.6-Opus-Reasoning-Distilled":{"id":"Gemma-4-31B-Claude-4.6-Opus-Reasoning-Distilled","name":"Gemma 4 31B Claude 4.6 Opus Reasoning Distilled","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"claude","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.306,"output":0.306,"cache_read":0.0306}},"doubao-1.5-pro-32k":{"id":"doubao-1.5-pro-32k","name":"Doubao 1.5 Pro 32k","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-11-20","last_updated":"2025-01-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"input":32000,"output":8192},"cost":{"input":0.1343,"output":0.3349,"cache_read":0.06715}},"gemini-2.5-flash-preview-04-17":{"id":"gemini-2.5-flash-preview-04-17","name":"Gemini 2.5 Flash Preview","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"release_date":"2025-04-17","last_updated":"2025-04-17","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.15,"output":0.6,"cache_read":0.015}},"hermes-high":{"id":"hermes-high","name":"Hermes High","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"hermes","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-11","last_updated":"2026-05-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2}},"ernie-5.1:thinking":{"id":"ernie-5.1:thinking","name":"ERNIE 5.1 Thinking","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"ernie","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2026-05-10","last_updated":"2026-05-10","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":119000,"input":119000,"output":64000},"cost":{"input":0.75,"output":3,"cache_read":0.75}},"gemma-4-26b-a4b-it-shadowsiren":{"id":"gemma-4-26b-a4b-it-shadowsiren","name":"Shadow Siren","description":"Shadow Siren is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"undi95/remm-slerp-l2-13b":{"id":"undi95/remm-slerp-l2-13b","name":"ReMM SLERP 13B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-06-30","release_date":"2023-07-22","last_updated":"2023-07-22","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":6144,"input":6144,"output":4096},"cost":{"input":0.799,"output":1.207,"cache_read":0.3995},"canonical_model_id":"undi95/remm-slerp-l2-13b"},"aion-labs/aion-3.0-mini":{"id":"aion-labs/aion-3.0-mini","name":"AionLabs: Aion 3.0 Mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-14","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.7,"output":1.4,"cache_read":0.18},"canonical_model_id":"aion-labs/aion-3.0-mini"},"aion-labs/aion-rp-llama-3.1-8b":{"id":"aion-labs/aion-rp-llama-3.1-8b","name":"Llama 3.1 8b (uncensored)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12-31","release_date":"2025-02-04","last_updated":"2025-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.8,"output":1.6,"cache_read":0.4},"canonical_model_id":"aion-labs/aion-rp-llama-3.1-8b"},"aion-labs/aion-3.5":{"id":"aion-labs/aion-3.5","name":"AionLabs: Aion 3.5","description":"A GLM-family collaborative generation model tuned for immersive roleplay and storytelling, with stronger narrative structure, tension, conflict, and nuanced mature themes.","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":3,"output":6,"cache_read":0.75}},"aion-labs/aion-3.5-mini":{"id":"aion-labs/aion-3.5-mini","name":"AionLabs: Aion 3.5 Mini","description":"A GLM-family collaborative generation model tuned for immersive roleplay and storytelling, with stronger narrative structure, tension, conflict, and nuanced mature themes.","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.7,"output":1.4,"cache_read":0.18}},"aion-labs/aion-2.0":{"id":"aion-labs/aion-2.0","name":"AionLabs: Aion-2.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-21","last_updated":"2025-12-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.8,"output":1.6,"cache_read":0.2},"canonical_model_id":"aion-labs/aion-2.0"},"aion-labs/aion-3.0":{"id":"aion-labs/aion-3.0","name":"AionLabs: Aion 3.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-05","last_updated":"2026-05-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":3,"output":6,"cache_read":0.75},"canonical_model_id":"aion-labs/aion-3.0"},"baseten/Kimi-K2-Instruct-FP4":{"id":"baseten/Kimi-K2-Instruct-FP4","name":"Kimi K2 0711 Instruct FP4","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":131072},"cost":{"input":0.4,"output":1.8,"cache_read":0.2}},"TheDrummer/Cydonia-24B-v4.1":{"id":"TheDrummer/Cydonia-24B-v4.1","name":"The Drummer Cydonia 24B v4.1","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04-30","release_date":"2025-09-27","last_updated":"2025-09-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":117964},"cost":{"input":0.3,"output":0.5,"cache_read":0.15},"canonical_model_id":"thedrummer/cydonia-24b-v4.1"},"TheDrummer/Magidonia-24B-v4.3":{"id":"TheDrummer/Magidonia-24B-v4.3","name":"The Drummer Magidonia 24B v4.3","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-12-25","last_updated":"2025-12-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":0.1003,"output":0.1207,"cache_read":0.05015}},"TheDrummer/UnslopNemo-12B-v4.1":{"id":"TheDrummer/UnslopNemo-12B-v4.1","name":"UnslopNemo 12b v4","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"input":8192,"output":26214},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"TheDrummer/Cydonia-24B-v4.3":{"id":"TheDrummer/Cydonia-24B-v4.3","name":"The Drummer Cydonia 24B v4.3","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-12-25","last_updated":"2025-12-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.12,"output":0.15,"cache_read":0.06}},"TheDrummer/Cydonia-24B-v4":{"id":"TheDrummer/Cydonia-24B-v4","name":"The Drummer Cydonia 24B v4","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":0.2006,"output":0.2414,"cache_read":0.1003}},"TheDrummer/Anubis-70B-v1":{"id":"TheDrummer/Anubis-70B-v1","name":"Anubis 70B v1","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":16384},"cost":{"input":0.31,"output":0.31,"cache_read":0.155}},"TheDrummer/Rocinante-12B-v1.1":{"id":"TheDrummer/Rocinante-12B-v1.1","name":"Rocinante 12b","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.408,"output":0.595,"cache_read":0.204}},"TheDrummer/Anubis-70B-v1.1":{"id":"TheDrummer/Anubis-70B-v1.1","name":"Anubis 70B v1.1","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"input":32000,"output":16384},"cost":{"input":0.31,"output":0.31,"cache_read":0.155}},"TheDrummer/skyfall-36b-v2":{"id":"TheDrummer/skyfall-36b-v2","name":"TheDrummer Skyfall 36B V2","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06-30","release_date":"2025-03-10","last_updated":"2025-03-10","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":29491},"cost":{"input":0.55,"output":0.8,"cache_read":0.25},"canonical_model_id":"thedrummer/skyfall-36b-v2"},"TheDrummer/orion-26b-a4b-v1.1":{"id":"TheDrummer/orion-26b-a4b-v1.1","name":"Orion 26B A4B v1.1","description":"TheDrummer's Gemma 4 26B A4B fine-tune for creative writing, roleplay, and expressive dialogue. Supports optional thinking with a 65,536-token context window and up to 8,192 output tokens. Text input only.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":false,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"TheDrummer/Cydonia-24B-v2":{"id":"TheDrummer/Cydonia-24B-v2","name":"The Drummer Cydonia 24B v2","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-02-17","last_updated":"2025-02-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":0.1003,"output":0.1207,"cache_read":0.05015}},"TheDrummer/Artemis-v1.1":{"id":"TheDrummer/Artemis-v1.1","name":"TheDrummer/Artemis v1.1","description":"TheDrummer's Artemis v1.1 is a Gemma 4 31B fine-tune for creative writing, expressive dialogue, and roleplay, with optional thinking and a 262K context window.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-06","last_updated":"2026-09-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"Salesforce/Llama-xLAM-2-70b-fc-r":{"id":"Salesforce/Llama-xLAM-2-70b-fc-r","name":"Llama-xLAM-2 70B fc-r","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-04-13","last_updated":"2025-04-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":2.5,"output":2.5,"cache_read":1.25}},"Gryphe/MythoMax-L2-13b":{"id":"Gryphe/MythoMax-L2-13b","name":"MythoMax 13B","description":"Llama 2 merge combining MythoLogic and Huginn for creative writing and roleplay","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2023-08-10","last_updated":"2023-08-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"input":4096,"output":3686},"cost":{"input":0.1003,"output":0.1003,"cache_read":0.05015},"canonical_model_id":"gryphe/mythomax-l2-13b"},"inception/mercury-2.5-preview":{"id":"inception/mercury-2.5-preview","name":"Mercury 2.5 Preview","description":"Mercury 2.5 Preview is Inception's latest and most intelligent diffusion language model. Instead of generating tokens strictly one at a time, it produces and refines multiple tokens in parallel, reaching up to 1,107 tokens per second on standard GPUs. It delivers a 10+ point intelligence gain over Mercury 2, with tunable reasoning, parallel tool calls, schema-aligned JSON output, and a 260K context window. It is built for latency-sensitive production work such as search agents, voice pipelines, customer support, rapid coding iteration, and coding subagents.","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":260000,"input":260000,"output":65536},"cost":{"input":0.04,"output":0.15,"cache_read":0.004}},"inception/mercury-3":{"id":"inception/mercury-3","name":"Mercury 3","description":"Mercury 3 is Inception's diffusion language model for fast chat, coding, and multi-step agents. It accepts and generates text, with adjustable reasoning, parallel tool calls, structured JSON output, and a shared 1,000,000-token context window. Responses can use up to 131,072 tokens, including reasoning, within the remaining context.","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":131072},"cost":{"input":0.1,"output":0.35,"cache_read":0.01}},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.1:thinking:1024":{"id":"anthropic/claude-opus-4.1:thinking:1024","name":"Claude 4.1 Opus Thinking (1K)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5},"canonical_model_id":"anthropic/claude-opus-4-1"},"anthropic/claude-opus-latest":{"id":"anthropic/claude-opus-latest","name":"Claude Opus Latest","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-29","last_updated":"2026-03-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5}},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude 4.7 Opus","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-sonnet-5:thinking":{"id":"anthropic/claude-sonnet-5:thinking","name":"Claude Sonnet 5 Thinking","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-opus-4.8:thinking":{"id":"anthropic/claude-opus-4.8:thinking","name":"Claude Opus 4.8 Thinking","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-opus-4.6:thinking:low":{"id":"anthropic/claude-opus-4.6:thinking:low","name":"Claude 4.6 Opus Thinking Low","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-sonnet-4.6:thinking":{"id":"anthropic/claude-sonnet-4.6:thinking","name":"Claude Sonnet 4.6 Thinking","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-sonnet-4.5:thinking":{"id":"anthropic/claude-sonnet-4.5:thinking","name":"Claude Sonnet 4.5 Thinking","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-haiku-latest":{"id":"anthropic/claude-haiku-latest","name":"Claude Haiku Latest","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-29","last_updated":"2026-03-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125}},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude 4.5 Opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4.5:thinking":{"id":"anthropic/claude-opus-4.5:thinking","name":"Claude 4.5 Opus Thinking","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-opus-4.6:thinking":{"id":"anthropic/claude-opus-4.6:thinking","name":"Claude 4.6 Opus Thinking","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-4.7:thinking":{"id":"anthropic/claude-opus-4.7:thinking","name":"Claude 4.7 Opus Thinking","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-sonnet-5.5:thinking":{"id":"anthropic/claude-sonnet-5.5:thinking","name":"Claude Sonnet 5.5 Thinking","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude 4.6 Opus","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-4.1:thinking:32768":{"id":"anthropic/claude-opus-4.1:thinking:32768","name":"Claude 4.1 Opus Thinking (32K)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5},"canonical_model_id":"anthropic/claude-opus-4-1"},"anthropic/claude-fable-latest":{"id":"anthropic/claude-fable-latest","name":"Claude Fable Latest","description":"Compatibility alias for Claude Fable.","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5}},"anthropic/claude-haiku-4.5:thinking":{"id":"anthropic/claude-haiku-4.5:thinking","name":"Claude Haiku 4.5 Thinking","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-opus-4.6:thinking:max":{"id":"anthropic/claude-opus-4.6:thinking:max","name":"Claude 4.6 Opus Thinking Max","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-4.1":{"id":"anthropic/claude-opus-4.1","name":"Claude 4.1 Opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5},"canonical_model_id":"anthropic/claude-opus-4-1"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-opus-4.6:thinking:medium":{"id":"anthropic/claude-opus-4.6:thinking:medium","name":"Claude 4.6 Opus Thinking Medium","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-4.1:thinking:8192":{"id":"anthropic/claude-opus-4.1:thinking:8192","name":"Claude 4.1 Opus Thinking (8K)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5},"canonical_model_id":"anthropic/claude-opus-4-1"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-sonnet-latest":{"id":"anthropic/claude-sonnet-latest","name":"Claude Sonnet Latest","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-01","last_updated":"2026-03-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5}},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-opus-4.1:thinking":{"id":"anthropic/claude-opus-4.1:thinking","name":"Claude 4.1 Opus Thinking","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5},"canonical_model_id":"anthropic/claude-opus-4-1"},"amazon/nova-2-lite-v1":{"id":"amazon/nova-2-lite-v1","name":"Amazon Nova 2 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":65535},"cost":{"input":0.51,"output":4.25,"cache_read":0.255},"canonical_model_id":"amazon/nova-2-lite-v1"},"amazon/nova-lite-v1":{"id":"amazon/nova-lite-v1","name":"Amazon Nova Lite 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"input":300000,"output":5120},"cost":{"input":0.0595,"output":0.238,"cache_read":0.02975},"canonical_model_id":"amazon/nova-lite-v1"},"amazon/nova-pro-v1":{"id":"amazon/nova-pro-v1","name":"Amazon Nova Pro 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"input":300000,"output":32000},"cost":{"input":0.799,"output":3.196,"cache_read":0.3995},"canonical_model_id":"amazon/nova-pro-v1"},"anthracite-org/magnum-v4-72b":{"id":"anthracite-org/magnum-v4-72b","name":"Magnum v4 72B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":2.006,"output":2.992,"cache_read":1.003},"canonical_model_id":"anthracite-org/magnum-v4-72b"},"anthracite-org/magnum-v2-72b":{"id":"anthracite-org/magnum-v2-72b","name":"Magnum V2 72B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":2.006,"output":2.992,"cache_read":1.003}},"LLM360/K2-Think":{"id":"LLM360/K2-Think","name":"K2-Think","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":32768},"cost":{"input":0.17,"output":0.68,"cache_read":0.085}},"ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0":{"id":"ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0","name":"Omega Directive 24B Unslop v2.0","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":0.5,"output":0.5,"cache_read":0.25}},"unsloth/gemma-3-12b-it":{"id":"unsloth/gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"unsloth","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-03-10","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":16384},"cost":{"input":0.272,"output":0.272,"cache_read":0.136}},"unsloth/gemma-3-27b-it":{"id":"unsloth/gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"unsloth","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-03-10","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":96000},"cost":{"input":0.2992,"output":0.2992,"cache_read":0.1496}},"unsloth/gemma-3-4b-it":{"id":"unsloth/gemma-3-4b-it","name":"Gemma 3 4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"unsloth","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-03-10","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.2006,"output":0.2006,"cache_read":0.1003}},"deepseek/deepseek-v4-flash-0423":{"id":"deepseek/deepseek-v4-flash-0423","name":"DeepSeek V4 Flash 0423","description":"Initial DeepSeek V4 Flash snapshot for economical reasoning, coding, and million-token agent workloads","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0423"},"deepseek/deepseek-v4-flash-0731:thinking":{"id":"deepseek/deepseek-v4-flash-0731:thinking","name":"DeepSeek V4 Flash 0731 (Thinking)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.05,"output":0.16,"cache_read":0.013},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.7,"output":1.4,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-pro:thinking":{"id":"deepseek/deepseek-v4-pro:thinking","name":"DeepSeek V4 Pro (Thinking)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.7,"output":1.4,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.05,"output":0.16,"cache_read":0.013},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-v4-flash-0423:thinking":{"id":"deepseek/deepseek-v4-flash-0423:thinking","name":"DeepSeek V4 Flash 0423 (Thinking)","description":"Initial DeepSeek V4 Flash snapshot for economical reasoning, coding, and million-token agent workloads","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0423"},"deepseek/deepseek-latest":{"id":"deepseek/deepseek-latest","name":"DeepSeek Latest","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-03","last_updated":"2026-05-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022}},"deepseek/deepseek-v4.1-flash-1017:thinking":{"id":"deepseek/deepseek-v4.1-flash-1017:thinking","name":"DeepSeek V4.1 Flash 1017 Thinking","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":393216},"cost":{"input":0.1,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-pro-0813:thinking":{"id":"deepseek/deepseek-v4-pro-0813:thinking","name":"DeepSeek V4 Pro 0813 Thinking","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4.1-flash:thinking":{"id":"deepseek/deepseek-v4.1-flash:thinking","name":"DeepSeek V4.1 Flash Thinking","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":384000},"cost":{"input":0.14,"output":0.42,"cache_read":0.0042},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.2156,"output":0.6468,"cache_read":0.00686},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v3.2:thinking":{"id":"deepseek/deepseek-v3.2:thinking","name":"DeepSeek V3.2 Thinking","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":163000,"input":163000,"output":65536},"cost":{"input":0.2088,"output":0.3096,"cache_read":0.0216},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":163000,"input":163000,"output":65536},"cost":{"input":0.2088,"output":0.3096,"cache_read":0.0216},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4-flash-latest":{"id":"deepseek/deepseek-v4-flash-latest","name":"DeepSeek V4 Flash Latest","description":"Compatibility alias that routes to the newest dated DeepSeek V4 Flash release. Currently routes to DeepSeek V4 Flash 0731. ⚠️ This route goes directly to DeepSeek, so privacy and logging guarantees are limited.","family":"deepseek","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-08-02","last_updated":"2026-08-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":393216},"cost":{"input":0.1,"output":0.6,"cache_read":0.003}},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":393216},"cost":{"input":0.1,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4-flash:thinking":{"id":"deepseek/deepseek-v4-flash:thinking","name":"DeepSeek V4 Flash (Thinking)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":384000},"cost":{"input":0.14,"output":0.42,"cache_read":0.0042},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"microsoft/wizardlm-2-8x22b":{"id":"microsoft/wizardlm-2-8x22b","name":"WizardLM-2 8x22B","description":"Microsoft WizardLM instruction-tuned Mixtral MoE for multilingual dialogue and complex tasks","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-04-15","last_updated":"2024-04-15","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465},"canonical_model_id":"microsoft/wizardlm-2-8x22b"},"nvidia/nemotron-3.5-content-safety":{"id":"nvidia/nemotron-3.5-content-safety","name":"Nemotron 3.5 Content Safety","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.05,"output":0.15,"cache_read":0.025},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"nvidia/nemotron-3-ultra-550b-a55b:thinking":{"id":"nvidia/nemotron-3-ultra-550b-a55b:thinking","name":"Nvidia Nemotron 3 Ultra 550B Thinking","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":0.5,"output":2.5,"cache_read":0.25},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/nemotron-3.5-lightning":{"id":"nvidia/nemotron-3.5-lightning","name":"Nvidia Nemotron 3.5 Lightning","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":0.05,"output":0.2,"cache_read":0.01},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-3.5-super-vl":{"id":"nvidia/nemotron-3.5-super-vl","name":"Nvidia Nemotron 3.5 Super VL","description":"NVIDIA’s open-weight 120B model (12B active) for reasoning, coding, tool use, document and GUI understanding, and video understanding. Text, images and video in; text out. Served in NVFP4 with a 262,144-token shared context window. NanoGPT currently limits responses to 16,384 tokens while separate hosted output limits remain unpublished. OpenMDW 1.1 license.","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.15,"output":0.5,"cache_read":0.02}},"nvidia/nemotron-3-ultra-550b-a55b":{"id":"nvidia/nemotron-3-ultra-550b-a55b","name":"Nvidia Nemotron 3 Ultra 550B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":0.5,"output":2.5,"cache_read":0.25},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/nemotron-3-super-120b-a12b:thinking":{"id":"nvidia/nemotron-3-super-120b-a12b:thinking","name":"Nvidia Nemotron 3 Super 120B Thinking","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-01","last_updated":"2026-03-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.05,"output":0.25,"cache_read":0.025},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/Llama-3.1-Nemotron-70B-Instruct-HF":{"id":"nvidia/Llama-3.1-Nemotron-70B-Instruct-HF","name":"Nvidia Nemotron 70b","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.357,"output":0.408,"cache_read":0.1785},"canonical_model_id":"nvidia/llama-3.1-nemotron-70b-instruct"},"nvidia/Llama-3.3-Nemotron-Super-49B-v1":{"id":"nvidia/Llama-3.3-Nemotron-Super-49B-v1","name":"Nvidia Nemotron Super 49B","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-08-08","last_updated":"2025-08-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.15,"output":0.15,"cache_read":0.075},"canonical_model_id":"nvidia/llama-3.3-nemotron-super-49b-v1"},"nvidia/nemotron-3-nano-30b-a3b":{"id":"nvidia/nemotron-3-nano-30b-a3b","name":"Nvidia Nemotron 3 Nano 30B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":235929},"cost":{"input":0.17,"output":0.68,"cache_read":0.085},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/nemotron-3-super-120b-a12b":{"id":"nvidia/nemotron-3-super-120b-a12b","name":"Nvidia Nemotron 3 Super 120B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-01","last_updated":"2026-03-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.05,"output":0.25,"cache_read":0.025},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/nemotron-3.5-lightning:thinking":{"id":"nvidia/nemotron-3.5-lightning:thinking","name":"Nvidia Nemotron 3.5 Lightning Thinking","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":0.05,"output":0.2,"cache_read":0.01},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"z-ai/glm-4.6v":{"id":"z-ai/glm-4.6v","name":"GLM 4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":24000},"cost":{"input":0.3,"output":0.9,"cache_read":0.15},"canonical_model_id":"zhipuai/glm-4.6v"},"z-ai/glm-5.3-flash-apex":{"id":"z-ai/glm-5.3-flash-apex","name":"Apex Flash 1","description":"Apex Flash 1 is Cantina Security and Yeta's cybersecurity-focused post-train of GLM 5.3 Flash for code investigation and tool use.","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":32768},"cost":{"input":0.15,"output":0.5,"cache_read":0.075}},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.1,"output":0.3,"cache_read":0.025},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.2:thinking":{"id":"z-ai/glm-5.2:thinking","name":"GLM 5.2 Thinking","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.42,"output":1.32,"cache_read":0.078},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-4.7-original:thinking":{"id":"z-ai/glm-4.7-original:thinking","name":"GLM 4.7 Original Thinking","description":"GLM-4.7 original with extended thinking capabilities for complex reasoning.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65535},"cost":{"input":0.6,"output":2.2,"cache_read":0.11}},"z-ai/glm-5:thinking":{"id":"z-ai/glm-5:thinking","name":"GLM 5 Thinking","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":128000},"cost":{"input":0.6,"output":1.9,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-4.5v":{"id":"z-ai/glm-4.5v","name":"GLM 4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":16384},"cost":{"input":0.6,"output":1.8,"cache_read":0.3},"canonical_model_id":"zhipuai/glm-4.5v"},"z-ai/glm-5.3-flash-uncensored":{"id":"z-ai/glm-5.3-flash-uncensored","name":"GLM 5.3 Flash Uncensored","description":"GLM 5.3 Flash Uncensored is an uncensored fine-tune of the efficient 320B mixture-of-experts reasoning model, built for unrestricted chat, creative writing, coding, agentic work, tool use, and long-context tasks.","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-07-29","last_updated":"2026-08-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":32768},"cost":{"input":0.2,"output":0.8,"cache_read":0.07}},"z-ai/glm-5.3:thinking":{"id":"z-ai/glm-5.3:thinking","name":"GLM 5.3 Thinking","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-5.3-flash-cybersecurity":{"id":"z-ai/glm-5.3-flash-cybersecurity","name":"GLM 5.3 Flash Cybersecurity","description":"GLM 5.3 Flash Cybersecurity is a cybersecurity-focused variant based on the uncensored model, with provider moderation for illegal activities. It supports always-on reasoning, image understanding, tool calling, and a 1,048,576-token context window.","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-09-19","last_updated":"2026-09-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":32768},"cost":{"input":0.15,"output":0.5,"cache_read":0.075}},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-5v-turbo:thinking":{"id":"z-ai/glm-5v-turbo:thinking","name":"GLM 5V Turbo Thinking","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":202800,"input":202800,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"z-ai/GLM-4.5-Air:thinking":{"id":"z-ai/GLM-4.5-Air:thinking","name":"GLM 4.5 Air (Thinking)","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":98304},"cost":{"input":0.12,"output":0.8,"cache_read":0.06},"canonical_model_id":"zhipuai/glm-4.5-air"},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"GLM 4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65535},"cost":{"input":0.4004,"output":1.9292,"cache_read":0.08008},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5.3-flash-apex-abliterated":{"id":"z-ai/glm-5.3-flash-apex-abliterated","name":"Apex Flash 1 Abliterated","description":"Apex Flash 1 Abliterated is an experimental derivative of Cantina Security and Yeta's Apex Flash 1 with modified refusal behavior.","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":32768},"cost":{"input":0.15,"output":0.5,"cache_read":0.075}},"z-ai/GLM-4.6-turbo:thinking":{"id":"z-ai/GLM-4.6-turbo:thinking","name":"GLM 4.6 Turbo (Thinking)","description":"GLM 4.6 Turbo with thinking mode enabled for enhanced reasoning; shows internal reasoning and supports long context.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2025-10-02","last_updated":"2025-10-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"input":204800,"output":131072},"cost":{"input":1,"output":3,"cache_read":0.5}},"z-ai/GLM-4.5:thinking":{"id":"z-ai/GLM-4.5:thinking","name":"GLM 4.5 (Thinking)","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.3,"output":1.3,"cache_read":0.15},"canonical_model_id":"zhipuai/glm-4.5"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM 5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":131072},"cost":{"input":0.7,"output":2.5,"cache_read":0.18},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/GLM-4.6-turbo":{"id":"z-ai/GLM-4.6-turbo","name":"GLM 4.6 Turbo","description":"Fast variant of GLM 4.6 for general chat, coding, and analysis with improved latency and strong reasoning.","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-10-02","last_updated":"2025-10-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"input":204800,"output":131072},"cost":{"input":1,"output":3,"cache_read":0.5}},"z-ai/glm-5-original":{"id":"z-ai/glm-5-original","name":"GLM 5 Original","description":"GLM-5 is Zhipu's latest flagship model with advanced reasoning and instruction following. Routed directly via Z-AI (Zhipu).","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":128000},"cost":{"input":0.6,"output":1.9,"cache_read":0.1}},"z-ai/glm-4.7-flash:thinking":{"id":"z-ai/glm-4.7-flash:thinking","name":"GLM 4.7 Flash Thinking","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":128000},"cost":{"input":0.07,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flash"},"z-ai/glm-4.6v-original":{"id":"z-ai/glm-4.6v-original","name":"GLM 4.6V Original","description":"GLM-4.6V scales its context window to 128k tokens in training, and achieves SoTA performance in visual understanding among models of similar parameter scales. Integrates native Function Calling capabilities, bridging 'visual perception' and 'executable action' for multimodal agents. Direct via Z-AI (Zhipu).","family":"glm","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":24000},"cost":{"input":0.6,"output":0.9,"cache_read":0.3}},"z-ai/glm-4.6-original":{"id":"z-ai/glm-4.6-original","name":"GLM 4.6 Original","description":"GLM-4.6, Zhipu's flagship text model with 256K context window and advanced reasoning capabilities. Direct via Z-AI (Zhipu).","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":65535},"cost":{"input":0.5,"output":2,"cache_read":0.1}},"z-ai/glm-4.7-flash":{"id":"z-ai/glm-4.7-flash","name":"GLM 4.7 Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":128000},"cost":{"input":0.07,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flash"},"z-ai/glm-4.7-original":{"id":"z-ai/glm-4.7-original","name":"GLM 4.7 Original","description":"GLM-4.7 is a next-gen GLM series text model with stronger reasoning, long-context chat, and reliable tool use. Routed directly via Z-AI (Zhipu).","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65535},"cost":{"input":0.6,"output":2.2,"cache_read":0.11}},"z-ai/glm-5.3-uncensored":{"id":"z-ai/glm-5.3-uncensored","name":"GLM 5.3 Uncensored","description":"An experimental uncensored version of the full GLM 5.3 reasoning model for chat, creative writing, coding, and tool use. It may show unusual behavior or inconsistent response quality.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-09-26","last_updated":"2026-09-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":1.25,"output":2.25,"cache_read":0.3}},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM 5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":128000},"cost":{"input":0.6,"output":1.9,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-4.6:thinking":{"id":"z-ai/glm-4.6:thinking","name":"GLM 4.6 Thinking","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65535},"cost":{"input":0.5,"output":2,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-4.6"},"z-ai/GLM-4.5-Air":{"id":"z-ai/GLM-4.5-Air","name":"GLM 4.5 Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":98304},"cost":{"input":0.12,"output":0.8,"cache_read":0.06},"canonical_model_id":"zhipuai/glm-4.5-air"},"z-ai/glm-4.7:thinking":{"id":"z-ai/glm-4.7:thinking","name":"GLM 4.7 Thinking","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65535},"cost":{"input":0.4004,"output":1.9292,"cache_read":0.08008},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5-original:thinking":{"id":"z-ai/glm-5-original:thinking","name":"GLM 5 Original Thinking","description":"GLM-5 original with extended thinking capabilities for complex reasoning.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":128000},"cost":{"input":0.6,"output":1.9,"cache_read":0.1}},"z-ai/glm-5v-turbo":{"id":"z-ai/glm-5v-turbo","name":"GLM 5V Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":202800,"input":202800,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM 5 Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202800,"input":202800,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"z-ai/glm-5.1:thinking":{"id":"z-ai/glm-5.1:thinking","name":"GLM 5.1 Thinking","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":131072},"cost":{"input":0.7,"output":2.5,"cache_read":0.18},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-4.6":{"id":"z-ai/glm-4.6","name":"GLM 4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65535},"cost":{"input":0.5,"output":2,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-4.6"},"z-ai/glm-latest":{"id":"z-ai/glm-latest","name":"GLM Latest","description":"Compatibility alias that routes to the newest thinking GLM model. Currently routes to GLM 5.2 Thinking.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-05-03","last_updated":"2026-05-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2}},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.42,"output":1.32,"cache_read":0.078},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-4.5":{"id":"z-ai/glm-4.5","name":"GLM 4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.3,"output":1.3,"cache_read":0.15},"canonical_model_id":"zhipuai/glm-4.5"},"z-ai/glm-4.5v:thinking":{"id":"z-ai/glm-4.5v:thinking","name":"GLM 4.5V Thinking","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":16384},"cost":{"input":0.6,"output":1.8,"cache_read":0.3},"canonical_model_id":"zhipuai/glm-4.5v"},"inclusionai/ling-3.0-flash-vl":{"id":"inclusionai/ling-3.0-flash-vl","name":"Ling 3.0 Flash VL","description":"Native multimodal hybrid MoE model for image and video understanding, reasoning, and visual agents","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012},"canonical_model_id":"inclusionai/ling-3.0-flash-vl"},"inclusionai/ling-3.0-flash":{"id":"inclusionai/ling-3.0-flash","name":"Ling 3.0 Flash","description":"Hybrid-linear MoE reasoning model with 124B total and 5.1B active parameters for efficient agents","family":"ling","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash"},"inclusionai/ling-3.0-flash:thinking":{"id":"inclusionai/ling-3.0-flash:thinking","name":"Ling 3.0 Flash Thinking","description":"Hybrid-linear MoE reasoning model with 124B total and 5.1B active parameters for efficient agents","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash"},"inclusionai/ling-3.1-flash":{"id":"inclusionai/ling-3.1-flash","name":"Ling 3.1 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.1-flash"},"meta/muse-spark-1.3-contributor":{"id":"meta/muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Meta's Muse Spark 1.3 Contributor is a frontier multimodal reasoning model for long-horizon coding and agentic workflows, with strong gains in computer use, browsing, professional tool use, codebase understanding, and million-token retrieval. It accepts text, images, audio, video, and files, supports tool calling and structured output, and always reasons before answering. Prompts and outputs may be used by Meta for training and to improve its products.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":943718},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/muse-glimmer-30b":{"id":"meta/muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":117964},"cost":{"input":0.35,"output":1.5,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"meta/muse-spark-1.2-contributor":{"id":"meta/muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor (Data Used for Training)","description":"A much cheaper opt-in version of Muse Spark 1.2 with the same multimodal coding and agentic capabilities. Prompts and outputs sent to this Contributor model may be used by Meta for training and to improve its products; use the standard Muse Spark 1.2 model if you do not want your data used for training.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-open":{"id":"meta/muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"deepseek-ai/DeepSeek-V3.1:thinking":{"id":"deepseek-ai/DeepSeek-V3.1:thinking","name":"DeepSeek V3.1 Thinking","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek-thinking","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.2,"output":0.7,"cache_read":0.1}},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.2,"output":0.7,"cache_read":0.1}},"deepseek-ai/deepseek-v3.2-exp":{"id":"deepseek-ai/deepseek-v3.2-exp","name":"DeepSeek V3.2 Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"input":163840,"output":65536},"cost":{"input":0.28,"output":0.42,"cache_read":0.14}},"deepseek-ai/DeepSeek-R1-0528":{"id":"deepseek-ai/DeepSeek-R1-0528","name":"DeepSeek R1 0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"input":163840,"output":32768},"cost":{"input":0.5,"output":2.15,"cache_read":0.35}},"deepseek-ai/DeepSeek-V3.1-Terminus":{"id":"deepseek-ai/DeepSeek-V3.1-Terminus","name":"DeepSeek V3.1 Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2025-08-02","last_updated":"2025-08-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.25,"output":0.7,"cache_read":0.125}},"deepseek-ai/DeepSeek-V3.1-Terminus:thinking":{"id":"deepseek-ai/DeepSeek-V3.1-Terminus:thinking","name":"DeepSeek V3.1 Terminus (Thinking)","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek-thinking","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2024-01-01","last_updated":"2025-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.25,"output":0.7,"cache_read":0.125}},"deepseek-ai/deepseek-v3.2-exp-thinking":{"id":"deepseek-ai/deepseek-v3.2-exp-thinking","name":"DeepSeek V3.2 Exp Thinking","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"input":163840,"output":65536},"cost":{"input":0.28,"output":0.42,"cache_read":0.14}},"bytedance-seed/seed-2-1-turbo":{"id":"bytedance-seed/seed-2-1-turbo","name":"ByteDance Seed 2.1 Turbo","description":"ByteDance Seed 2.1 Turbo is a multimodal model for coding and long-horizon agent workflows, including end-to-end software delivery and multi-step task execution. It supports text, image, and video input with a 262k context window.","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":235929},"cost":{"input":0.5,"output":2.5,"cache_read":0.25}},"bytedance-seed/seed-2.0-code":{"id":"bytedance-seed/seed-2.0-code","name":"ByteDance Seed 2.0 Code","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.5,"output":3,"cache_read":0.25},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"bytedance-seed/seed-2.0-lite":{"id":"bytedance-seed/seed-2.0-lite","name":"ByteDance Seed 2.0 Lite","description":"Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation","family":"seed","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.25,"output":2,"cache_read":0.125},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"meta-llama/llama-3.1-8b-instruct":{"id":"meta-llama/llama-3.1-8b-instruct","name":"Llama 3.1 8b Instruct","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":16384},"cost":{"input":0.02,"output":0.04,"cache_read":0.0272},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta-llama/llama-3.2-3b-instruct":{"id":"meta-llama/llama-3.2-3b-instruct","name":"Llama 3.2 3b Instruct","description":"Open Llama multimodal model for image understanding and text reasoning","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-09-25","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":8192},"cost":{"input":0.0306,"output":0.0493,"cache_read":0.0153}},"meta-llama/llama-3.3-70b-instruct":{"id":"meta-llama/llama-3.3-70b-instruct","name":"Llama 3.3 70b Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":16384},"cost":{"input":0.05,"output":0.23,"cache_read":0.025},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/llama-4-maverick":{"id":"meta-llama/llama-4-maverick","name":"Llama 4 Maverick","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.1875,"output":0.6525,"cache_read":0.05}},"meta-llama/llama-4-scout":{"id":"meta-llama/llama-4-scout","name":"Llama 4 Scout","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":328000,"input":328000,"output":65536},"cost":{"input":0.085,"output":0.46,"cache_read":0.0425}},"Steelskull/L3.3-Electra-R1-70b":{"id":"Steelskull/L3.3-Electra-R1-70b","name":"Steelskull Electra R1 70b","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.69989,"output":0.69989,"cache_read":0.349945}},"Steelskull/L3.3-Nevoria-R1-70b":{"id":"Steelskull/L3.3-Nevoria-R1-70b","name":"Steelskull Nevoria R1 70b","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"Steelskull/L3.3-Cu-Mai-R1-70b":{"id":"Steelskull/L3.3-Cu-Mai-R1-70b","name":"Llama 3.3 70B Cu Mai","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"Steelskull/L3.3-MS-Nevoria-70b":{"id":"Steelskull/L3.3-MS-Nevoria-70b","name":"Steelskull Nevoria 70b","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"qwen/qwen3.7-flash":{"id":"qwen/qwen3.7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":65536},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.038},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen/qwen3.8-27b-kimiko":{"id":"qwen/qwen3.8-27b-kimiko","name":"Qwen 3.8 27B Kimiko","description":"Qwen 3.8 27B Kimiko is a creative finetune served by Aoru in NVFP4 with a 262,144-token context window. It supports image understanding, tool calling, and optional reasoning.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3.8-27b-obliterated:thinking":{"id":"qwen/qwen3.8-27b-obliterated:thinking","name":"Qwen 3.8 27B Obliterated Thinking","description":"Qwen 3.8 27B Obliterated with thinking enabled for more deliberate creative work, coding, multimodal analysis, tool use, and long-context problem solving.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-07-29","last_updated":"2026-08-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"input":524288,"output":32768},"cost":{"input":0.25,"output":1.75,"cache_read":0.225}},"qwen/qwen3.5-9b":{"id":"qwen/qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":81920}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.05,"output":0.15,"cache_read":0.025},"canonical_model_id":"alibaba/qwen3.5-9b"},"qwen/qwen3.8-27b:thinking":{"id":"qwen/qwen3.8-27b:thinking","name":"Qwen3.8 27B Thinking","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1001536,"input":1001536,"output":32768},"cost":{"input":0.15,"output":0.7,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.8-27b-queen":{"id":"qwen/qwen3.8-27b-queen","name":"Qwen 3.8 27B Queen","description":"Qwen 3.8 27B Queen is an open-weight roleplay finetune with image understanding, tool calling, optional reasoning, and a 262,144-token context window.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-09","last_updated":"2026-09-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"input":524288,"output":32768},"cost":{"input":0.25,"output":1.5,"cache_read":0.125}},"qwen/qwen3.8-27b-obliterated":{"id":"qwen/qwen3.8-27b-obliterated","name":"Qwen 3.8 27B Obliterated","description":"Qwen 3.8 27B Obliterated is an open-weight multimodal model LoRA-tuned for fewer refusals across chat, coding, reasoning, tool use, and long-context work.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-07-29","last_updated":"2026-08-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"input":524288,"output":32768},"cost":{"input":0.25,"output":1.75,"cache_read":0.225}},"qwen/qwen3.7-max:thinking":{"id":"qwen/qwen3.7-max:thinking","name":"Qwen3.7 Max Thinking","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":1.475,"output":4.425,"cache_read":0.295,"cache_write":1.84375},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.8-27b-cybersecurity":{"id":"qwen/qwen3.8-27b-cybersecurity","name":"Qwen 3.8 27B Cybersecurity","description":"Qwen 3.8 27B Cybersecurity is a cybersecurity-focused variant based on the uncensored model, with provider moderation for illegal activities. It supports optional reasoning, image understanding, tool calling, and a 262,144-token context window.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-19","last_updated":"2026-09-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3-vl-235b-a22b-instruct":{"id":"qwen/qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.3,"output":1.9,"cache_read":0.15},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"qwen/qwen-plus":{"id":"qwen/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":995904,"input":995904,"output":32768},"cost":{"input":0.3995,"output":1.2002,"cache_read":0.19975},"canonical_model_id":"alibaba/qwen-plus"},"qwen/qwen3.5-omni-flash":{"id":"qwen/qwen3.5-omni-flash","name":"Qwen3.5 Omni Flash","description":"Qwen3.5 Omni Flash is Qwen's fast multimodal model. We verified live support for text prompts, images, audio files, and direct video URLs on Alibaba's chat-completions-compatible API. Alibaba describes Flash as a fully evolved version of Qwen3 Omni with audio input support across 60+ languages.","family":"qwen3.5","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-03-30","last_updated":"2026-03-30","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":49152,"input":49152,"output":16384}},"qwen/qwen3.7-flash:thinking":{"id":"qwen/qwen3.7-flash:thinking","name":"Qwen3.7 Flash Thinking","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"input":983616,"output":65536},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.038},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.26,"output":2.08,"cache_read":0.103788},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3-8b":{"id":"qwen/qwen3-8b","name":"Qwen 3 8B","description":"Qwen 3 8B is a 8B model. Supports switching between thinking and non thinking: trigger thinking with /think and /no_think anywhere in a prompt or system message to toggle chain-of-thought reasoning.","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":41000,"input":41000,"output":32768},"cost":{"input":0.47,"output":0.47,"cache_read":0.235}},"qwen/qwen3.5-plus":{"id":"qwen/qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"input":983616,"output":65536},"cost":{"input":0.4,"output":2.4,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen/qwen3.6-35b-a3b":{"id":"qwen/qwen3.6-35b-a3b","name":"Qwen3.6 35B A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.112,"output":0.8,"cache_read":0.056},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3.8-omni-flash":{"id":"qwen/qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"qwen/qwen3.5-35b-a3b:thinking":{"id":"qwen/qwen3.5-35b-a3b:thinking","name":"Qwen3.5 35B A3B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":260096,"input":260096,"output":65536},"cost":{"input":0.14,"output":1,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3.8-27b-swift":{"id":"qwen/qwen3.8-27b-swift","name":"Qwen 3.8 27B Swift","description":"Qwen 3.8 27B Swift is a finetune for fast responses, served by Aoru in NVFP4 with a 262,144-token context window. It supports image understanding, tool calling, and optional reasoning.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen-max":{"id":"qwen/qwen-max","name":"Qwen 2.5 Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"input":32000,"output":8192},"cost":{"input":1.5997,"output":6.392,"cache_read":0.79985},"canonical_model_id":"alibaba/qwen-max"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen 3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":65536},"cost":{"input":0.325,"output":1.95,"cache_read":0.0325,"cache_write":0.40625},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen/qwen3-235b-a22b-instruct-2507":{"id":"qwen/qwen3-235b-a22b-instruct-2507","name":"Qwen 3 235b A22B 2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.0875,"output":0.35,"cache_read":0.0175},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"qwen/qwen3.5-122b-a10b:thinking":{"id":"qwen/qwen3.5-122b-a10b:thinking","name":"Qwen3.5 122B A10B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.26,"output":2.08,"cache_read":0.103788},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":65536},"cost":{"input":0.19,"output":1.16,"cache_read":0.02,"cache_write":0.24},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3.5-flash:thinking":{"id":"qwen/qwen3.5-flash:thinking","name":"Qwen3.5 Flash Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.5-flash"},"qwen/qwen3.5-flash":{"id":"qwen/qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.5-flash"},"qwen/qwen3-235b-a22b":{"id":"qwen/qwen3-235b-a22b","name":"Qwen 3 235b A22B","description":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.0875,"output":0.35,"cache_read":0.0175},"canonical_model_id":"alibaba/qwen3-235b-a22b"},"qwen/qwen3-coder-flash":{"id":"qwen/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.3,"output":1.5,"cache_read":0.15},"canonical_model_id":"alibaba/qwen3-coder-flash"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":131072},"cost":{"input":0.14,"output":0.42,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3.8-max-prime":{"id":"qwen/qwen3.8-max-prime","name":"Qwen3.8 Max Prime","description":"High-throughput edition of Qwen3.8 Max for coding, professional work, multimodal understanding, and long-running agent workflows","family":"qwen3.8-max","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":131072},"cost":{"input":4,"output":12,"cache_read":0.5},"canonical_model_id":"alibaba/qwen3.8-max-prime"},"qwen/qwen3.5-plus:thinking":{"id":"qwen/qwen3.5-plus:thinking","name":"Qwen3.5 Plus Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"input":983616,"output":65536},"cost":{"input":0.4,"output":2.4,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen/qwen3.8-27b-uncensored":{"id":"qwen/qwen3.8-27b-uncensored","name":"Qwen 3.8 27B Uncensored","description":"Qwen 3.8 27B Uncensored is an NVFP4 open-weight multimodal model LoRA-tuned for fewer refusals across chat, coding, reasoning, tool use, and long-context work.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-07-29","last_updated":"2026-08-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":32768},"cost":{"input":1,"output":3,"cache_read":0.22}},"qwen/qwen3-30b-a3b":{"id":"qwen/qwen3-30b-a3b","name":"Qwen3 30B A3B","description":"Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":41000,"input":41000,"output":32768},"cost":{"input":0.1,"output":0.3,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"qwen/qwen3-vl-235b-a22b-thinking":{"id":"qwen/qwen3-vl-235b-a22b-thinking","name":"Qwen3 VL 235B A22B Thinking","description":"Qwen vision-language thinking model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-thinking"},"qwen/qwen3.7-plus:thinking":{"id":"qwen/qwen3.7-plus:thinking","name":"Qwen3.7 Plus Thinking","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"input":983616,"output":65536},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3.8-27b-kiwen":{"id":"qwen/qwen3.8-27b-kiwen","name":"Qwen 3.8 27B Kiwen","description":"Qwen 3.8 27B Kiwen is a creative finetune served by Aoru in NVFP4 with a 262,144-token context window. It supports image understanding, tool calling, and optional reasoning.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":65536},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen-turbo":{"id":"qwen/qwen-turbo","name":"Qwen Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-11-01","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":8192},"cost":{"input":0.04998,"output":0.2006,"cache_read":0.02499},"canonical_model_id":"alibaba/qwen-turbo"},"qwen/qwen3-next-80b-a3b-instruct":{"id":"qwen/qwen3-next-80b-a3b-instruct","name":"Qwen3 Next 80B A3B (Instruct)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":1.1,"cache_read":0.07},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":260096,"input":260096,"output":65536},"cost":{"input":0.25,"output":2,"cache_read":0.135},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3-next-80b-a3b-thinking":{"id":"qwen/qwen3-next-80b-a3b-thinking","name":"Qwen3 Next 80B A3B (Thinking)","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":0.15,"output":0.65,"cache_read":0.075},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"qwen/qwen3.6-27b:thinking":{"id":"qwen/qwen3.6-27b:thinking","name":"Qwen3.6 27B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.203,"output":2.24,"cache_read":0.1015},"canonical_model_id":"alibaba/qwen3.6-27b"},"qwen/qwen3-max-2026-01-23":{"id":"qwen/qwen3-max-2026-01-23","name":"Qwen3 Max 2026-01-23","description":"Qwen3 Max is Alibaba's flagship Qwen 3 reasoning model with native tool use (web search, web extractor, code interpreter) and a 256K context window.","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-01-26","last_updated":"2026-01-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":1.2002,"output":6.001,"cache_read":0.6001}},"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":260096,"input":260096,"output":65536},"cost":{"input":0.14,"output":1,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3.6-35b-a3b:thinking":{"id":"qwen/qwen3.6-35b-a3b:thinking","name":"Qwen3.6 35B A3B Thinking","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.112,"output":0.8,"cache_read":0.056},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3-30b-a3b-instruct-2507":{"id":"qwen/qwen3-30b-a3b-instruct-2507","name":"Qwen3 30B A3B Instruct 2507","description":"Qwen3-30B-A3B-Instruct-2507 is a 30.5B-parameter mixture-of-experts language model from Qwen, with 3.3B active parameters per inference. Significant improvements in general capabilities, including instruction following, logical reasoning, text comprehension, mathematics, science, coding and tool usage.","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-02-20","last_updated":"2025-02-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":0.2,"output":0.5,"cache_read":0.1}},"qwen/qwen3.8-27b-fimi":{"id":"qwen/qwen3.8-27b-fimi","name":"Qwen 3.8 27B Fimi","description":"Qwen 3.8 27B Fimi is a creative finetune served by Aoru in NVFP4 with a 262,144-token context window. It supports image understanding, tool calling, and optional reasoning.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3-32b":{"id":"qwen/qwen3-32b","name":"Qwen 3 32b","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":41000,"input":41000,"output":32768},"cost":{"input":0.1,"output":0.3,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3-32b"},"qwen/qwen3.8-27b-uncensored:thinking":{"id":"qwen/qwen3.8-27b-uncensored:thinking","name":"Qwen 3.8 27B Uncensored Thinking","description":"Qwen 3.8 27B Uncensored with thinking enabled for more deliberate creative work, coding, multimodal analysis, tool use, and long-context problem solving.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-07-29","last_updated":"2026-08-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":32768},"cost":{"input":1,"output":3,"cache_read":0.22}},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1001536,"input":1001536,"output":32768},"cost":{"input":0.15,"output":0.7,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.8-27b-synthia":{"id":"qwen/qwen3.8-27b-synthia","name":"Qwen 3.8 27B Synthia","description":"Qwen 3.8 27B Synthia is a text-only roleplay finetune served by Aoru in NVFP4 with a 262,144-token context window.","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3-coder-30b-a3b-instruct":{"id":"qwen/qwen3-coder-30b-a3b-instruct","name":"Qwen3 Coder 30B A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"qwen/qwen3.8-27b-novelist":{"id":"qwen/qwen3.8-27b-novelist","name":"Qwen 3.8 27B Novelist","description":"Qwen 3.8 27B Novelist is a creative finetune served by Aoru in NVFP4 with a 262,144-token context window. It supports image understanding, tool calling, and optional reasoning.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.39,"output":2.34,"cache_read":0.3},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen-long":{"id":"qwen/qwen-long","name":"Qwen Long 10M","description":"Alibaba's huge context window model. Takes in up to 10 million tokens, which is equivalent to dozens of books.","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-08-01","last_updated":"2024-08-01","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":10000000,"input":10000000,"output":8192},"cost":{"input":0.1003,"output":0.408,"cache_read":0.05015}},"qwen/qwen3.6-max-preview":{"id":"qwen/qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":245760,"input":245760,"output":65536},"cost":{"input":1.04,"output":6.24,"cache_read":0.52},"canonical_model_id":"alibaba/qwen3.6-max-preview"},"qwen/qwen2.5-coder-32b-instruct":{"id":"qwen/qwen2.5-coder-32b-instruct","name":"Qwen 2.5 Coder 32b","description":"Open coding-focused Qwen model for code generation, repair, and repository reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"input":32000,"output":8192},"cost":{"input":0.2006,"output":0.2006,"cache_read":0.1003},"canonical_model_id":"alibaba/qwen2.5-coder-32b-instruct"},"qwen/qwen3.5-397b-a17b:thinking":{"id":"qwen/qwen3.5-397b-a17b:thinking","name":"Qwen3.5 397B A17B Thinking","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.39,"output":2.34,"cache_read":0.3},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3.5-27b:thinking":{"id":"qwen/qwen3.5-27b:thinking","name":"Qwen3.5 27B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":260096,"input":260096,"output":65536},"cost":{"input":0.25,"output":2,"cache_read":0.135},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3-coder-next":{"id":"qwen/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.250251,"output":0.80293,"cache_read":0.079764},"canonical_model_id":"alibaba/qwen3-coder-next"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.203,"output":2.24,"cache_read":0.1015},"canonical_model_id":"alibaba/qwen3.6-27b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":1.475,"output":4.425,"cache_read":0.295,"cache_write":1.84375},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.8-27b-darkidol":{"id":"qwen/qwen3.8-27b-darkidol","name":"Qwen 3.8 27B DarkIdol","description":"Qwen 3.8 27B DarkIdol is a creative finetune served by Aoru in NVFP4 with a 262,144-token context window. It supports image understanding, tool calling, and optional reasoning.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3-coder":{"id":"qwen/qwen3-coder","name":"Qwen 3 Coder 480B","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"input":262000,"output":65536},"cost":{"input":0.13,"output":0.5,"cache_read":0.065}},"qwen/qwen3.8-max:thinking":{"id":"qwen/qwen3.8-max:thinking","name":"Qwen3.8 Max Thinking","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"input":991000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen/qwen-2.5-72b-instruct":{"id":"qwen/qwen-2.5-72b-instruct","name":"Qwen2.5 72B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2025-07-03","last_updated":"2025-07-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":8192},"cost":{"input":0.36,"output":0.4,"cache_read":0.1785}},"qwen/qwen3.8-max":{"id":"qwen/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"input":991000,"output":65536},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen/qwen3.8-27b-antislop":{"id":"qwen/qwen3.8-27b-antislop","name":"Qwen 3.8 27B Antislop","description":"Qwen 3.8 27B Antislop is a creative finetune served by Aoru in NVFP4 with a 262,144-token context window. It supports image understanding, tool calling, and optional reasoning.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3-max":{"id":"qwen/qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":1.2002,"output":6.001,"cache_read":0.6001},"canonical_model_id":"alibaba/qwen3-max"},"qwen/qwen3.8-27b-hemmingway":{"id":"qwen/qwen3.8-27b-hemmingway","name":"Qwen 3.8 27B Hemingway","description":"Qwen 3.8 27B Hemingway is an open-weight NVFP4 multimodal creative finetune for long-form prose, character dialogue, storytelling, and roleplay.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.25,"output":1.5,"cache_read":0.125}},"qwen/qwen3.8-27b-fable":{"id":"qwen/qwen3.8-27b-fable","name":"Qwen 3.8 27B Fable","description":"Qwen 3.8 27B Fable is an open-weight multimodal creative finetune for expressive dialogue, long-form storytelling, character work, and roleplay.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-08-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"input":524288,"output":32768},"cost":{"input":0.25,"output":1.5,"cache_read":0.125}},"qwen/qwen3-14b":{"id":"qwen/qwen3-14b","name":"Qwen 3 14b","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":41000,"input":41000,"output":32768},"cost":{"input":0.08,"output":0.24,"cache_read":0.04}},"qwen/qwen3.8-max-0902":{"id":"qwen/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"qwen/qwen3.5-omni-plus":{"id":"qwen/qwen3.5-omni-plus","name":"Qwen3.5 Omni Plus","description":"Qwen3.5 Omni Plus is Qwen's stronger general multimodal model. We verified live support for text prompts, images, audio files, and direct video URLs on Alibaba's chat-completions-compatible API. Alibaba describes Plus as a comprehensive evolution of Qwen3 Omni with support for over 10 hours of audio input.","family":"qwen3.5","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-03-30","last_updated":"2026-03-30","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"input":983616,"output":65536}},"inference-net/schematron-v2-small":{"id":"inference-net/schematron-v2-small","name":"Schematron V2 Small","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":4096},"cost":{"input":0.05,"output":0.23,"cache_read":0.025},"canonical_model_id":"inference-net/schematron-v2-small"},"inference-net/schematron-v2-turbo":{"id":"inference-net/schematron-v2-turbo","name":"Schematron V2 Turbo","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.03,"output":0.15,"cache_read":0.015},"canonical_model_id":"inference-net/schematron-v2-turbo"},"moonshotai/kimi-k2-instruct-0711":{"id":"moonshotai/kimi-k2-instruct-0711","name":"Kimi K2 0711","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.4,"output":1.8,"cache_read":0.2}},"moonshotai/Kimi-K2-Instruct-0905":{"id":"moonshotai/Kimi-K2-Instruct-0905","name":"Kimi K2 0905","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":100352},"cost":{"input":0.4,"output":1.8,"cache_read":0.2}},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.45,"output":1.95,"cache_read":0.045},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":943718},"cost":{"input":1.65,"output":8.25,"cache_read":0.165},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2.7-code-highspeed":{"id":"moonshotai/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code High-Speed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":1.9,"output":8,"cache_read":0.32},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshotai/kimi-k2.5:thinking":{"id":"moonshotai/kimi-k2.5:thinking","name":"Kimi K2.5 Thinking","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.375,"output":2,"cache_read":0.07},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/kimi-k2-instruct":{"id":"moonshotai/kimi-k2-instruct","name":"Kimi K2 Instruct","description":"Non-thinking Kimi K2 MoE instruction model for coding and tool-using agents","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":8192},"cost":{"input":0.4,"output":1.8,"cache_read":0.2},"canonical_model_id":"moonshotai/kimi-k2-instruct"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.6712,"output":3.35,"cache_read":0.18},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-latest":{"id":"moonshotai/kimi-latest","name":"Kimi Latest","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-03","last_updated":"2026-05-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":943718},"cost":{"input":1.65,"output":8.25,"cache_read":0.165}},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.375,"output":2,"cache_read":0.07},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/kimi-k2.6:thinking":{"id":"moonshotai/kimi-k2.6:thinking","name":"Kimi K2.6 Thinking","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.45,"output":1.95,"cache_read":0.045},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":98304},"cost":{"input":0.6,"output":2.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"MiniMaxAI/MiniMax-M1-80k":{"id":"MiniMaxAI/MiniMax-M1-80k","name":"MiniMax M1 80K","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-01-08","last_updated":"2025-06-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":131072},"cost":{"input":0.6052,"output":2.4225,"cache_read":0.3026}},"Sao10K/L3.1-70B-Euryale-v2.2":{"id":"Sao10K/L3.1-70B-Euryale-v2.2","name":"Llama 3.1 70B Euryale","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":20480,"input":20480,"output":16384},"cost":{"input":0.306,"output":0.357,"cache_read":0.153}},"Sao10K/L3.1-70B-Hanami-x1":{"id":"Sao10K/L3.1-70B-Hanami-x1","name":"Llama 3.1 70B Hanami","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"Sao10K/L3-8B-Stheno-v3.2":{"id":"Sao10K/L3-8B-Stheno-v3.2","name":"Sao10K Stheno 8b","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-11-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.2006,"output":0.2006,"cache_read":0.1003}},"Sao10K/L3.3-70B-Euryale-v2.3":{"id":"Sao10K/L3.3-70B-Euryale-v2.3","name":"Llama 3.3 70B Euryale","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":20480,"input":20480,"output":16384},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"poolside/laguna-s-2.1:thinking":{"id":"poolside/laguna-s-2.1:thinking","name":"Laguna S 2.1 Thinking","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.01},"canonical_model_id":"poolside/laguna-s-2.1"},"poolside/laguna-s-2.1":{"id":"poolside/laguna-s-2.1","name":"Laguna S 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.01},"canonical_model_id":"poolside/laguna-s-2.1"},"fireworks/ember-1":{"id":"fireworks/ember-1","name":"Ember 1","description":"Specialized model from Fireworks built on Kimi K3, producing shorter reasoning traces with approximately 40% fewer tokens while maintaining comparable quality","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"fireworks/ember-1"},"stepfun-ai/step-3.5-flash-2603":{"id":"stepfun-ai/step-3.5-flash-2603","name":"Step 3.5 Flash 2603","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.1,"output":0.3,"cache_read":0.05},"canonical_model_id":"stepfun/step-3.5-flash-2603"},"stepfun-ai/step-3.5-flash":{"id":"stepfun-ai/step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.1,"output":0.3,"cache_read":0.05},"canonical_model_id":"stepfun/step-3.5-flash"},"bytedance/doubao-seed-2.1-pro":{"id":"bytedance/doubao-seed-2.1-pro","name":"Doubao Seed 2.1 Pro","description":"Higher-capability model in the Doubao Seed 2.1 family for agentic coding, long-context analysis, complex instruction following, and productivity workflows. Supports a 256k context window and up to 128k output tokens. Note: privacy and logging guarantees may be limited.","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":128000},"cost":{"input":1,"output":5,"cache_read":0.5}},"bytedance/doubao-seed-character":{"id":"bytedance/doubao-seed-character","name":"Doubao Seed Character","description":"ByteDance's character-focused Doubao Seed model for roleplay, persona consistency, dialogue, and creative character interactions. It supports text and image input with a 128k context window. Requests route through ZenMux to ByteDance; ZenMux does not publish a model-API zero-retention or training guarantee, so avoid sensitive data.","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":false,"release_date":"2026-07-18","last_updated":"2026-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":32768},"cost":{"input":0.15,"output":0.37,"cache_read":0.03,"cache_write":0.003}},"bytedance/doubao-seed-2.1-turbo":{"id":"bytedance/doubao-seed-2.1-turbo","name":"Doubao Seed 2.1 Turbo","description":"Fast, lower-cost model in the Doubao Seed 2.1 family for everyday chat, coding assistance, document work, and high-throughput productivity tasks. Supports a 256k context window and up to 128k output tokens. Note: privacy and logging guarantees may be limited.","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":128000},"cost":{"input":0.5,"output":2.5,"cache_read":0.25}},"GalrionSoftworks/MN-LooseCannon-12B-v1":{"id":"GalrionSoftworks/MN-LooseCannon-12B-v1","name":"MN-LooseCannon-12B-v1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"MarinaraSpaghetti/NemoMix-Unleashed-12B":{"id":"MarinaraSpaghetti/NemoMix-Unleashed-12B","name":"NemoMix 12B Unleashed","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"mlabonne/NeuralDaredevil-8B-abliterated":{"id":"mlabonne/NeuralDaredevil-8B-abliterated","name":"Neural Daredevil 8B abliterated","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"input":8192,"output":8192},"cost":{"input":0.44,"output":0.44,"cache_read":0.22}},"EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2":{"id":"EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2","name":"EVA-Qwen2.5-32B-v0.2","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.799,"output":0.799,"cache_read":0.3995}},"EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2":{"id":"EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2","name":"EVA-Qwen2.5-72B-v0.2","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.799,"output":0.799,"cache_read":0.3995}},"EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.0":{"id":"EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.0","name":"EVA Llama 3.33 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":2.006,"output":2.006,"cache_read":1.003}},"EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.1":{"id":"EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.1","name":"EVA-LLaMA-3.33-70B-v0.1","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":2.006,"output":2.006,"cache_read":1.003}},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048000,"input":1048000,"output":32768},"cost":{"input":0.95,"output":4.05,"cache_read":0.16},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/inkling:thinking":{"id":"thinkingmachines/inkling:thinking","name":"Inkling Thinking","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048000,"input":1048000,"output":32768},"cost":{"input":0.95,"output":4.05,"cache_read":0.16},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/Inkling-Small:thinking":{"id":"thinkingmachines/Inkling-Small:thinking","name":"Inkling Small Thinking","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"input":524288,"output":32768},"cost":{"input":0.45,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/Inkling-Small":{"id":"thinkingmachines/Inkling-Small","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"input":524288,"output":32768},"cost":{"input":0.45,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"reflection/beam-1.0":{"id":"reflection/beam-1.0","name":"Beam 1.0 501B A23B","description":"Reflection's Apache 2.0 open-weight text model with 501.3B total and 23.4B active parameters, built for coding, reasoning and tool use. It supports a shared 1,048,576-token context window, up to 131,072 output tokens, JSON mode and structured output. Reasoning is always enabled with low, medium, high, xhigh and max effort; NanoGPT defaults to max. Knowledge cutoff: June 30, 2026.","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.5,"output":1.5,"cache_read":0.05}},"ornith-ai/ornith-1.5-35b-a3b":{"id":"ornith-ai/ornith-1.5-35b-a3b","name":"Ornith 1.5 35B","description":"Ornith 1.5 35B A3B is an open-weight mixture-of-experts model for agentic coding, tool use, image understanding, and long-context work. This variant disables thinking for faster direct responses.","family":"ornith","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-07-29","last_updated":"2026-08-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.05}},"ornith-ai/ornith-1.5-35b-a3b:thinking":{"id":"ornith-ai/ornith-1.5-35b-a3b:thinking","name":"Ornith 1.5 35B Thinking","description":"Ornith 1.5 35B A3B is an open-weight mixture-of-experts model for agentic coding, reasoning, tool use, image understanding, and long-context work. This variant enables thinking by default.","family":"ornith","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-07-29","last_updated":"2026-08-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.05}},"reka/reka-flash-3":{"id":"reka/reka-flash-3","name":"Reka Flash 3","description":"Reka's 21B reasoning model for chat and coding, served directly by Reka.","family":"reka","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":65536},"cost":{"input":0.105,"output":0.21}},"reka/reka-edge-2603":{"id":"reka/reka-edge-2603","name":"Reka Edge 2603","description":"Reka's compact multimodal model for visual understanding and tool use, served directly by Reka.","family":"reka","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"input":16384,"output":16384},"cost":{"input":0.105,"output":0.105}},"stepfun/step-3.7-flash:thinking":{"id":"stepfun/step-3.7-flash:thinking","name":"Step 3.7 Flash Thinking","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"stepfun/step-5-preview":{"id":"stepfun/step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":1000000},"cost":{"input":1,"output":2.7,"cache_read":0.05},"canonical_model_id":"stepfun/step-5-preview"},"lightonai/LightOnOCR-2-1B":{"id":"lightonai/LightOnOCR-2-1B","name":"LightOnOCR 2","description":"LightOnOCR 2 hosted by IONOS in Berlin, Germany. Zero data retention.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.1785,"output":0.3465}},"LatitudeGames/Wayfarer-Large-70B-Llama-3.3":{"id":"LatitudeGames/Wayfarer-Large-70B-Llama-3.3","name":"Llama 3.3 70B Wayfarer","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-02-20","last_updated":"2025-02-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.7,"output":0.7,"cache_read":0.35}},"Doctor-Shotgun/MS3.2-24B-Magnum-Diamond":{"id":"Doctor-Shotgun/MS3.2-24B-Magnum-Diamond","name":"MS3.2 24B Magnum Diamond","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"perceptron/perceptron-mk1":{"id":"perceptron/perceptron-mk1","name":"Perceptron Mk1","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.15,"output":1.5,"cache_read":0.075},"canonical_model_id":"perceptron/perceptron-mk1"},"perceptron/perceptron-mk1.5":{"id":"perceptron/perceptron-mk1.5","name":"Perceptron Mk1.5","description":"Perceptron Mk1.5 is a multimodal reasoning model for visual analysis and physical agents. It accepts text, images, video, and audio, supports tool calling and structured output, and can return spatial annotations such as points, boxes, polygons, and tracks.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":36864,"input":36864,"output":8192},"cost":{"input":0.15,"output":1.5}},"nothingiisreal/L3.1-70B-Celeste-V0.1-BF16":{"id":"nothingiisreal/L3.1-70B-Celeste-V0.1-BF16","name":"Llama 3.1 70B Celeste v0.1","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"soob3123/amoral-gemma3-27B-v2":{"id":"soob3123/amoral-gemma3-27B-v2","name":"Amoral Gemma3 27B v2","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-05-23","last_updated":"2025-05-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.3,"output":0.3,"cache_read":0.15}},"soob3123/Veiled-Calla-12B":{"id":"soob3123/Veiled-Calla-12B","name":"Veiled Calla 12B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-04-13","last_updated":"2025-04-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.3,"output":0.3,"cache_read":0.15}},"abacusai/Dracarys-72B-Instruct":{"id":"abacusai/Dracarys-72B-Instruct","name":"Llama 3.1 70B Dracarys 2","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-08-02","last_updated":"2025-08-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"heabsy/qwen3.8-27b-uncensored-fast":{"id":"heabsy/qwen3.8-27b-uncensored-fast","name":"Qwen 3.8 27B Uncensored Fast by Heabsy","description":"Heabsy's NVFP4 build of Qwen3.8-27B by Alibaba Cloud, with refusal removal based on huihui-ai's Huihui-Qwen3.8-27B-abliterated. Quantized and calibrated by Heabsy, with unpublished weights under the Heabsy Model License. Supports text and image input, tools, and optional reasoning capped at 16,384 tokens within the output budget. This named build is separate from Qwen 3.8 27B Uncensored.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"input":524288,"output":235929},"cost":{"input":1,"output":3,"cache_read":0.22}},"heabsy/cyberheabsy":{"id":"heabsy/cyberheabsy","name":"CyberHeabsy","description":"A low-refusal text model for authorized security research, roleplay, and creative writing. It preserves tool calling after refusal removal, supports optional reasoning, and has a 1M-token context window.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-09-27","last_updated":"2026-09-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":1.5,"output":4.5,"cache_read":0.275}},"mistralai/mistral-small-3.1-24b-instruct":{"id":"mistralai/mistral-small-3.1-24b-instruct","name":"Mistral Small 3.1 24B (2503)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":102400},"cost":{"input":0.1,"output":0.3,"cache_read":0.05},"canonical_model_id":"mistral/mistral-small-3.1-24b-instruct"},"mistralai/mistral-medium-3":{"id":"mistralai/mistral-medium-3","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.4,"output":2,"cache_read":0.2},"canonical_model_id":"mistral/mistral-medium-3"},"mistralai/mistral-medium-3.5":{"id":"mistralai/mistral-medium-3.5","name":"Mistral Medium 3.5","description":"Mistral Medium 3.5 is a 128B dense open-weights flagship model for instruction-following, reasoning, coding, long-horizon agentic work, tool use, structured output, and multimodal prompts. It supports a 256k context window and configurable reasoning effort.","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":1.5,"output":7.5,"cache_read":0.75}},"mistralai/mixtral-8x22b-instruct-v0.1":{"id":"mistralai/mixtral-8x22b-instruct-v0.1","name":"Mixtral 8x22B","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mixtral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":52428},"cost":{"input":2,"output":6,"cache_read":0.2}},"mistralai/ministral-14b-instruct-2512":{"id":"mistralai/ministral-14b-instruct-2512","name":"Ministral 3 14B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.05}},"mistralai/mistral-small-4-119b-2603:thinking":{"id":"mistralai/mistral-small-4-119b-2603:thinking","name":"Mistral Small 4 119B Thinking","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.4,"output":1.4,"cache_read":0.2}},"mistralai/devstral-small-2505":{"id":"mistralai/devstral-small-2505","name":"Mistral Devstral Small 2505","description":"OpenHands+Devstral is 100% local 100% open, and is SOTA for the category on SWE-Bench Verified: 46.8% accuracy.","family":"devstral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-08-02","last_updated":"2025-08-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.06,"output":0.06,"cache_read":0.03}},"mistralai/mistral-large":{"id":"mistralai/mistral-large","name":"Mistral Large 2411","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-11-30","release_date":"2024-02-26","last_updated":"2024-02-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":102400},"cost":{"input":2.006,"output":6.001,"cache_read":0.2},"canonical_model_id":"mistral/mistral-large"},"mistralai/mistral-small-3.2-24b-instruct":{"id":"mistralai/mistral-small-3.2-24b-instruct","name":"Mistral Small 3.2 24B (2506)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.2,"output":0.4,"cache_read":0.1},"canonical_model_id":"mistral/mistral-small-3.2-24b-instruct"},"mistralai/mistral-saba":{"id":"mistralai/mistral-saba","name":"Mistral Saba","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-02-17","last_updated":"2025-02-17","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":26214},"cost":{"input":0.1989,"output":0.595,"cache_read":0.09945},"canonical_model_id":"mistral/mistral-saba"},"mistralai/devstral-2-123b-instruct-2512":{"id":"mistralai/devstral-2-123b-instruct-2512","name":"Devstral 2 123B","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.4,"output":1.4,"cache_read":0.2}},"mistralai/ministral-8b-2512":{"id":"mistralai/ministral-8b-2512","name":"Ministral 8B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.15,"output":0.15,"cache_read":0.075},"canonical_model_id":"mistral/ministral-8b-2512"},"mistralai/mistral-nemo-instruct-2407":{"id":"mistralai/mistral-nemo-instruct-2407","name":"Mistral Nemo","description":"12B parameter model with multilingual support.","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.1003,"output":0.1207,"cache_read":0.05015}},"mistralai/mistral-medium-3.1":{"id":"mistralai/mistral-medium-3.1","name":"Mistral Medium 3.1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-06-30","release_date":"2025-08-12","last_updated":"2025-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.4,"output":2,"cache_read":0.2},"canonical_model_id":"mistral/mistral-medium-3.1"},"mistralai/ministral-3b-2512":{"id":"mistralai/ministral-3b-2512","name":"Ministral 3B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.1,"output":0.1,"cache_read":0.05},"canonical_model_id":"mistral/ministral-3b-2512"},"mistralai/mistral-small-4-119b-2603":{"id":"mistralai/mistral-small-4-119b-2603","name":"Mistral Small 4 119B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.4,"output":1.4,"cache_read":0.2}},"mistralai/mistral-medium-3.5:thinking":{"id":"mistralai/mistral-medium-3.5:thinking","name":"Mistral Medium 3.5 Thinking","description":"Mistral Medium 3.5 with reasoning enabled by default (reasoning_effort=high), for complex coding, agentic, and multi-step reasoning prompts.","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":1.5,"output":7.5,"cache_read":0.75}},"mistralai/ministral-14b-2512":{"id":"mistralai/ministral-14b-2512","name":"Ministral 14B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.2,"output":0.2,"cache_read":0.1},"canonical_model_id":"mistral/ministral-14b-2512"},"mistralai/mistral-large-4":{"id":"mistralai/mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"input":524288,"output":262144},"cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"mistralai/codestral-2508":{"id":"mistralai/codestral-2508","name":"Codestral 2508","description":"Mistral code model for low-latency completion, fill-in-the-middle, and code generation","family":"codestral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.15},"canonical_model_id":"mistral/codestral-2508"},"mistralai/mistral-small-24b-instruct-2501":{"id":"mistralai/mistral-small-24b-instruct-2501","name":"Mistral Small 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-01-30","last_updated":"2025-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.1155,"output":0.3465},"canonical_model_id":"mistral/mistral-small-24b-instruct-2501"},"abliteration-ai/abliterated-model":{"id":"abliteration-ai/abliterated-model","name":"Abliterated Model","description":"Abliteration.ai's multimodal reasoning model supports text and image input, structured output, automatic prompt caching, and a 262K-token context window.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":false,"structured_output":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":262134},"cost":{"input":1,"output":3,"cache_read":0.1}},"abliteration-ai/abliterated-model-large-v2":{"id":"abliteration-ai/abliterated-model-large-v2","name":"Abliterated Model Large V2","description":"Abliteration.ai's default large text reasoning model is derived from GLM-5.3 for harder reasoning and evaluation workloads, with automatic prompt caching and a one-million-token context window.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-08-31","last_updated":"2026-08-31","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":999990},"cost":{"input":3,"output":5,"cache_read":0.3}},"abliteration-ai/abliterated-model-large":{"id":"abliteration-ai/abliterated-model-large","name":"Abliterated Model Large","description":"Abliteration.ai's large text reasoning model is derived from GLM-5.2 and supports native tool calling, structured output, automatic prompt caching, and a one-million-token context window.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":999990},"cost":{"input":3,"output":5,"cache_read":0.3}},"minimax/minimax-m2.7-turbo":{"id":"minimax/minimax-m2.7-turbo","name":"MiniMax M2.7 Turbo","description":"Efficient MiniMax model for quick assistance, coding, and routine automation","family":"minimax-m2.7","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"input":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.06}},"minimax/minimax-m2-her":{"id":"minimax/minimax-m2-her","name":"MiniMax M2-her","description":"MiniMax M2 variant tuned for conversational and character-driven agent interactions","family":"minimax","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-01-23","last_updated":"2026-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65532,"input":65532,"output":2048},"cost":{"input":0.302,"output":1.207,"cache_read":0.151},"canonical_model_id":"minimax/MiniMax-M2-Her"},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"input":204800,"output":131072},"cost":{"input":0.21,"output":0.84,"cache_read":0.042},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":131072},"cost":{"input":0.17,"output":1.53,"cache_read":0.085},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-latest":{"id":"minimax/minimax-latest","name":"MiniMax Latest","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"release_date":"2026-05-03","last_updated":"2026-05-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"input":512000,"output":80000},"cost":{"input":0.23,"output":0.96,"cache_read":0.05}},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":131072},"cost":{"input":0.33,"output":1.32,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-m3.1-flash":{"id":"minimax/minimax-m3.1-flash","name":"MiniMax M3.1 Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":524288},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"minimax/minimax-m3:thinking":{"id":"minimax/minimax-m3:thinking","name":"MiniMax M3 Thinking","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"input":512000,"output":80000},"cost":{"input":0.23,"output":0.96,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"input":512000,"output":80000},"cost":{"input":0.23,"output":0.96,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-01":{"id":"minimax/minimax-01","name":"MiniMax 01","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-01-15","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000192,"input":1000192,"output":16384},"cost":{"input":0.1394,"output":1.122,"cache_read":0.0697}},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"input":204800,"output":131072},"cost":{"input":0.27,"output":1.08,"cache_read":0.027},"canonical_model_id":"minimax/MiniMax-M2.5"},"huihui-ai/Qwen2.5-32B-Instruct-abliterated":{"id":"huihui-ai/Qwen2.5-32B-Instruct-abliterated","name":"Qwen 2.5 32B Abliterated","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-01-06","last_updated":"2025-01-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.7,"output":0.7,"cache_read":0.35}},"huihui-ai/Llama-3.3-70B-Instruct-abliterated":{"id":"huihui-ai/Llama-3.3-70B-Instruct-abliterated","name":"Llama 3.3 70B Instruct abliterated","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-08-08","last_updated":"2025-08-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.7,"output":0.7,"cache_read":0.35}},"huihui-ai/DeepSeek-R1-Distill-Qwen-32B-abliterated":{"id":"huihui-ai/DeepSeek-R1-Distill-Qwen-32B-abliterated","name":"DeepSeek R1 Qwen Abliterated","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-01-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":1.4,"output":1.4,"cache_read":0.7}},"huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated":{"id":"huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated","name":"DeepSeek R1 Llama 70B Abliterated","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-01-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.7,"output":0.7,"cache_read":0.35}},"prism-ml/ternary-bonsai-2-27b":{"id":"prism-ml/ternary-bonsai-2-27b","name":"Ternary Bonsai 2 27B","description":"Ternary Bonsai 2 27B is a 27B-parameter reasoning model from PrismML derived from Qwen3.8-27B. It supports coding, mathematics, tool calling, and image understanding with a 262K-token context window.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","xhigh"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.075,"output":0.5,"cache_read":0.0375}},"NeverSleep/Lumimaid-v0.2-70B":{"id":"NeverSleep/Lumimaid-v0.2-70B","name":"Lumimaid v0.2","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":1,"output":1.5,"cache_read":0.5}},"inflatebot/MN-12B-Mag-Mell-R1":{"id":"inflatebot/MN-12B-Mag-Mell-R1","name":"Mag Mell R1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"featherless-ai/Qwerky-72B":{"id":"featherless-ai/Qwerky-72B","name":"Qwerky 72B","description":"General-purpose chat model for instruction following, writing, and analysis","family":"qwerky","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-03-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"input":32000,"output":8192},"cost":{"input":0.5,"output":0.5,"cache_read":0.25}},"elorian/glimpse":{"id":"elorian/glimpse","name":"Elorian Glimpse","description":"Glimpse is Elorian's visual reasoning model for understanding images and video and answering in text. It can analyze charts, count objects, follow events over time, and reason about spatial relationships and physical cause and effect. It supports tool calling and structured output across a shared 262,144-token context window, with up to 256,000 output tokens within the remaining context.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":256000},"cost":{"input":0.63,"output":3.15,"cache_read":0.063}},"liquid/lfm-2.5-2.6b":{"id":"liquid/lfm-2.5-2.6b","name":"LFM2.5 2.6B","description":"Liquid AI's compact 2.6B reasoning model for agent workflows, data extraction, RAG, and long-context processing. It supports tool calling and structured output, but Liquid advises against using it for agentic coding. Warning: prompts and responses may be logged and used for model training or service improvement; do not send sensitive data.","family":"liquid","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":32768},"cost":{"input":0.1,"output":0.2,"cache_read":0.05}},"liquid/lfm-2.5-24b-a2b":{"id":"liquid/lfm-2.5-24b-a2b","name":"LFM2.5 24B A2B","description":"Liquid AI's text-only mixture-of-experts model for reasoning and tool-using agents, with 24B total and 2.45B active parameters. It always thinks before answering and supports structured output. Pair it with retrieval for knowledge-heavy tasks. Open weights use the LFM Open License v1.0, with a $10M annual-revenue commercial threshold; hosted API use is subject to separate service terms that permit retention and use of prompts and responses for model training.","family":"liquid","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-10-14","last_updated":"2026-10-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.05,"output":0.7,"cache_read":0.025}},"upstage/solar-mini4:thinking":{"id":"upstage/solar-mini4:thinking","name":"Solar Mini 4 Thinking","description":"Upstage's compact, cost-efficient model for agentic use","family":"solar-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"input":524288,"output":131072},"cost":{"input":0.05,"output":0.2,"cache_read":0.005},"canonical_model_id":"upstage/solar-mini4"},"upstage/solar-pro4:thinking":{"id":"upstage/solar-pro4:thinking","name":"Solar Pro 4 Thinking","description":"Upstage's flagship model, specialized for agentic use","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"input":524288,"output":131072},"cost":{"input":0.03,"output":0.12,"cache_read":0.006},"canonical_model_id":"upstage/solar-pro4"},"upstage/solar-pro-3":{"id":"upstage/solar-pro-3","name":"Solar Pro 3","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"input":131072,"output":117964},"cost":{"input":0.15,"output":0.6,"cache_read":0.015}},"upstage/solar-pro4":{"id":"upstage/solar-pro4","name":"Solar Pro 4","description":"Upstage's flagship model, specialized for agentic use","family":"solar-pro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"input":524288,"output":131072},"cost":{"input":0.03,"output":0.12,"cache_read":0.006},"canonical_model_id":"upstage/solar-pro4"},"upstage/solar-mini4":{"id":"upstage/solar-mini4","name":"Solar Mini 4","description":"Upstage's compact, cost-efficient model for agentic use","family":"solar-mini","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"input":524288,"output":131072},"cost":{"input":0.05,"output":0.2,"cache_read":0.005},"canonical_model_id":"upstage/solar-mini4"},"ibm-granite/granite-4.2-8b":{"id":"ibm-granite/granite-4.2-8b","name":"Granite 4.2 8B","description":"IBM Granite 4.2 8B is an Apache 2.0-licensed dense model with native step-by-step reasoning and specialized training for agentic work. It can plan before acting, sequence tools, navigate codebases, work in terminals, and verify results across coding, search, mathematics, science, and complex instruction-following tasks.","family":"granite","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-08-31","last_updated":"2026-08-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":117964},"cost":{"input":0.1,"output":0.15,"cache_read":0.05}},"cohere/command-r-plus-08-2024":{"id":"cohere/command-r-plus-08-2024","name":"Cohere: Command R+","description":"Cohere's RAG workhorse for long-context enterprise search and tool use","family":"command-r","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":4096},"cost":{"input":2.856,"output":14.246,"cache_read":1.428},"canonical_model_id":"cohere/command-r-plus-08-2024"},"failspy/Meta-Llama-3-70B-Instruct-abliterated-v3.5":{"id":"failspy/Meta-Llama-3-70B-Instruct-abliterated-v3.5","name":"Llama 3 70B abliterated","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"input":8192,"output":8192},"cost":{"input":0.7,"output":0.7,"cache_read":0.35}},"xiaomi/mimo-v2.5-pro":{"id":"xiaomi/mimo-v2.5-pro","name":"MiMo V2.5 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.6-flash-uncensored":{"id":"xiaomi/mimo-v2.6-flash-uncensored","name":"MiMo V2.6 Flash Uncensored","description":"A text-only, lower-refusal MiMo V2.6 Flash finetune served by Heabsy with a 1M-token context window, separate reasoning output, and tool calling.","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.5,"output":1.5,"cache_read":0.1995}},"xiaomi/mimo-v2.6-pro-ultraspeed":{"id":"xiaomi/mimo-v2.6-pro-ultraspeed","name":"MiMo V2.6 Pro UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":4.35,"output":8.7,"cache_read":0.036,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"xiaomi/mimo-v2.5-pro:thinking":{"id":"xiaomi/mimo-v2.5-pro:thinking","name":"MiMo V2.5 Pro Thinking","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2.5:thinking":{"id":"xiaomi/mimo-v2.5:thinking","name":"MiMo V2.5 Thinking","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2.6-flash-uncensored:thinking":{"id":"xiaomi/mimo-v2.6-flash-uncensored:thinking","name":"MiMo V2.6 Flash Uncensored Thinking","description":"MiMo V2.6 Flash Uncensored with maximum thinking enabled. Supports a 1M-token context window and tool calling.","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-09-26","last_updated":"2026-09-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.5,"output":1.5,"cache_read":0.1995}},"xiaomi/mimo-v2.6-flash":{"id":"xiaomi/mimo-v2.6-flash","name":"MiMo V2.6 Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo V2.6 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"xiaomi/mimo-v2.6-flash-abliterated":{"id":"xiaomi/mimo-v2.6-flash-abliterated","name":"MiMo V2.6 Flash Abliterated","description":"A MiMo V2.6 Flash finetune with refusal-direction ablation, image input, a 1M-token context window, separate reasoning output, and tool calling.","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.5,"output":1.5,"cache_read":0.275}},"unbiased/pareto-26.9":{"id":"unbiased/pareto-26.9","name":"Pareto 26.9","description":"Pareto 26.9 is an experimental multimodal composite model from Unbiased AI for research, coding, agentic workflows, and general-purpose tasks. It runs several frontier and open models for each request and returns a single selected or synthesized answer. This experimental release supports text and image input, tool calling, a 262K-token context window, and up to 131K output tokens. Its underlying model mix and behavior may change as it is evaluated. Request content may be retained for abuse prevention and security monitoring for up to 30 days, and is not used for model training without explicit consent.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.25}},"unbiased/pareto-26.10":{"id":"unbiased/pareto-26.10","name":"Pareto 26.10","description":"Pareto 26.10 is Unbiased AI's multimodal composite model for research, coding, and agentic workflows. It accepts text and images, supports tool calling, and provides a 1M-token context window with up to 131K output tokens. This is the stable release of Pareto 26.10. Tool selection is automatic; forced tool selection, structured output, and adjustable reasoning controls are not supported.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.8,"output":3.2,"cache_read":0.03}},"pamanseau/OpenReasoning-Nemotron-32B":{"id":"pamanseau/OpenReasoning-Nemotron-32B","name":"OpenReasoning Nemotron 32B","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.05}},"NousResearch/hermes-4-405b:thinking":{"id":"NousResearch/hermes-4-405b:thinking","name":"Hermes 4 Large (Thinking)","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"hermes","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.15},"canonical_model_id":"nousresearch/hermes-4-405b"},"NousResearch/hermes-4-405b":{"id":"NousResearch/hermes-4-405b","name":"Hermes 4 Large","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"hermes","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.15},"canonical_model_id":"nousresearch/hermes-4-405b"},"NousResearch/hermes-3-llama-3.1-70b":{"id":"NousResearch/hermes-3-llama-3.1-70b","name":"Hermes 3 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-18","last_updated":"2024-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.408,"output":0.408,"cache_read":0.204},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-70b"},"moondream/moondream3-preview":{"id":"moondream/moondream3-preview","name":"Moondream 3 Preview","description":"Moondream 3 Preview is the original 9B mixture-of-experts vision-language checkpoint with 2B active parameters, for visual question answering, captioning, OCR, and visual reasoning. It accepts text or one image with optional reasoning, a 32,000-token context, and up to 4,096 output tokens.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"input":32000,"output":4096},"cost":{"input":0.3,"output":2.5}},"moondream/moondream3.1-9b-a2b":{"id":"moondream/moondream3.1-9b-a2b","name":"Moondream 3.1 9B A2B","description":"Moondream 3.1 is the successor to Moondream 3 Preview: a distinct 9B mixture-of-experts vision-language checkpoint with 2B active parameters, for visual question answering, captioning, OCR, and visual reasoning. It accepts text or one image with optional reasoning, a 32,000-token context, and up to 4,096 output tokens.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"input":32000,"output":4096},"cost":{"input":0.3,"output":2.5}},"THUDM/GLM-4-9B-0414":{"id":"THUDM/GLM-4-9B-0414","name":"GLM 4 9B 0414","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-04-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"input":32000,"output":8000},"cost":{"input":0.2,"output":0.2,"cache_read":0.1}},"THUDM/GLM-Z1-9B-0414":{"id":"THUDM/GLM-Z1-9B-0414","name":"GLM Z1 9B 0414","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm-z","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-04-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"input":32000,"output":8000},"cost":{"input":0.2,"output":0.2,"cache_read":0.1}},"THUDM/GLM-4-32B-0414":{"id":"THUDM/GLM-4-32B-0414","name":"GLM 4 32B 0414","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-04-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.2,"output":0.2,"cache_read":0.1}},"x-ai/grok-4.20-multi-agent":{"id":"x-ai/grok-4.20-multi-agent","name":"Grok 4.20 Multi-Agent","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"input":2000000,"output":131072},"cost":{"input":1.25,"output":2.5,"cache_read":0.2}},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":900000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"x-ai/grok-4.7":{"id":"x-ai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"input":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"x-ai/grok-build-0.1":{"id":"x-ai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":230400},"cost":{"input":1,"output":2,"cache_read":0.2},"canonical_model_id":"xai/grok-build-0.1"},"x-ai/grok-4.5":{"id":"x-ai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"input":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.3},"canonical_model_id":"xai/grok-4.5"},"x-ai/grok-latest":{"id":"x-ai/grok-latest","name":"Grok Latest","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-03","last_updated":"2026-05-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"input":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5}},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"input":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"x-ai/grok-4.20":{"id":"x-ai/grok-4.20","name":"Grok 4.20","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"input":2000000,"output":131072},"cost":{"input":1.25,"output":2.5,"cache_read":0.2}},"deepcogito/cogito-v1-preview-qwen-32B":{"id":"deepcogito/cogito-v1-preview-qwen-32B","name":"Cogito v1 Preview Qwen 32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-05-10","last_updated":"2025-05-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":32768},"cost":{"input":1.8,"output":1.8,"cache_read":0.9}},"google/gemini-pro-latest":{"id":"google/gemini-pro-latest","name":"Gemini Pro Latest","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-29","last_updated":"2026-03-29","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":0.375}},"google/gemma-4-26b-a4b-it:thinking":{"id":"google/gemma-4-26b-a4b-it:thinking","name":"Gemma 4 26B A4B Thinking","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.09,"output":0.3,"cache_read":0.05},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-flash-lite-latest":{"id":"google/gemini-flash-lite-latest","name":"Gemini Flash Lite Latest","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.08333},"canonical_model_id":"google/gemini-flash-lite-latest"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash (Preview)","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-3.1-pro-preview-low":{"id":"google/gemini-3.1-pro-preview-low","name":"Gemini 3.1 Pro (Preview Low)","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-02-21","last_updated":"2026-02-21","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":0.375}},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro (Preview Custom Tools)","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemini-3.5-flash-thinking":{"id":"google/gemini-3.5-flash-thinking","name":"Gemini 3.5 Flash Thinking","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro (Preview)","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.8-flash"},"google/diffusiongemma":{"id":"google/diffusiongemma","name":"DiffusionGemma","description":"DiffusionGemma is a high-speed diffusion-based version of Gemma 4 26B A4B. It supports optional reasoning and a 262,144-token context window.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","xhigh"]}],"tool_call":false,"structured_output":false,"release_date":"2026-09-19","last_updated":"2026-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.05,"output":0.15,"cache_read":0.025}},"google/gemini-3.1-pro-preview-high":{"id":"google/gemini-3.1-pro-preview-high","name":"Gemini 3.1 Pro (Preview High)","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-02-21","last_updated":"2026-02-21","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":0.375}},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemma-4-26b-a4b-it-cybersecurity":{"id":"google/gemma-4-26b-a4b-it-cybersecurity","name":"Gemma 4 26B A4B Cybersecurity","description":"Gemma 4 26B A4B Cybersecurity is a cybersecurity-focused variant based on the uncensored model, with provider moderation for illegal activities. It supports optional reasoning, image understanding, tool calling, and a 262,144-token context window.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-19","last_updated":"2026-09-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1056,"output":0.3344,"cache_read":0.0528}},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":0.1875},"canonical_model_id":"google/gemini-4-argon"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.09,"output":0.3,"cache_read":0.05},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-flash-latest":{"id":"google/gemini-flash-latest","name":"Gemini Flash Latest","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-flash-latest"},"google/gemini-3-flash-preview-thinking":{"id":"google/gemini-3-flash-preview-thinking","name":"Gemini 3 Flash Thinking","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemma4-31b-splituntied":{"id":"google/gemma4-31b-splituntied","name":"Gemma 4 31B Split-Untied","description":"Blazed-Forge's Split-Untied is a text-only Gemma 4 31B community finetune with an untied BF16 output head, built for creative writing, roleplay, expressive dialogue, and tool use.","family":"gemma","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.3,"cache_read":0.05}},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemma-4-31b-it:thinking":{"id":"google/gemma-4-31b-it:thinking","name":"Gemma 4 31B Thinking","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.1,"output":0.34,"cache_read":0.1},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.14,"output":0.4,"cache_read":0.05},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.6-flash"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Tencent Hy4 Preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Tencent Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":128000},"cost":{"input":0.066,"output":0.26,"cache_read":0.029},"canonical_model_id":"tencent/hy3"},"TEE/glm-5.3-flash":{"id":"TEE/glm-5.3-flash","name":"GLM 5.3 Flash TEE","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"TEE/glm-5.2:thinking":{"id":"TEE/glm-5.2:thinking","name":"GLM 5.2 Thinking TEE","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.7},"canonical_model_id":"zhipuai/glm-5.2"},"TEE/nemotron-3.5-lightning":{"id":"TEE/nemotron-3.5-lightning","name":"Nvidia Nemotron 3.5 Lightning TEE","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.08,"output":0.2,"cache_read":0.04},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"TEE/qwen3.6-35b-a3b":{"id":"TEE/qwen3.6-35b-a3b","name":"Qwen3.6 35B A3B TEE","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":262144},"cost":{"input":0.2,"output":1.27,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"TEE/kimi-k2.6":{"id":"TEE/kimi-k2.6","name":"Kimi K2.6 TEE","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":1.5,"output":5.25,"cache_read":0.375},"canonical_model_id":"moonshotai/kimi-k2.6"},"TEE/kimi-k3":{"id":"TEE/kimi-k3","name":"Kimi K3 TEE","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":65535},"cost":{"input":3,"output":15,"cache_read":1.5},"canonical_model_id":"moonshotai/kimi-k3"},"TEE/muse-glimmer-30b":{"id":"TEE/muse-glimmer-30b","name":"Muse Glimmer 30B TEE","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":131072},"cost":{"input":0.35,"output":1.5,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"TEE/gemma4-31b:thinking":{"id":"TEE/gemma4-31b:thinking","name":"Gemma 4 31B Thinking TEE","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-05-02","last_updated":"2026-05-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.4,"output":1,"cache_read":0.4}},"TEE/glm-5.3":{"id":"TEE/glm-5.3","name":"GLM 5.3 TEE","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"TEE/qwen3.5-27b":{"id":"TEE/qwen3.5-27b","name":"Qwen3.5 27B TEE","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.3,"output":2.4,"cache_read":0.15},"canonical_model_id":"alibaba/qwen3.5-27b"},"TEE/glm-5.1":{"id":"TEE/glm-5.1","name":"GLM 5.1 TEE","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"input":202752,"output":65535},"cost":{"input":0.7,"output":2.5,"cache_read":0.18},"canonical_model_id":"zhipuai/glm-5.1"},"TEE/gemma-4-26b-a4b-uncensored":{"id":"TEE/gemma-4-26b-a4b-uncensored","name":"Gemma 4 26B A4B Uncensored TEE","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-05-23","last_updated":"2026-05-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":65536},"cost":{"input":0.15,"output":0.7,"cache_read":0.075}},"TEE/llama3-3-70b":{"id":"TEE/llama3-3-70b","name":"Llama 3.3 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-03","last_updated":"2025-07-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":1.75,"output":2.75,"cache_read":1.75}},"TEE/qwen3.8-27b":{"id":"TEE/qwen3.8-27b","name":"Qwen3.8 27B TEE","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":262144},"cost":{"input":0.4,"output":3,"cache_read":0.15},"canonical_model_id":"alibaba/qwen3.8-27b"},"TEE/kimi-k2.7-code":{"id":"TEE/kimi-k2.7-code","name":"Kimi K2.7 Code TEE","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"TEE/qwen3.5-397b-a17b":{"id":"TEE/qwen3.5-397b-a17b","name":"Qwen3.5 397B A17B TEE","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.55,"output":3.5,"cache_read":0.275},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"TEE/deepseek-v3.2":{"id":"TEE/deepseek-v3.2","name":"DeepSeek V3.2 TEE","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"input":164000,"output":65536},"cost":{"input":0.5,"output":1,"cache_read":0.25},"canonical_model_id":"deepseek/deepseek-v3.2"},"TEE/gpt-oss-120b":{"id":"TEE/gpt-oss-120b","name":"GPT-OSS 120B TEE","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":16384},"cost":{"input":2,"output":2,"cache_read":2},"canonical_model_id":"openai/gpt-oss-120b"},"TEE/gemma-4-31b-it":{"id":"TEE/gemma-4-31b-it","name":"Gemma 4 31B IT TEE","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":262144},"cost":{"input":0.15,"output":0.46,"cache_read":0.075},"canonical_model_id":"google/gemma-4-31b-it"},"TEE/qwen3.6-27b":{"id":"TEE/qwen3.6-27b","name":"Qwen3.6 27B TEE","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.32,"output":2.7,"cache_read":0.16},"canonical_model_id":"alibaba/qwen3.6-27b"},"TEE/glm-5.2":{"id":"TEE/glm-5.2","name":"GLM 5.2 TEE","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.7},"canonical_model_id":"zhipuai/glm-5.2"},"TEE/deepseek-v4.1-flash":{"id":"TEE/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash TEE","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.65,"output":1.45,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"TEE/qwen2.5-vl-72b-instruct":{"id":"TEE/qwen2.5-vl-72b-instruct","name":"Qwen2.5 VL 72B TEE","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-02-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.7,"output":0.7,"cache_read":0.35}},"TEE/gemma4-31b":{"id":"TEE/gemma4-31b","name":"Gemma 4 31B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-04-04","last_updated":"2026-04-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.4,"output":1,"cache_read":0.4}},"TEE/glm-5.1-thinking":{"id":"TEE/glm-5.1-thinking","name":"GLM 5.1 Thinking TEE","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"input":202752,"output":65535},"cost":{"input":0.7,"output":2.5,"cache_read":0.18},"canonical_model_id":"zhipuai/glm-5.1"},"sakana/fugu-max":{"id":"sakana/fugu-max","name":"Fugu Max","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"sakana/fugu-max"},"sakana/fugu-ultra-v1.1":{"id":"sakana/fugu-ultra-v1.1","name":"Fugu Ultra v1.1","description":"Sakana AI's upgraded Fugu Ultra release with stronger coding, agentic task execution, and advanced reasoning through dynamic orchestration of frontier models.","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":16384},"cost":{"input":5,"output":30,"cache_read":0.5}},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":16384},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"sakana/fugu-ultra"},"openai/gpt-luna-latest":{"id":"openai/gpt-luna-latest","name":"GPT Luna Latest","description":"Compatibility alias that routes to GPT 5.6 Luna, the latest supported GPT Luna model.","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125}},"openai/o3-mini-high":{"id":"openai/o3-mini-high","name":"OpenAI o3-mini (High)","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2024-01-01","last_updated":"2025-01-31","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55}},"openai/o4-mini":{"id":"openai/o4-mini","name":"OpenAI o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o4-mini"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/o3-pro-2025-06-10":{"id":"openai/o3-pro-2025-06-10","name":"OpenAI o3-pro (2025-06-10)","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2024-01-01","last_updated":"2025-06-10","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":22,"output":88,"cache_read":11}},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT 5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-5.6-luna-pro":{"id":"openai/gpt-5.6-luna-pro","name":"GPT 5.6 Luna Pro","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT 4.1 Nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"input":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-sol-latest":{"id":"openai/gpt-sol-latest","name":"GPT Sol Latest","description":"Compatibility alias that routes to GPT 5.6 Sol, the latest supported GPT Sol model.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5}},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT 4.1 Mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"input":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT 5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT 5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":15,"output":120,"cache_read":1.5},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-astra-latest":{"id":"openai/gpt-astra-latest","name":"GPT Astra Latest","description":"Compatibility alias that routes to GPT 6 Astra, the latest supported GPT Astra model.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5}},"openai/gpt-6-luna-pro":{"id":"openai/gpt-6-luna-pro","name":"GPT 6 Luna Pro","description":"GPT-6 Luna Pro uses the same underlying model as GPT-6 Luna with Pro reasoning mode enabled for higher-quality responses on complex tasks.","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125}},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT 6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT 5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT 5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.1-2025-11-13":{"id":"openai/gpt-5.1-2025-11-13","name":"GPT-5.1 (2025-11-13)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2024-01-01","last_updated":"2025-11-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"openai/o3-mini":{"id":"openai/o3-mini","name":"OpenAI o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"input":16385,"output":4096},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT 4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"input":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-latest":{"id":"openai/gpt-latest","name":"GPT Latest","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-29","last_updated":"2026-03-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5}},"openai/gpt-chat-latest":{"id":"openai/gpt-chat-latest","name":"GPT Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-03","last_updated":"2026-05-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5}},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT 5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-4o-2024-08-06":{"id":"openai/gpt-4o-2024-08-06","name":"GPT-4o (2024-08-06)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-08-06"},"openai/o3-mini-low":{"id":"openai/o3-mini-low","name":"OpenAI o3-mini (Low)","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-01-31","last_updated":"2025-01-31","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55}},"openai/gpt-6-astra-pro":{"id":"openai/gpt-6-astra-pro","name":"GPT 6 Astra Pro","description":"GPT 6 Astra in Pro reasoning mode. Uses additional model work for difficult tasks, with higher latency and token usage at the same per-token rates. Reasoning effort remains independently configurable.","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5}},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT 5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-6-sol-pro":{"id":"openai/gpt-6-sol-pro","name":"GPT 6 Sol Pro","description":"GPT-6 Sol Pro uses the same underlying model as GPT-6 Sol with Pro reasoning mode enabled for higher-quality responses on complex tasks.","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5}},"openai/gpt-oss-safeguard-20b":{"id":"openai/gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.075,"output":0.3},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT 5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":16384},"cost":{"input":0.02,"output":0.1},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT 5.2 Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/o3":{"id":"openai/o3","name":"OpenAI o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":1},"canonical_model_id":"openai/o3"},"openai/o1":{"id":"openai/o1","name":"OpenAI o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT 5.1 Codex Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/gpt-5.6-terra-pro":{"id":"openai/gpt-5.6-terra-pro","name":"GPT 5.6 Terra Pro","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT 5.1 Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT 6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/o4-mini-high":{"id":"openai/o4-mini-high","name":"OpenAI o4-mini high","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2025-12-04","last_updated":"2025-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55}},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT 5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.6-sol-pro":{"id":"openai/gpt-5.6-sol-pro","name":"GPT 5.6 Sol Pro","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT 6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT 5.4 Nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-terra-latest":{"id":"openai/gpt-terra-latest","name":"GPT Terra Latest","description":"Compatibility alias that routes to GPT 5.6 Terra, the latest supported GPT Terra model.","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5}},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.03,"output":0.15,"cache_read":0.03},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT 5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT 6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1-codex-max":{"id":"openai/gpt-5.1-codex-max","name":"GPT 5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":2.5,"output":20,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.1-codex-max"},"openai/o1-pro":{"id":"openai/o1-pro","name":"OpenAI o1 Pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2023-09","release_date":"2025-03-19","last_updated":"2025-03-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":150,"output":600,"cache_read":75},"canonical_model_id":"openai/o1-pro"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT 5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT 5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-6.1-sol-pro":{"id":"openai/gpt-6.1-sol-pro","name":"GPT 6.1 Sol Pro","description":"GPT-6.1 Sol Pro uses the same underlying model as GPT-6.1 Sol with Pro reasoning mode enabled for higher-quality responses on complex tasks.","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5}},"openai/gpt-4o-2024-11-20":{"id":"openai/gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT 5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"aisingapore/Gemma-SEA-LION-v4-27B-IT":{"id":"aisingapore/Gemma-SEA-LION-v4-27B-IT","name":"Gemma SEA-LION v4 27B","description":"Gemma 3 27B tuned by AI Singapore for Southeast Asian languages and instruction following","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":128000},"cost":{"input":0.24,"output":0.48},"canonical_model_id":"aisingapore/gemma-sea-lion-v4-27b-it"},"nanogpt/coding-router:max":{"id":"nanogpt/coding-router:max","name":"Coding Router Max","description":"Automatic model router for matching prompts to suitable backends and budgets","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5}},"nanogpt/coding-router:low":{"id":"nanogpt/coding-router:low","name":"Coding Router Low","description":"Automatic model router for matching prompts to suitable backends and budgets","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":0.14,"output":0.28,"cache_read":0.028}},"nanogpt/coding-router":{"id":"nanogpt/coding-router","name":"Coding Router","description":"Automatic model router for matching prompts to suitable backends and budgets","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":1.1,"output":2.2,"cache_read":0.11}},"nanogpt/coding-router:medium":{"id":"nanogpt/coding-router:medium","name":"Coding Router Medium","description":"Automatic model router for matching prompts to suitable backends and budgets","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":0.14,"output":0.28,"cache_read":0.028}},"nanogpt/coding-router:high":{"id":"nanogpt/coding-router:high","name":"Coding Router High","description":"Automatic model router for matching prompts to suitable backends and budgets","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":1.1,"output":2.2,"cache_read":0.11}},"meganova-ai/manta-flash-1.0":{"id":"meganova-ai/manta-flash-1.0","name":"Manta Flash 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-09-20","last_updated":"2025-12-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"input":16384,"output":16384},"cost":{"input":0.02,"output":0.16,"cache_read":0.01}},"meganova-ai/manta-mini-1.0":{"id":"meganova-ai/manta-mini-1.0","name":"Manta Mini 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-09-20","last_updated":"2025-12-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"input":8192,"output":8192},"cost":{"input":0.02,"output":0.16,"cache_read":0.01}},"meganova-ai/manta-pro-1.0":{"id":"meganova-ai/manta-pro-1.0","name":"Manta Pro 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-09-20","last_updated":"2025-12-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"input":65536,"output":32768},"cost":{"input":0.06,"output":0.5,"cache_read":0.03}},"shisa-ai/shisa-v2.1-llama3.3-70b":{"id":"shisa-ai/shisa-v2.1-llama3.3-70b","name":"Shisa V2.1 Llama 3.3 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":4096},"cost":{"input":0.5,"output":0.5,"cache_read":0.25}},"shisa-ai/shisa-v2-llama3.3-70b":{"id":"shisa-ai/shisa-v2-llama3.3-70b","name":"Shisa V2 Llama 3.3 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.5,"output":0.5,"cache_read":0.25}},"Envoid/Llama-3.05-NT-Storybreaker-Ministral-70B":{"id":"Envoid/Llama-3.05-NT-Storybreaker-Ministral-70B","name":"Llama 3.05 Storybreaker Ministral 70b","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"Envoid/Llama-3.05-Nemotron-Tenyxchat-Storybreaker-70B":{"id":"Envoid/Llama-3.05-Nemotron-Tenyxchat-Storybreaker-70B","name":"Nemotron Tenyxchat Storybreaker 70b","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"arcee-ai/trinity-large-thinking":{"id":"arcee-ai/trinity-large-thinking","name":"Trinity Large Thinking","description":"Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use","family":"trinity","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":80000},"cost":{"input":0.25,"output":0.9,"cache_read":0.125},"canonical_model_id":"arcee-ai/trinity-large-thinking"},"VongolaChouko/Starcannon-Unleashed-12B-v1.0":{"id":"VongolaChouko/Starcannon-Unleashed-12B-v1.0","name":"Mistral Nemo Starcannon 12b v1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}}}},"novita-ai":{"id":"novita-ai","env":["NOVITA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.novita.ai/openai","name":"Novita AI","doc":"https://novita.ai/docs/guides/introduction","models":{"xiaomimimo/mimo-v2.5-pro":{"id":"xiaomimimo/mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.522,"output":1.044,"cache_read":0.0043},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomimimo/mimo-v2.5":{"id":"xiaomimimo/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.168,"output":0.336,"cache_read":0.0034},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomimimo/mimo-v2.6-flash":{"id":"xiaomimimo/mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"xiaomimimo/mimo-v2.6-pro":{"id":"xiaomimimo/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"baichuan/baichuan-m2-32b":{"id":"baichuan/baichuan-m2-32b","name":"Baichuan M2 32B","description":"Open medical model trained for clinical reasoning and patient interaction with on, off, and auto thinking modes","family":"baichuan","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"temperature":true,"release_date":"2025-08-13","last_updated":"2025-08-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.07,"output":0.07},"canonical_model_id":"baichuan/baichuan-m2-32b"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":1.6,"output":3.2,"cache_read":0.135},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-r1-0528-qwen3-8b":{"id":"deepseek/deepseek-r1-0528-qwen3-8b","name":"DeepSeek R1 0528 Qwen3 8B","description":"DeepSeek R1 0528 distilled into Qwen3 8B for mathematics, coding, and reasoning","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"release_date":"2025-05-29","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.06,"output":0.09},"canonical_model_id":"deepseek/deepseek-r1-0528-qwen3-8b"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.44,"output":1.32,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek_v3":{"id":"deepseek/deepseek_v3","name":"DeepSeek-V3","description":"Open DeepSeek MoE chat model for coding, math, and general reasoning","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":16000},"cost":{"input":0.89,"output":0.89},"canonical_model_id":"deepseek/deepseek-v3"},"deepseek/deepseek-r1-0528":{"id":"deepseek/deepseek-r1-0528","name":"DeepSeek R1 0528","description":"May 2025 DeepSeek R1 update with improved mathematics, coding, and reasoning","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.7,"output":2.5,"cache_read":0.35},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-r1-turbo":{"id":"deepseek/deepseek-r1-turbo","name":"DeepSeek R1 (Turbo)","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":16000},"cost":{"input":0.7,"output":2.5},"canonical_model_id":"deepseek/deepseek-r1"},"deepseek/deepseek-ocr-2":{"id":"deepseek/deepseek-ocr-2","name":"DeepSeek OCR 2","description":"High-accuracy OCR model for extracting text from documents, screenshots, receipts, and natural scenes","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.03,"output":0.03},"canonical_model_id":"deepseek/deepseek-ocr-2"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"microsoft/wizardlm-2-8x22b":{"id":"microsoft/wizardlm-2-8x22b","name":"WizardLM 2 8x22B","description":"Microsoft WizardLM instruction-tuned Mixtral MoE for multilingual dialogue and complex tasks","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2024-04-15","last_updated":"2024-04-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65535,"output":8000},"cost":{"input":0.62,"output":0.62},"canonical_model_id":"microsoft/wizardlm-2-8x22b"},"gryphe/mythomax-l2-13b":{"id":"gryphe/mythomax-l2-13b","name":"MythoMax L2 13B","description":"Llama 2 merge combining MythoLogic and Huginn for creative writing and roleplay","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2023-08-10","last_updated":"2023-08-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":3200},"cost":{"input":0.09,"output":0.09},"canonical_model_id":"gryphe/mythomax-l2-13b"},"nvidia/nemotron-3-nano-30b-a3b":{"id":"nvidia/nemotron-3-nano-30b-a3b","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.05,"output":0.2},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"inclusionai/ling-3.0-flash-sante":{"id":"inclusionai/ling-3.0-flash-sante","name":"Ling 3.0 Flash Sante","description":"Health and medicine specialization of Ling 3.0 Flash for medical reasoning and evidence-based retrieval","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash-sante"},"inclusionai/ling-3.0-flash-vl":{"id":"inclusionai/ling-3.0-flash-vl","name":"Ling 3.0 Flash VL","description":"Native multimodal hybrid MoE model for image and video understanding, reasoning, and visual agents","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash-vl"},"inclusionai/ling-3.0-flash":{"id":"inclusionai/ling-3.0-flash","name":"Ling 3.0 Flash","description":"Hybrid-linear MoE reasoning model with 124B total and 5.1B active parameters for efficient agents","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012},"canonical_model_id":"inclusionai/ling-3.0-flash"},"inclusionai/ling-3.0-flash-fin":{"id":"inclusionai/ling-3.0-flash-fin","name":"Ling 3.0 Flash Fin","description":"Finance-enhanced model for financial research, multi-step investment workflows, and long-horizon planning and execution","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-27","last_updated":"2026-09-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash-fin"},"inclusionai/ling-3.1-flash":{"id":"inclusionai/ling-3.1-flash","name":"Ling 3.1 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":0.9,"cache_read":0.06},"canonical_model_id":"inclusionai/ling-3.1-flash"},"meta-llama/llama-3.1-8b-instruct":{"id":"meta-llama/llama-3.1-8b-instruct","name":"Llama-3.1-8B-Instruct","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":0.02,"output":0.05},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta-llama/llama-3.3-70b-instruct":{"id":"meta-llama/llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":12288,"output":12288},"cost":{"input":0.135,"output":0.4},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/llama-4-maverick-17b-128e-instruct-fp8":{"id":"meta-llama/llama-4-maverick-17b-128e-instruct-fp8","name":"Llama 4 Maverick 17B Instruct","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":8192},"cost":{"input":0.27,"output":0.85},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"meta-llama/llama-3.2-1b-instruct":{"id":"meta-llama/llama-3.2-1b-instruct","name":"Llama 3.2 1B Instruct","description":"Compact instruction-tuned Llama model for multilingual dialogue and on-device assistants","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":32000},"cost":{"input":0.02,"output":0.02},"canonical_model_id":"meta/llama-3.2-1b-instruct"},"meta-llama/llama-4-scout-17b-16e-instruct":{"id":"meta-llama/llama-4-scout-17b-16e-instruct","name":"Llama 4 Scout 17B Instruct","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.18,"output":0.59},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"qwen/qwen3-vl-30b-a3b-instruct":{"id":"qwen/qwen3-vl-30b-a3b-instruct","name":"Qwen3 VL 30B A3B Instruct","description":"Open vision-language MoE model for image and video understanding, documents, and visual agents","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.2,"output":0.7},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":3.2},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3.6-35b-a3b":{"id":"qwen/qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.248,"output":1.485},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3-omni-30b-a3b-thinking":{"id":"qwen/qwen3-omni-30b-a3b-thinking","name":"Qwen3 Omni 30B A3B Thinking","description":"Multimodal reasoning model accepting text, images, video, and audio with text output","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-22","last_updated":"2025-09-22","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":0.25,"output":0.97,"input_audio":2.2},"canonical_model_id":"alibaba/qwen3-omni-30b-a3b-thinking"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3.8-2.4t-a95b":{"id":"qwen/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":2.4},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3-embedding-8b":{"id":"qwen/qwen3-embedding-8b","type":"embedding","name":"Qwen3 Embedding 8B","description":"Multilingual text embedding model for retrieval, code search, classification, and clustering","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":0},"cost":{"input":0.07,"output":0},"canonical_model_id":"alibaba/qwen3-embedding-8b"},"qwen/qwen3-reranker-8b":{"id":"qwen/qwen3-reranker-8b","type":"reranking","name":"Qwen3 Reranker 8B","description":"Multilingual query-document reranking model supporting more than 100 languages","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":0},"cost":{"input":0.05,"output":0},"canonical_model_id":"alibaba/qwen3-reranker-8b"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.42,"output":3,"cache_read":0.085},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.6-27b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.25,"output":3.75,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen-mt-plus":{"id":"qwen/qwen-mt-plus","name":"Qwen MT Plus","description":"Flagship Qwen translation model with terminology control and support for 92 languages","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-07-24","last_updated":"2025-07-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":8192},"cost":{"input":0.25,"output":0.75},"canonical_model_id":"alibaba/qwen-mt-plus"},"qwen/qwen3.8-max":{"id":"qwen/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max"},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 Instruct 0905","description":"September 2025 Kimi K2 instruction update with a 256K context and improved agentic coding","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.6,"output":2.5},"canonical_model_id":"moonshotai/kimi-k2-0905"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.8,"output":3.4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2-instruct":{"id":"moonshotai/kimi-k2-instruct","name":"Kimi K2 Instruct","description":"Non-thinking Kimi K2 MoE instruction model for coding and tool-using agents","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.57,"output":2.3},"canonical_model_id":"moonshotai/kimi-k2-instruct"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.6,"output":2.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"mindai/macaron-v1-tall":{"id":"mindai/macaron-v1-tall","name":"Macaron V1 Tall","description":"Qwen3.6-based Mixture-of-LoRA model with specialists for chat, personal agents, coding, and generative UI","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2475,"output":1.43,"cache_read":0.044},"canonical_model_id":"mindlab/macaron-v1-tall"},"mindai/macaron-v1-venti":{"id":"mindai/macaron-v1-venti","name":"Macaron V1 Venti","description":"GLM-5.2-based Mixture-of-LoRA model with specialists for chat, personal agents, coding, and generative UI","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.825,"output":2.475,"cache_read":0.165},"canonical_model_id":"mindlab/macaron-v1-venti"},"paddlepaddle/paddleocr-vl":{"id":"paddlepaddle/paddleocr-vl","name":"PaddleOCR VL","description":"Compact vision-language model for multilingual document parsing, OCR, tables, formulas, and charts","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-10-16","last_updated":"2025-10-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":0.02,"output":0.02},"canonical_model_id":"paddlepaddle/paddleocr-vl"},"baidu/cobuddy":{"id":"baidu/cobuddy","name":"CoBuddy","description":"Baidu Qianfan model optimized for code generation and tool-using coding agents","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-05-06","last_updated":"2026-05-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":0.28,"output":1.13,"cache_read":0.07},"canonical_model_id":"baidu/cobuddy"},"baidu/ernie-4.5-21B-a3b":{"id":"baidu/ernie-4.5-21B-a3b","name":"ERNIE 4.5 21B A3B","description":"Open ERNIE text MoE with 21B total and 3B active parameters","family":"ernie","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-06-30","last_updated":"2025-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":120000,"output":8000},"cost":{"input":0.07,"output":0.28},"canonical_model_id":"baidu/ernie-4.5-21b-a3b"},"minimaxai/minimax-m1-80k":{"id":"minimaxai/minimax-m1-80k","name":"MiniMax M1 80K","description":"Open hybrid-attention MoE reasoning model with a million-token context and an 80K thinking budget","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":40000},"cost":{"input":0.55,"output":2.2},"canonical_model_id":"minimax/minimax-m1-80k"},"stepfun/step-3.7-flash":{"id":"stepfun/step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":256000,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"stepfun/step-5-preview":{"id":"stepfun/step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":64000},"cost":{"input":1,"output":2.7,"cache_read":0.1},"canonical_model_id":"stepfun/step-5-preview"},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"tiers":[{"input":0.6,"output":2.4,"cache_read":0.12,"tier":{"type":"context","size":524288}}],"context_over_200k":{"input":0.6,"output":2.4,"cache_read":0.12}},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"nousresearch/hermes-2-pro-llama-3-8b":{"id":"nousresearch/hermes-2-pro-llama-3-8b","name":"Hermes 2 Pro Llama 3 8B","description":"Llama 3 fine-tune optimized for function calling, structured JSON responses, and dialogue","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-04-30","last_updated":"2024-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.14,"output":0.14},"canonical_model_id":"nousresearch/hermes-2-pro-llama-3-8b"},"zai-org/glm-4.6v":{"id":"zai-org/glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.055},"canonical_model_id":"zhipuai/glm-4.6v"},"zai-org/glm-5.3-flash":{"id":"zai-org/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai-org/glm-4.5v":{"id":"zai-org/glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":0.6,"output":1.8,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5v"},"zai-org/glm-4.5-air":{"id":"zai-org/glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.13,"output":0.85,"cache_read":0.025},"canonical_model_id":"zhipuai/glm-4.5-air"},"zai-org/glm-5.3":{"id":"zai-org/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/glm-4.7":{"id":"zai-org/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"zai-org/glm-5.1":{"id":"zai-org/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1.38,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/glm-4.7-flash":{"id":"zai-org/glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.07,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flash"},"zai-org/glm-5":{"id":"zai-org/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"zai-org/glm-4.6":{"id":"zai-org/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.55,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"zai-org/glm-5.2":{"id":"zai-org/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/autoglm-phone-9b-multilingual":{"id":"zai-org/autoglm-phone-9b-multilingual","name":"AutoGLM Phone 9B Multilingual","description":"Multilingual visual model for operating smartphone interfaces and mobile agents","family":"glm","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-10","last_updated":"2025-12-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.035,"output":0.138},"canonical_model_id":"zhipuai/autoglm-phone-9b-multilingual"},"apodex/apodex-1.1-mini":{"id":"apodex/apodex-1.1-mini","name":"Apodex 1.1 Mini","description":"Open-weight 35B MoE model for long-horizon research, forecasting, and tool-assisted work","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-24","last_updated":"2026-08-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"apodex/apodex-1.1-mini"},"baai/bge-m3":{"id":"baai/bge-m3","type":"embedding","name":"BGE M3","description":"Multilingual embedding model combining dense, sparse, and multi-vector retrieval","family":"bge","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-01-30","last_updated":"2024-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":0},"cost":{"input":0.01,"output":0},"canonical_model_id":"baai/bge-m3"},"baai/bge-reranker-v2-m3":{"id":"baai/bge-reranker-v2-m3","type":"reranking","name":"BGE Reranker V2 M3","description":"Lightweight multilingual cross-encoder that scores query-document relevance","family":"bge","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-03-15","last_updated":"2024-03-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"output":0},"cost":{"input":0.01,"output":0},"canonical_model_id":"baai/bge-reranker-v2-m3"},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.05,"output":0.1},"canonical_model_id":"google/gemma-3-12b-it"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":98304,"output":16384},"cost":{"input":0.119,"output":0.2},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.13,"output":0.4},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":262144},"cost":{"input":0.14,"output":0.58,"cache_read":0.035},"canonical_model_id":"tencent/hy3"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.05,"output":0.25},"canonical_model_id":"openai/gpt-oss-120b"}}},"togetherai":{"id":"togetherai","env":["TOGETHER_API_KEY"],"npm":"@ai-sdk/togetherai","name":"Together AI","doc":"https://docs.together.ai/docs/serverless-models","models":{"nvidia/NVIDIA-Nemotron-3.5-Super-VL":{"id":"nvidia/NVIDIA-Nemotron-3.5-Super-VL","name":"Nemotron 3.5 Super VL 120B A12B","description":"NVIDIA vision-language reasoning model for text, image, and video understanding and agentic tool use","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.15,"output":0.5,"cache_read":0.02},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/nemotron-3-ultra-550b-a55b":{"id":"nvidia/nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512300,"output":512300},"cost":{"input":0.6,"output":3.6,"cache_read":0.2},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"LiquidAI/LFM2-24B-A2B":{"id":"LiquidAI/LFM2-24B-A2B","name":"LFM2-24B-A2B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"liquid","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-02-25","last_updated":"2026-02-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.03,"output":0.12}},"deepseek-ai/DeepSeek-V3":{"id":"deepseek-ai/DeepSeek-V3","name":"DeepSeek-V3","description":"Legacy model retained for compatibility with older integrations","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-12-26","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":1.25,"output":1.25}},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-V4.1-Flash":{"id":"deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-ai/DeepSeek-V3-1":{"id":"deepseek-ai/DeepSeek-V3-1","name":"DeepSeek V3.1","description":"Legacy model retained for compatibility with older integrations","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-08","release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":0.6,"output":1.7}},"deepseek-ai/DeepSeek-R1":{"id":"deepseek-ai/DeepSeek-R1","name":"DeepSeek-R1","description":"Legacy model retained for compatibility with older integrations","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163839,"output":163839},"status":"deprecated","cost":{"input":3,"output":7}},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"meta-llama/Llama-3.3-70B-Instruct-Turbo":{"id":"meta-llama/Llama-3.3-70B-Instruct-Turbo","name":"Llama 3.3 70B","description":"Compact Llama instruction model for fast chat and local deployment","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":1.04,"output":1.04}},"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":250000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06}},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax-M2.5","description":"Legacy model retained for compatibility with older integrations","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.06}},"MiniMaxAI/MiniMax-M2.7":{"id":"MiniMaxAI/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"thinkingmachines/Inkling":{"id":"thinkingmachines/Inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["max","xhigh","high","medium","low","none"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":131072},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"reflection/Beam-1.0-501B-A23B":{"id":"reflection/Beam-1.0-501B-A23B","name":"Beam 1.0 501B A23B","description":"Sparse mixture-of-experts reasoning model with 23B active parameters for coding and agentic workflows","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.4,"output":1.2,"cache_read":0.04},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"meta-models/Muse-Spark-Open-816B-A42B":{"id":"meta-models/Muse-Spark-Open-816B-A42B","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"Qwen/Qwen2.5-7B-Instruct-Turbo":{"id":"Qwen/Qwen2.5-7B-Instruct-Turbo","name":"Qwen 2.5 7B Instruct Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-09-19","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":0.3}},"Qwen/Qwen3-Coder-Next-FP8":{"id":"Qwen/Qwen3-Coder-Next-FP8","name":"Qwen3 Coder Next FP8","description":"Legacy model retained for compatibility with older integrations","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2026-02-03","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.5,"output":1.2}},"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8":{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8","name":"Qwen3 Coder 480B A35B Instruct","description":"Legacy model retained for compatibility with older integrations","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":2,"output":2}},"Qwen/Qwen3.7-Max":{"id":"Qwen/Qwen3.7-Max","name":"Qwen3.7 Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.25,"output":3.75,"cache_read":0.125}},"Qwen/Qwen3.5-9B":{"id":"Qwen/Qwen3.5-9B","name":"Qwen3.5 9B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.17,"output":0.25}},"Qwen/Qwen3-235B-A22B-Instruct-2507-tput":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507-tput","name":"Qwen3 235B A22B Instruct 2507 FP8","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.2,"output":0.6}},"Qwen/Qwen3.5-397B-A17B":{"id":"Qwen/Qwen3.5-397B-A17B","name":"Qwen3.5 397B A17B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-16","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":130000},"status":"deprecated","cost":{"input":0.6,"output":3.6,"cache_read":0.35}},"Qwen/Qwen3.6-Plus":{"id":"Qwen/Qwen3.6-Plus","name":"Qwen3.6 Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3}},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-11","release_date":"2026-04-07","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":0.26}},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5":{"id":"zai-org/GLM-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"status":"deprecated","cost":{"input":1,"output":3.2},"canonical_model_id":"zhipuai/glm-5"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-16","last_updated":"2026-06-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048575,"output":164000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048575,"output":400000},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"essentialai/Rnj-1-Instruct":{"id":"essentialai/Rnj-1-Instruct","name":"Rnj-1 Instruct","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"rnj","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-12-05","last_updated":"2025-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.15,"output":0.15}},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.15,"output":0.6}}}},"llama":{"id":"llama","env":["LLAMA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.llama.com/compat/v1/","name":"Llama","doc":"https://llama.developer.meta.com/docs/models","models":{"cerebras-llama-4-scout-17b-16e-instruct":{"id":"cerebras-llama-4-scout-17b-16e-instruct","name":"Cerebras-Llama-4-Scout-17B-16E-Instruct","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0}},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0}},"llama-4-maverick-17b-128e-instruct-fp8":{"id":"llama-4-maverick-17b-128e-instruct-fp8","name":"Llama-4-Maverick-17B-128E-Instruct-FP8","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0}},"llama-4-scout-17b-16e-instruct-fp8":{"id":"llama-4-scout-17b-16e-instruct-fp8","name":"Llama-4-Scout-17B-16E-Instruct-FP8","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0}},"groq-llama-4-maverick-17b-128e-instruct":{"id":"groq-llama-4-maverick-17b-128e-instruct","name":"Groq-Llama-4-Maverick-17B-128E-Instruct","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0}},"llama-3.3-8b-instruct":{"id":"llama-3.3-8b-instruct","name":"Llama-3.3-8B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0}},"cerebras-llama-4-maverick-17b-128e-instruct":{"id":"cerebras-llama-4-maverick-17b-128e-instruct","name":"Cerebras-Llama-4-Maverick-17B-128E-Instruct","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0}}}},"lucidquery":{"id":"lucidquery","env":["LUCIDQUERY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.lucidquery.com/v1","name":"LucidQuery","doc":"https://lucidquery.com/docs","models":{"lucidquery-nexus-coder":{"id":"lucidquery-nexus-coder","name":"LucidQuery Nexus Coder","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"lucid","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2025-08-01","release_date":"2025-09-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":250000,"output":60000},"cost":{"input":2,"output":5}},"lucidquery-agi-01-swift":{"id":"lucidquery-agi-01-swift","name":"AGI-01 Swift","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"agi","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2026-06-05","release_date":"2026-06-16","last_updated":"2026-06-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":120000},"cost":{"input":2.5,"output":15}},"lucidnova-rf1-100b":{"id":"lucidnova-rf1-100b","name":"LucidNova RF1 100B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2025-09-16","release_date":"2024-12-28","last_updated":"2025-09-10","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":120000,"output":8000},"cost":{"input":2,"output":5}},"lucidquery-agi-01-frontier":{"id":"lucidquery-agi-01-frontier","name":"AGI-01 Frontier","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"agi","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2026-06-05","release_date":"2026-06-16","last_updated":"2026-06-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":120000},"cost":{"input":4.5,"output":22}}}},"databricks":{"id":"databricks","env":["DATABRICKS_HOST","DATABRICKS_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"https://${DATABRICKS_HOST}/ai-gateway/mlflow/v1","name":"Databricks","doc":"https://docs.databricks.com/aws/en/machine-learning/foundation-models/","models":{"databricks-gemini-3-1-flash-lite":{"id":"databricks-gemini-3-1-flash-lite","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"databricks-gemini-2-5-flash":{"id":"databricks-gemini-2-5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"databricks-claude-opus-4-7":{"id":"databricks-claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"databricks-gemini-3-1-pro":{"id":"databricks-gemini-3-1-pro","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"databricks-gemini-3-5-flash":{"id":"databricks-gemini-3-5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"databricks-gpt-5-6-luna":{"id":"databricks-gpt-5-6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1,"output":6,"cache_read":0.1,"tiers":[{"input":2,"output":9,"cache_read":0.2,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":2,"output":9,"cache_read":0.2}},"canonical_model_id":"openai/gpt-5.6-luna"},"databricks-claude-sonnet-4-6":{"id":"databricks-claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"databricks-gpt-5-6-terra":{"id":"databricks-gpt-5-6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.6-terra"},"databricks-kimi-k2-7-code":{"id":"databricks-kimi-k2-7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"databricks-gpt-5-4-mini":{"id":"databricks-gpt-5-4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":1.5,"output":9,"cache_read":0.15},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"databricks-gpt-5-6-sol":{"id":"databricks-gpt-5-6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.6-sol"},"databricks-gemini-3-pro":{"id":"databricks-gemini-3-pro","name":"Gemini 3 Pro Preview","description":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3-pro-preview"},"databricks-claude-opus-4-5":{"id":"databricks-claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"databricks-claude-opus-4-1":{"id":"databricks-claude-opus-4-1","name":"Claude Opus 4.1 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"databricks-gpt-oss-120b":{"id":"databricks-gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.072,"output":0.28}},"databricks-gpt-oss-20b":{"id":"databricks-gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.05,"output":0.2}},"databricks-gemini-2-5-pro":{"id":"databricks-gemini-2-5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"databricks-gpt-5-2":{"id":"databricks-gpt-5-2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"databricks-gpt-5-4":{"id":"databricks-gpt-5-4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"databricks-gemini-3-flash":{"id":"databricks-gemini-3-flash","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"databricks-claude-opus-4-6":{"id":"databricks-claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"databricks-gpt-5-5":{"id":"databricks-gpt-5-5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"databricks-gpt-5-4-nano":{"id":"databricks-gpt-5-4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"databricks-claude-sonnet-4":{"id":"databricks-claude-sonnet-4","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"databricks-gpt-5-1":{"id":"databricks-gpt-5-1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"databricks-claude-haiku-4-5":{"id":"databricks-claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"databricks-gpt-5":{"id":"databricks-gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"databricks-glm-5-2":{"id":"databricks-glm-5-2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"databricks-gpt-5-mini":{"id":"databricks-gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"databricks-gpt-5-nano":{"id":"databricks-gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"databricks-claude-sonnet-4-5":{"id":"databricks-claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"}}},"cloudflare-workers-ai":{"id":"cloudflare-workers-ai","env":["CLOUDFLARE_ACCOUNT_ID","CLOUDFLARE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.cloudflare.com/client/v4/accounts/${CLOUDFLARE_ACCOUNT_ID}/ai/v1","name":"Cloudflare Workers AI","doc":"https://developers.cloudflare.com/workers-ai/models/","models":{"@cf/nvidia/nemotron-3-120b-a12b":{"id":"@cf/nvidia/nemotron-3-120b-a12b","name":"Nemotron 3 Super 120B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"@cf/meta/llama-3.2-11b-vision-instruct":{"id":"@cf/meta/llama-3.2-11b-vision-instruct","name":"Llama 3.2 11B Vision Instruct","description":"Open Llama multimodal model for image understanding and text reasoning","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.0485,"output":0.676},"canonical_model_id":"meta/llama-3.2-11b-vision-instruct"},"@cf/meta/llama-3.2-3b-instruct":{"id":"@cf/meta/llama-3.2-3b-instruct","name":"Llama 3.2 3B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":80000,"output":80000},"cost":{"input":0.0509,"output":0.335},"canonical_model_id":"meta/llama-3.2-3b"},"@cf/meta/llama-3.3-70b-instruct-fp8-fast":{"id":"@cf/meta/llama-3.3-70b-instruct-fp8-fast","name":"Llama 3.3 70B Instruct fp8 Fast","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":24000,"output":24000},"cost":{"input":0.293,"output":2.253},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"@cf/meta/llama-guard-3-8b":{"id":"@cf/meta/llama-guard-3-8b","name":"Llama Guard 3 8B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.484,"output":0.03},"canonical_model_id":"meta/llama-guard-3-8b"},"@cf/meta/llama-3.1-8b-instruct-fp8":{"id":"@cf/meta/llama-3.1-8b-instruct-fp8","name":"Llama 3.1 8B Instruct fp8","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000},"cost":{"input":0.152,"output":0.287},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"@cf/meta/llama-3.2-1b-instruct":{"id":"@cf/meta/llama-3.2-1b-instruct","name":"Llama 3.2 1B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":60000,"output":60000},"cost":{"input":0.027,"output":0.201},"canonical_model_id":"meta/llama-3.2-1b"},"@cf/meta/llama-4-scout-17b-16e-instruct":{"id":"@cf/meta/llama-4-scout-17b-16e-instruct","name":"Llama 4 Scout 17B 16E Instruct","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":16384},"cost":{"input":0.27,"output":0.85},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"@cf/deepseek-ai/deepseek-r1-distill-qwen-32b":{"id":"@cf/deepseek-ai/deepseek-r1-distill-qwen-32b","name":"Deepseek R1 Distill Qwen 32B","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":80000,"output":80000},"cost":{"input":0.497,"output":4.881},"canonical_model_id":"deepseek/deepseek-r1"},"@cf/deepseek-ai/deepseek-v4-flash-0731":{"id":"@cf/deepseek-ai/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"@cf/deepseek-ai/deepseek-v4-pro-0813":{"id":"@cf/deepseek-ai/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"@cf/deepseek-ai/deepseek-v4.1-flash-1017":{"id":"@cf/deepseek-ai/deepseek-v4.1-flash-1017","name":"Deepseek V4.1 Flash 1017","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.3,"output":1.2,"cache_read":0.006}},"@cf/qwen/qwq-32b":{"id":"@cf/qwen/qwq-32b","name":"Qwq 32B","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-03-05","last_updated":"2025-03-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":24000,"output":24000},"cost":{"input":0.66,"output":1},"canonical_model_id":"alibaba/qwq-32b"},"@cf/qwen/qwen3.8-27b":{"id":"@cf/qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.45,"output":3.2,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.8-27b"},"@cf/qwen/qwen2.5-coder-32b-instruct":{"id":"@cf/qwen/qwen2.5-coder-32b-instruct","name":"Qwen2.5 Coder 32B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.66,"output":1},"canonical_model_id":"alibaba/qwen2.5-coder-32b-instruct"},"@cf/qwen/qwen3-30b-a3b-fp8":{"id":"@cf/qwen/qwen3-30b-a3b-fp8","name":"Qwen3 30B A3b fp8","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.0509,"output":0.335},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"@cf/moonshotai/kimi-k2.6":{"id":"@cf/moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"@cf/moonshotai/kimi-k2.7-code":{"id":"@cf/moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"@cf/mistralai/mistral-small-3.1-24b-instruct":{"id":"@cf/mistralai/mistral-small-3.1-24b-instruct","name":"Mistral Small 3.1 24B Instruct","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.351,"output":0.555},"canonical_model_id":"mistral/mistral-small-3-1-24b-instruct-2503"},"@cf/ibm-granite/granite-4.0-h-micro":{"id":"@cf/ibm-granite/granite-4.0-h-micro","name":"Granite 4.0 H Micro","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"granite","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-10-02","last_updated":"2025-10-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.017,"output":0.112},"canonical_model_id":"ibm/granite-4-h-micro"},"@cf/zai-org/glm-5.3-flash":{"id":"@cf/zai-org/glm-5.3-flash","name":"Glm 5.3 Flash","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"@cf/zai-org/glm-5.3":{"id":"@cf/zai-org/glm-5.3","name":"Glm 5.3","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"@cf/zai-org/glm-4.7-flash":{"id":"@cf/zai-org/glm-4.7-flash","name":"GLM-4.7-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.0605,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"@cf/zai-org/glm-5.2":{"id":"@cf/zai-org/glm-5.2","name":"Glm 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"@cf/google/gemma-4-26b-a4b-it":{"id":"@cf/google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.1,"output":0.3,"cache_read":0.05},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"@cf/openai/gpt-oss-20b":{"id":"@cf/openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.2,"output":0.3},"canonical_model_id":"openai/gpt-oss-20b"},"@cf/openai/gpt-oss-120b":{"id":"@cf/openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.35,"output":0.75},"canonical_model_id":"openai/gpt-oss-120b"},"@cf/aisingapore/gemma-sea-lion-v4-27b-it":{"id":"@cf/aisingapore/gemma-sea-lion-v4-27b-it","name":"Gemma Sea Lion V4 27B It","description":"Gemma 3 27B tuned by AI Singapore for Southeast Asian languages and instruction following","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.351,"output":0.555},"canonical_model_id":"aisingapore/gemma-sea-lion-v4-27b-it"}}},"ofox":{"id":"ofox","env":["OFOX_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.ofox.ai/v1","name":"Ofox","doc":"https://ofox.ai/docs","models":{"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"bailian/qwen-plus":{"id":"bailian/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.12,"output":0.29,"cache_read":0.023},"canonical_model_id":"alibaba/qwen-plus"},"bailian/qwen-flash":{"id":"bailian/qwen-flash","name":"Qwen Flash","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.022,"output":0.22,"cache_read":0.0043,"cache_write":0.027},"canonical_model_id":"alibaba/qwen-flash"},"bailian/qwen3-coder-plus":{"id":"bailian/qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Hosted Qwen coder for software agents, repo edits, and long-context code","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.8,"output":9,"cache_read":0.2,"cache_write":1},"canonical_model_id":"alibaba/qwen3-coder-plus"},"bailian/qwen3.5-122b-a10b":{"id":"bailian/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.29,"output":2.29,"cache_read":0.29},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"bailian/qwen3.5-plus":{"id":"bailian/qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.4,"output":2.4,"cache_read":0.04,"cache_write":0.4},"canonical_model_id":"alibaba/qwen3.5-plus"},"bailian/qwen-max":{"id":"bailian/qwen-max","name":"Qwen Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":0.35,"output":1.38,"cache_read":0.069},"canonical_model_id":"alibaba/qwen-max"},"bailian/qwen3.6-plus":{"id":"bailian/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625},"canonical_model_id":"alibaba/qwen3.6-plus"},"bailian/qwen3.6-flash":{"id":"bailian/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.31},"canonical_model_id":"alibaba/qwen3.6-flash"},"bailian/qwen3.5-flash":{"id":"bailian/qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"alibaba/qwen3.5-flash"},"bailian/qwen3-coder-flash":{"id":"bailian/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":2.5,"cache_read":0.06,"cache_write":0.27},"canonical_model_id":"alibaba/qwen3-coder-flash"},"bailian/qwen3.8-flash":{"id":"bailian/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"bailian/qwen3.7-plus":{"id":"bailian/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen3.7-plus"},"bailian/qwen-vl-max":{"id":"bailian/qwen-vl-max","name":"Qwen-VL Max","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-08","last_updated":"2025-08-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.23,"output":0.58,"cache_read":0.046},"canonical_model_id":"alibaba/qwen-vl-max"},"bailian/qwen-turbo":{"id":"bailian/qwen-turbo","name":"Qwen Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-11-01","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.05,"output":0.09,"cache_read":0.0086},"canonical_model_id":"alibaba/qwen-turbo"},"bailian/qwen3.5-27b":{"id":"bailian/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.29,"output":2.05,"cache_read":0.29},"canonical_model_id":"alibaba/qwen3.5-27b"},"bailian/qwen3.5-35b-a3b":{"id":"bailian/qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.29,"output":1.83,"cache_read":0.29},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"bailian/qwen3.8-27b":{"id":"bailian/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1131072,"output":131072},"cost":{"input":0.45,"output":3.2,"cache_read":0.05,"cache_write":0.5625},"canonical_model_id":"alibaba/qwen3.8-27b"},"bailian/qwen3.5-397b-a17b":{"id":"bailian/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.55,"output":3.5,"cache_read":0.55},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"bailian/qwen3.6-max-preview":{"id":"bailian/qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":2.15,"output":12.86,"cache_read":0.2,"cache_write":1.17},"canonical_model_id":"alibaba/qwen3.6-max-preview"},"bailian/qwen3-coder-next":{"id":"bailian/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.2,"output":1.5},"canonical_model_id":"alibaba/qwen3-coder-next"},"bailian/qwen3.6-27b":{"id":"bailian/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.6-27b"},"bailian/qwen3.7-max":{"id":"bailian/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.5,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.7-max"},"bailian/qwen3.8-max":{"id":"bailian/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"bailian/qwen3-max":{"id":"bailian/qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.36,"output":1.43,"cache_read":0.072},"canonical_model_id":"alibaba/qwen3-max"},"bailian/qwen3.8-max-0902":{"id":"bailian/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"deepseek/deepseek-v4-flash-0423":{"id":"deepseek/deepseek-v4-flash-0423","name":"DeepSeek V4 Flash 0423","description":"Initial DeepSeek V4 Flash snapshot for economical reasoning, coding, and million-token agent workloads","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.19,"output":0.51,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0423"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"beta","cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4-pro-0423":{"id":"deepseek/deepseek-v4-pro-0423","name":"DeepSeek V4 Pro 0423","description":"DeepSeek V4 Pro initial snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.15},"canonical_model_id":"deepseek/deepseek-v4-pro-0423"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.29,"output":0.43,"cache_read":0.06},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.4,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-5v-turbo":{"id":"z-ai/glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"z-ai/glm-5.3-flashx":{"id":"z-ai/glm-5.3-flashx","name":"GLM-5.3-FlashX","description":"High-speed GLM-5.3-Flash serving option for coding and agent workflows","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"z-ai/glm-4.6":{"id":"z-ai/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-4.7-flashx":{"id":"z-ai/glm-4.7-flashx","name":"GLM-4.7-FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.072,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"qwen/qwen-plus":{"id":"qwen/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":0.12,"output":0.29,"cache_read":0.023},"canonical_model_id":"alibaba/qwen-plus"},"qwen/qwen-flash":{"id":"qwen/qwen-flash","name":"Qwen Flash","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":0.022,"output":0.22,"cache_read":0.0043,"cache_write":0.027},"canonical_model_id":"alibaba/qwen-flash"},"qwen/qwen3-coder-plus":{"id":"qwen/qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Hosted Qwen coder for software agents, repo edits, and long-context code","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":1.8,"output":9,"cache_read":0.2,"cache_write":1},"canonical_model_id":"alibaba/qwen3-coder-plus"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.29,"output":2.29,"cache_read":0.29},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3.5-plus":{"id":"qwen/qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":2.4,"cache_read":0.04,"cache_write":0.4},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen/qwen-max":{"id":"qwen/qwen-max","name":"Qwen Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":8000},"cost":{"input":0.35,"output":1.38,"cache_read":0.069},"canonical_model_id":"alibaba/qwen-max"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.31},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3.5-flash":{"id":"qwen/qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"alibaba/qwen3.5-flash"},"qwen/qwen3-coder-flash":{"id":"qwen/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.5,"output":2.5,"cache_read":0.06,"cache_write":0.27},"canonical_model_id":"alibaba/qwen3-coder-flash"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.11,"output":0.39,"cache_read":0.011,"cache_write":0.14},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1064000,"output":64000},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen-vl-max":{"id":"qwen/qwen-vl-max","name":"Qwen-VL Max","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-08","last_updated":"2025-08-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8000},"cost":{"input":0.23,"output":0.58,"cache_read":0.023},"canonical_model_id":"alibaba/qwen-vl-max"},"qwen/qwen-turbo":{"id":"qwen/qwen-turbo","name":"Qwen Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-11-01","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000},"cost":{"input":0.043,"output":0.09,"cache_read":0.0086},"canonical_model_id":"alibaba/qwen-turbo"},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.29,"output":2.05,"cache_read":0.29},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.29,"output":1.83,"cache_read":0.29},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1131072,"output":131072},"cost":{"input":0.5,"output":1.71,"cache_read":0.043,"cache_write":0.63},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.55,"output":3.5,"cache_read":0.55},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3.6-max-preview":{"id":"qwen/qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":2.15,"output":12.86,"cache_read":0.2,"cache_write":1.17},"canonical_model_id":"alibaba/qwen3.6-max-preview"},"qwen/qwen3-coder-next":{"id":"qwen/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.2,"output":1.5},"canonical_model_id":"alibaba/qwen3-coder-next"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.43,"output":2.57},"canonical_model_id":"alibaba/qwen3.6-27b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1064000,"output":64000},"cost":{"input":1.71,"output":5.14,"cache_read":0.17,"cache_write":2.14},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.8-max":{"id":"qwen/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.71,"output":5.14,"cache_read":0.17,"cache_write":2.14},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen/qwen3-max":{"id":"qwen/qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":0.36,"output":1.43,"cache_read":0.072},"canonical_model_id":"alibaba/qwen3-max"},"qwen/qwen3.8-max-0902":{"id":"qwen/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.71,"output":5.14,"cache_read":0.17,"cache_write":2.14},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2.7-code-highspeed":{"id":"moonshotai/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"minimax/minimax-m2.1-lightning":{"id":"minimax/minimax-m2.1-lightning","name":"MiniMax-M2.1 Lightning","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":2.4,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-m2.5-lightning":{"id":"minimax/minimax-m2.5-lightning","name":"MiniMax-M2.5 Lightning","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":2.4,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5-highspeed"},"minimax/m2-her":{"id":"minimax/m2-her","name":"MiniMax-M2 Her","description":"MiniMax M2 variant tuned for conversational and character-driven agent interactions","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-01-23","last_updated":"2026-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":2048},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2-Her"},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax-M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5"},"volcengine/doubao-seed-1-6-flash":{"id":"volcengine/doubao-seed-1-6-flash","name":"Seed 1.6 Flash","description":"Low-latency ByteDance Seed model for high-throughput chat, extraction, and lightweight tool use","family":"seed","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.03,"output":0.22,"cache_read":0.0043},"canonical_model_id":"bytedance-seed/seed-1-6-flash"},"volcengine/doubao-seed-1-8":{"id":"volcengine/doubao-seed-1-8","name":"Seed 1.8","description":"ByteDance Seed model for multimodal reasoning, long-context analysis, and agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-28","last_updated":"2025-12-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":0.12,"output":0.29,"cache_read":0.023},"canonical_model_id":"bytedance-seed/seed-1-8"},"volcengine/doubao-seed-1-6":{"id":"volcengine/doubao-seed-1-6","name":"Seed 1.6","description":"ByteDance Seed model for long-context reasoning, instruction following, and tool-assisted tasks","family":"seed","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":0.12,"output":0.29,"cache_read":0.023},"canonical_model_id":"bytedance-seed/seed-1-6"},"volcengine/doubao-seed-2.1-pro":{"id":"volcengine/doubao-seed-2.1-pro","name":"Seed 2.1 Pro","description":"Flagship ByteDance Seed 2.1 model for complex multimodal reasoning, coding, and agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.7072,"output":3.536,"cache_read":0.1416,"cache_write":0.002},"canonical_model_id":"bytedance-seed/seed-2.1-pro"},"volcengine/doubao-seed-character":{"id":"volcengine/doubao-seed-character","name":"Seed Character","description":"ByteDance Seed model optimized for character-driven dialogue and consistent conversational behavior","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.177,"output":0.884,"cache_read":0.024,"cache_write":0.0025},"canonical_model_id":"bytedance-seed/seed-character"},"volcengine/doubao-seed-1-6-vision":{"id":"volcengine/doubao-seed-1-6-vision","name":"Seed 1.6 Vision","description":"ByteDance Seed multimodal model for image understanding, visual reasoning, and tool-assisted tasks","family":"seed","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-15","last_updated":"2025-08-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.12,"output":1.15,"cache_read":0.023},"canonical_model_id":"bytedance-seed/seed-1-6-vision"},"volcengine/doubao-seed-2.0-mini":{"id":"volcengine/doubao-seed-2.0-mini","name":"Seed 2.0 Mini","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.06,"output":0.56,"cache_read":0.02,"cache_write":0.0024},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"volcengine/doubao-seed-2.0-code":{"id":"volcengine/doubao-seed-2.0-code","name":"Seed 2.0 Code","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.67,"output":3.36,"cache_read":0.14,"cache_write":0.0024},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"volcengine/doubao-seed-2.0-lite":{"id":"volcengine/doubao-seed-2.0-lite","name":"Seed 2.0 Lite","description":"Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.13,"output":0.76,"cache_read":0.03,"cache_write":0.0024},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"volcengine/doubao-seed-2.0-pro":{"id":"volcengine/doubao-seed-2.0-pro","name":"Seed 2.0 Pro","description":"Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.67,"output":3.36,"cache_read":0.14,"cache_write":0.0024},"canonical_model_id":"bytedance-seed/seed-2.0-pro"},"volcengine/doubao-seed-evolving":{"id":"volcengine/doubao-seed-evolving","name":"Seed Evolving","description":"Rolling ByteDance Seed model for rapidly updated reasoning, coding, and agent capabilities","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.884,"output":4.42,"cache_read":0.177,"cache_write":0.0025},"canonical_model_id":"bytedance-seed/seed-evolving"},"volcengine/doubao-seed-2.1-turbo":{"id":"volcengine/doubao-seed-2.1-turbo","name":"Seed 2.1 Turbo","description":"Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.3536,"output":1.7696,"cache_read":0.068,"cache_write":0.0019},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"x-ai/grok-4.1-fast":{"id":"x-ai/grok-4.1-fast","name":"Grok 4.1 Fast","description":"xAI's fast agentic tool-calling model with a 2M context window; non-reasoning variant for low-latency responses","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast"},"x-ai/grok-4.7":{"id":"x-ai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":65536},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"x-ai/grok-4.5":{"id":"x-ai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":65536},"cost":{"input":2,"output":6,"cache_read":0.3},"canonical_model_id":"xai/grok-4.5"},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":65536},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"x-ai/grok-4.20":{"id":"x-ai/grok-4.20","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":128000},"cost":{"input":4,"output":12,"cache_read":0.4},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":1},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":1,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"cache_write":4.5},"canonical_model_id":"google/gemini-2.5-pro"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":1,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://api.ofox.ai/gemini/v1beta"},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.0415,"input_audio":1.5},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://api.ofox.ai/gemini/v1beta"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":4.5},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://api.ofox.ai/gemini/v1beta"},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.0415,"input_audio":0.75},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://api.ofox.ai/gemini/v1beta"},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":0.083,"input_audio":3},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://api.ofox.ai/gemini/v1beta"},"cost":{"input":2,"output":10,"cache_read":0.1,"input_audio":2},"canonical_model_id":"google/gemini-4-argon"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://api.ofox.ai/gemini/v1beta"},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.083},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":1,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://api.ofox.ai/gemini/v1beta"},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.0415,"input_audio":1.5},"canonical_model_id":"google/gemini-3.6-flash"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.18},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32768},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.25,"output":2,"cache_read":0.03},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.18},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.25,"output":2,"cache_read":0.03},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1-codex-max":{"id":"openai/gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5.1-codex-max"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.05,"output":0.4,"cache_read":0.01},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.18},"canonical_model_id":"openai/gpt-5.3-codex"}}},"agentrouter":{"id":"agentrouter","env":["AGENTROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://agentrouter.org/v1","name":"AgentRouter","doc":"https://agentrouter.org/docs/opencode.html","models":{"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"canonical_model_id":"zhipuai/glm-5.3"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://agentrouter.org/v1"},"canonical_model_id":"anthropic/claude-opus-4-8"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"canonical_model_id":"deepseek/deepseek-v4-flash"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://agentrouter.org/v1"},"canonical_model_id":"anthropic/claude-opus-5"}}},"deepseek":{"id":"deepseek","env":["DEEPSEEK_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.deepseek.com","name":"DeepSeek","doc":"https://api-docs.deepseek.com/quick_start/pricing","models":{"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.66,"output":1.98,"reasoning":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek-flash":{"id":"deepseek-flash","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.15,"output":0.6,"reasoning":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-v4-flash-vision-exp":{"id":"deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"status":"deprecated","cost":{"input":0.15,"output":0.6,"reasoning":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"status":"deprecated","cost":{"input":0.15,"output":0.6,"reasoning":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"}}},"alibaba-coding-plan":{"id":"alibaba-coding-plan","env":["ALIBABA_CODING_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://coding-intl.dashscope.aliyuncs.com/v1","name":"Alibaba Coding Plan","doc":"https://www.alibabacloud.com/help/en/model-studio/coding-plan","models":{"qwen3-coder-plus":{"id":"qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3.5-plus":{"id":"qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"input":196601,"output":24576},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.7-plus"},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3-max-2026-01-23":{"id":"qwen3-max-2026-01-23","name":"Qwen3 Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-23","last_updated":"2026-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"glm-5":{"id":"glm-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":16384},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3-coder-next":{"id":"qwen3-coder-next","name":"Qwen3 Coder Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}}}},"modelis":{"id":"modelis","env":["MODELIS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://modelishub.com/v1","name":"Modelis","doc":"https://modelishub.com/pricing","models":{"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.435,"output":0.87},"canonical_model_id":"deepseek/deepseek-v4-pro"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50},"canonical_model_id":"anthropic/claude-fable-5"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10},"canonical_model_id":"google/gemini-2.5-pro"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"toggle"}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-8"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.0983,"output":0.1966},"canonical_model_id":"deepseek/deepseek-v4-flash"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","max"]},{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5},"canonical_model_id":"google/gemini-2.5-flash"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.768,"output":3.072},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":3,"output":9},"canonical_model_id":"alibaba/qwen3.7-max"}}},"moonshotai-cn":{"id":"moonshotai-cn","env":["MOONSHOT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.moonshot.cn/v1","name":"Moonshot AI (China)","doc":"https://platform.moonshot.cn/docs/api/chat","models":{"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.9711,"output":4.0337,"cache_read":0.1643},"canonical_model_id":"moonshotai/kimi-k2.6"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":2.9879,"output":14.9395,"cache_read":0.2988,"cache_write":2.9879},"canonical_model_id":"moonshotai/kimi-k3"},"kimi-k2.7-code-highspeed":{"id":"kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code HighSpeed","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9421,"output":8.0673,"cache_read":0.3884},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.9711,"output":4.0337,"cache_read":0.1942},"canonical_model_id":"moonshotai/kimi-k2.7-code"}}},"ebcloud":{"id":"ebcloud","env":["EBCLOUD_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://maas-api.ebcloud.com/v1","name":"EBCloud","doc":"https://docs.ebtech.com/ai/model-api.html","models":{"Kimi-K2.6":{"id":"Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.9286,"output":3.8571},"canonical_model_id":"moonshotai/kimi-k2.6"},"GLM-5.1":{"id":"GLM-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.8571,"output":3.4286},"canonical_model_id":"zhipuai/glm-5.1"},"DeepSeek-V4-Flash":{"id":"DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.143,"output":0.2857},"canonical_model_id":"deepseek/deepseek-v4-flash"},"DeepSeek-V4-Pro":{"id":"DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.4286,"output":0.8571},"canonical_model_id":"deepseek/deepseek-v4-pro"}}},"nvidia":{"id":"nvidia","env":["NVIDIA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://integrate.api.nvidia.com/v1","name":"Nvidia","doc":"https://docs.api.nvidia.com/nim/","models":{"microsoft/phi-4-multimodal-instruct":{"id":"microsoft/phi-4-multimodal-instruct","name":"Phi 4 Multimodal","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":16384},"status":"deprecated","cost":{"input":0,"output":0}},"microsoft/phi-4-mini-instruct":{"id":"microsoft/phi-4-mini-instruct","name":"Phi-4-Mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"phi","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2024-12-01","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/streampetr":{"id":"nvidia/streampetr","name":"streampetr","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0}},"nvidia/nemotron-mini-4b-instruct":{"id":"nvidia/nemotron-mini-4b-instruct","name":"nemotron-mini-4b-instruct","description":"Compact Nemotron model for efficient reasoning and deployable AI agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-08-21","last_updated":"2024-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-mini-4b-instruct"},"nvidia/usdcode":{"id":"nvidia/usdcode","name":"usdcode","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-01","last_updated":"2026-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0}},"nvidia/nemotron-3.5-content-safety":{"id":"nvidia/nemotron-3.5-content-safety","name":"Nemotron 3.5 Content Safety","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"nvidia/nemotron-3.5-lightning-30b-a3b":{"id":"nvidia/nemotron-3.5-lightning-30b-a3b","name":"Nemotron 3.5 Lightning 30B A3B","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/llama-3_2-nemoretriever-300m-embed-v1":{"id":"nvidia/llama-3_2-nemoretriever-300m-embed-v1","name":"llama-3_2-nemoretriever-300m-embed-v1","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-07-24","last_updated":"2025-07-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":2048},"cost":{"input":0,"output":0}},"nvidia/studiovoice":{"id":"nvidia/studiovoice","name":"studiovoice","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-10-03","last_updated":"2025-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0}},"nvidia/nemotron-voicechat":{"id":"nvidia/nemotron-voicechat","name":"nemotron-voicechat","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-voicechat"},"nvidia/llama-3.1-nemotron-safety-guard-8b-v3":{"id":"nvidia/llama-3.1-nemotron-safety-guard-8b-v3","name":"llama-3.1-nemotron-safety-guard-8b-v3","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-28","last_updated":"2025-10-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/llama-3.1-nemotron-safety-guard-8b-v3"},"nvidia/llama-3.1-nemotron-70b-instruct":{"id":"nvidia/llama-3.1-nemotron-70b-instruct","name":"Llama 3.1 Nemotron 70B Instruct","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/llama-3.1-nemotron-70b-instruct"},"nvidia/cosmos-predict1-5b":{"id":"nvidia/cosmos-predict1-5b","name":"cosmos-predict1-5b","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-03-18","last_updated":"2025-03-18","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":true,"limit":{"context":0,"output":4096},"cost":{"input":0,"output":0}},"nvidia/nvidia-nemotron-nano-9b-v2":{"id":"nvidia/nvidia-nemotron-nano-9b-v2","name":"nvidia-nemotron-nano-9b-v2","description":"Compact Nemotron model for efficient reasoning and deployable AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2025-08-18","last_updated":"2025-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-nano-9b-v2"},"nvidia/llama-3.3-nemotron-super-49b-v1":{"id":"nvidia/llama-3.3-nemotron-super-49b-v1","name":"Llama 3.3 Nemotron Super 49B v1","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-03-18","last_updated":"2025-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/llama-3.3-nemotron-super-49b-v1"},"nvidia/nemotron-content-safety-reasoning-4b":{"id":"nvidia/nemotron-content-safety-reasoning-4b","name":"nemotron-content-safety-reasoning-4b","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":false,"release_date":"2026-01-22","last_updated":"2026-01-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-content-safety-reasoning-4b"},"nvidia/magpie-tts-zeroshot":{"id":"nvidia/magpie-tts-zeroshot","name":"magpie-tts-zeroshot","description":"Speech generation model for controllable voice, narration, and audio delivery","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-05-22","last_updated":"2025-06-12","modalities":{"input":["text","audio"],"output":["audio"]},"open_weights":true,"limit":{"context":0,"output":4096},"cost":{"input":0,"output":0}},"nvidia/nemotron-3-ultra-550b-a55b":{"id":"nvidia/nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":2.5,"cache_read":0.15},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/llama-3.3-nemotron-super-49b-v1.5":{"id":"nvidia/llama-3.3-nemotron-super-49b-v1.5","name":"Llama 3.3 Nemotron Super 49B v1.5","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/llama-3.3-nemotron-super-49b-v1.5"},"nvidia/llama-nemotron-embed-vl-1b-v2":{"id":"nvidia/llama-nemotron-embed-vl-1b-v2","type":"embedding","name":"llama-nemotron-embed-vl-1b-v2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"nemotron","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-10","last_updated":"2026-02-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":2048},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/llama-nemotron-embed-vl-1b-v2"},"nvidia/gliner-pii":{"id":"nvidia/gliner-pii","name":"gliner-pii","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/nemotron-3.5-super-vl-120b-a12b":{"id":"nvidia/nemotron-3.5-super-vl-120b-a12b","name":"Nemotron 3.5 Super VL 120B A12B","description":"NVIDIA vision-language reasoning model for text, image, and video understanding and agentic tool use","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/cosmos-reason2-8b":{"id":"nvidia/cosmos-reason2-8b","name":"Cosmos Reason2 8B","description":"Vision language model for physical-world understanding with structured reasoning on video and images","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0,"output":0}},"nvidia/llama-3.1-nemotron-ultra-253b-v1":{"id":"nvidia/llama-3.1-nemotron-ultra-253b-v1","name":"Llama 3.1 Nemotron Ultra 253B","description":"Flagship Nemotron model for high-throughput reasoning and complex agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-04-07","last_updated":"2025-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/llama-3.1-nemotron-ultra-253b"},"nvidia/cosmos-transfer1-7b":{"id":"nvidia/cosmos-transfer1-7b","name":"cosmos-transfer1-7b","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-13","last_updated":"2025-06-30","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":true,"limit":{"context":0,"output":4096},"cost":{"input":0,"output":0}},"nvidia/usdvalidate":{"id":"nvidia/usdvalidate","name":"usdvalidate","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-07-24","last_updated":"2025-01-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":0,"output":4096},"cost":{"input":0,"output":0}},"nvidia/nv-embedcode-7b-v1":{"id":"nvidia/nv-embedcode-7b-v1","name":"nv-embedcode-7b-v1","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-03-17","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":2048},"cost":{"input":0,"output":0}},"nvidia/riva-translate-4b-instruct-v1.1":{"id":"nvidia/riva-translate-4b-instruct-v1.1","name":"riva-translate-4b-instruct-v1_1","description":"Translation model for multilingual conversion, localization, and cross-language workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-12","last_updated":"2025-12-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0}},"nvidia/llama-3.1-nemotron-nano-vl-8b-v1":{"id":"nvidia/llama-3.1-nemotron-nano-vl-8b-v1","name":"Llama 3.1 Nemotron Nano VL 8B v1","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-04-10","last_updated":"2025-04-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/bevformer":{"id":"nvidia/bevformer","name":"bevformer","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-03-18","last_updated":"2025-07-20","modalities":{"input":["video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0}},"nvidia/nemotron-3-nano-30b-a3b":{"id":"nvidia/nemotron-3-nano-30b-a3b","name":"nemotron-3-nano-30b-a3b","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2024-12","last_updated":"2024-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/nv-embed-v1":{"id":"nvidia/nv-embed-v1","name":"nv-embed-v1","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-06-07","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":2048},"cost":{"input":0,"output":0}},"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning":{"id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning","name":"Nemotron 3 Nano Omni","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":-1,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"nvidia/llama-3.1-nemotron-nano-8b-v1":{"id":"nvidia/llama-3.1-nemotron-nano-8b-v1","name":"Llama 3.1 Nemotron Nano 8B v1","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-03-18","last_updated":"2025-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/nemotron-nano-12b-v2-vl":{"id":"nvidia/nemotron-nano-12b-v2-vl","name":"Nemotron Nano 12B v2 VL","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-10-28","last_updated":"2025-10-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-nano-12b-v2-vl"},"nvidia/active-speaker-detection":{"id":"nvidia/active-speaker-detection","name":"Active Speaker Detection","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["video"],"output":["text"]},"open_weights":true,"limit":{"context":0,"output":4096},"cost":{"input":0,"output":0}},"nvidia/nemotron-3-content-safety":{"id":"nvidia/nemotron-3-content-safety","name":"nemotron-3-content-safety","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-content-safety"},"nvidia/sparsedrive":{"id":"nvidia/sparsedrive","name":"sparsedrive","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-03-18","last_updated":"2025-07-20","modalities":{"input":["video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0}},"nvidia/cosmos-transfer2_5-2b":{"id":"nvidia/cosmos-transfer2_5-2b","name":"cosmos-transfer2.5-2b","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":true,"limit":{"context":0,"output":4096},"cost":{"input":0,"output":0}},"nvidia/synthetic-video-detector":{"id":"nvidia/synthetic-video-detector","name":"synthetic-video-detector","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["video"],"output":["text"]},"open_weights":true,"limit":{"context":0,"output":4096},"cost":{"input":0,"output":0}},"nvidia/llama-nemotron-rerank-vl-1b-v2":{"id":"nvidia/llama-nemotron-rerank-vl-1b-v2","type":"reranking","name":"llama-nemotron-rerank-vl-1b-v2","description":"Reranking model for improving retrieval quality in search and recommendation systems","family":"nemotron","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/llama-nemotron-rerank-vl-1b-v2"},"nvidia/rerank-qa-mistral-4b":{"id":"nvidia/rerank-qa-mistral-4b","name":"rerank-qa-mistral-4b","description":"Reranking model for improving retrieval quality in search and recommendation systems","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-03-17","last_updated":"2025-01-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0}},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.2"},"meta/llama-4-maverick-17b-128e-instruct":{"id":"meta/llama-4-maverick-17b-128e-instruct","name":"Llama 4 Maverick 17b 128e Instruct","description":"Open multimodal Llama model for strong reasoning and fast responses","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-02","release_date":"2025-04-01","last_updated":"2025-04-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"meta/llama-3.1-8b-instruct":{"id":"meta/llama-3.1-8b-instruct","name":"Llama 3.1 8B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"meta/esmfold":{"id":"meta/esmfold","name":"esmfold","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-03-15","last_updated":"2025-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0}},"meta/esm2-650m":{"id":"meta/esm2-650m","name":"esm2-650m","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-08-29","last_updated":"2025-03-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0}},"meta/llama-3.2-90b-vision-instruct":{"id":"meta/llama-3.2-90b-vision-instruct","name":"Llama-3.2-90B-Vision-Instruct","description":"Open Llama multimodal model for image understanding and text reasoning","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0}},"meta/llama-3.2-11b-vision-instruct":{"id":"meta/llama-3.2-11b-vision-instruct","name":"Llama 3.2 11b Vision Instruct","description":"Open Llama multimodal model for image understanding and text reasoning","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-18","last_updated":"2024-09-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0}},"meta/llama-3.2-3b-instruct":{"id":"meta/llama-3.2-3b-instruct","name":"Llama 3.2 3B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2024-09-18","last_updated":"2024-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32000},"status":"deprecated","cost":{"input":0,"output":0}},"meta/llama-3.3-70b-instruct":{"id":"meta/llama-3.3-70b-instruct","name":"Llama 3.3 70b Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-11-26","last_updated":"2024-11-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"meta/muse-glimmer-30b":{"id":"meta/muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"meta/muse-glimmer-30b"},"meta/llama-3.2-1b-instruct":{"id":"meta/llama-3.2-1b-instruct","name":"Llama 3.2 1b Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-18","last_updated":"2024-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"meta/llama-guard-4-12b":{"id":"meta/llama-guard-4-12b","name":"Llama Guard 4 12B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-04-05","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0,"output":0}},"meta/llama-3.1-70b-instruct":{"id":"meta/llama-3.1-70b-instruct","name":"Llama 3.1 70b Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-07-16","last_updated":"2024-07-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"deepseek-ai/deepseek-v4-pro":{"id":"deepseek-ai/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"status":"deprecated","cost":{"input":0.435,"output":0.87,"cache_read":0.003625},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/deepseek-v4-flash-0731":{"id":"deepseek-ai/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/deepseek-v4-pro-0813":{"id":"deepseek-ai/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek-ai/deepseek-v4-flash":{"id":"deepseek-ai/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-ai/deepseek-v4.1-flash":{"id":"deepseek-ai/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0,"output":0}},"qwen/qwen3-next-80b-a3b-instruct":{"id":"qwen/qwen3-next-80b-a3b-instruct","name":"Qwen3-Next-80B-A3B-Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2024-12-01","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":16384},"status":"deprecated","cost":{"input":0,"output":0}},"qwen/qwen-image":{"id":"qwen/qwen-image","name":"Qwen Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}},"qwen/qwen-image-edit":{"id":"qwen/qwen-image-edit","name":"Qwen Image Edit","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-08-19","last_updated":"2025-08-19","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}},"qwen/qwen3-coder-480b-a35b-instruct":{"id":"qwen/qwen3-coder-480b-a35b-instruct","name":"Qwen3 Coder 480B A35B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":66536},"status":"deprecated","cost":{"input":0,"output":0}},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5-397B-A17B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"status":"deprecated","cost":{"input":0,"output":0}},"qwen/qwen2.5-coder-32b-instruct":{"id":"qwen/qwen2.5-coder-32b-instruct","name":"Qwen2.5 Coder 32b Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-11-06","last_updated":"2024-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"sarvamai/sarvam-m":{"id":"sarvamai/sarvam-m","name":"sarvam-m","description":"Efficient Indian-language reasoning model for chat, coding, and multilingual work","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0,"output":0}},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0,"output":0}},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"moonshotai/kimi-k3"},"poolside/laguna-xs-2.1":{"id":"poolside/laguna-xs-2.1","name":"Laguna XS 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-02","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-xs-2.1"},"stepfun-ai/step-3.7-flash":{"id":"stepfun-ai/step-3.7-flash","name":"Step 3.7 Flash","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"status":"deprecated","cost":{"input":0,"output":0}},"stepfun-ai/step-5-preview":{"id":"stepfun-ai/step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":1048576},"cost":{"input":0,"output":0},"canonical_model_id":"stepfun/step-5-preview"},"stepfun-ai/step-3.5-flash":{"id":"stepfun-ai/step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-02-02","last_updated":"2026-02-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"status":"deprecated","cost":{"input":0,"output":0}},"bytedance/seed-oss-36b-instruct":{"id":"bytedance/seed-oss-36b-instruct","name":"ByteDance-Seed/Seed-OSS-36B-Instruct","description":"Tool-capable chat model for instruction following and agentic application workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-04","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"status":"deprecated","cost":{"input":0,"output":0}},"minimaxai/minimax-m2.7":{"id":"minimaxai/minimax-m2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-04-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0}},"minimaxai/minimax-m3":{"id":"minimaxai/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":16384},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"minimax/MiniMax-M3"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":16384},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"thinkingmachines/inkling"},"abacusai/dracarys-llama-3.1-70b-instruct":{"id":"abacusai/dracarys-llama-3.1-70b-instruct","name":"dracarys-llama-3.1-70b-instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-09-11","last_updated":"2025-05-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0,"output":0}},"mistralai/mistral-7b-instruct-v0.3":{"id":"mistralai/mistral-7b-instruct-v0.3","name":"Mistral-7B-Instruct-v0.3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-01","last_updated":"2025-04-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0,"output":0}},"mistralai/ministral-14b-instruct-2512":{"id":"mistralai/ministral-14b-instruct-2512","name":"Ministral 3 14B Instruct 2512","description":"Compact Mistral VLM for chat and instruction-based workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"status":"deprecated","cost":{"input":0,"output":0}},"mistralai/mixtral-8x7b-instruct":{"id":"mistralai/mixtral-8x7b-instruct","name":"Mistral: Mixtral 8x7B Instruct","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2023-12-10","last_updated":"2026-03-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"status":"deprecated","cost":{"input":0,"output":0}},"mistralai/mistral-large-3-675b-instruct-2512":{"id":"mistralai/mistral-large-3-675b-instruct-2512","name":"Mistral Large 3 675B Instruct 2512","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0,"output":0}},"mistralai/mistral-nemotron":{"id":"mistralai/mistral-nemotron","name":"mistral-nemotron","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"nemotron","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-06-11","last_updated":"2025-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/mistral-nemotron"},"mistralai/mistral-medium-3-instruct":{"id":"mistralai/mistral-medium-3-instruct","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0,"output":0}},"mistralai/mixtral-8x22b-instruct":{"id":"mistralai/mixtral-8x22b-instruct","name":"Mistral: Mixtral 8x22B Instruct","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-04-17","last_updated":"2024-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":13108},"status":"deprecated","cost":{"input":0,"output":0}},"mistralai/mistral-small-4-119b-2603":{"id":"mistralai/mistral-small-4-119b-2603","name":"mistral-small-4-119b-2603","description":"Efficient Mistral model for fast chat, extraction, and production assistants","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0,"output":0}},"mistralai/mistral-medium-3.5-128b":{"id":"mistralai/mistral-medium-3.5-128b","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"mistral/mistral-medium-2604"},"mistralai/magistral-small-2506":{"id":"mistralai/magistral-small-2506","name":"Magistral Small 2506","description":"Mistral reasoning model for transparent analysis, math, and complex decisions","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":0,"output":0}},"black-forest-labs/flux.1-dev":{"id":"black-forest-labs/flux.1-dev","name":"FLUX.1-dev","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-08","release_date":"2024-08-01","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":4096,"output":0},"cost":{"input":0,"output":0}},"black-forest-labs/flux_1-kontext-dev":{"id":"black-forest-labs/flux_1-kontext-dev","name":"FLUX.1-Kontext-dev","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-08-12","last_updated":"2025-08-12","modalities":{"input":["text","image"],"output":["image"]},"open_weights":true,"limit":{"context":40960,"output":40960},"cost":{"input":0,"output":0}},"black-forest-labs/flux_1-schnell":{"id":"black-forest-labs/flux_1-schnell","name":"FLUX.1-schnell","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2024-07","release_date":"2024-08-01","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["image"]},"open_weights":true,"limit":{"context":77,"input":77,"output":0},"cost":{"input":0,"output":0}},"black-forest-labs/flux_2-klein-4b":{"id":"black-forest-labs/flux_2-klein-4b","name":"FLUX.2 Klein 4B","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-06","release_date":"2026-01-14","last_updated":"2026-01-31","modalities":{"input":["image","text"],"output":["image"]},"open_weights":true,"limit":{"context":40960,"output":40960},"cost":{"input":0,"output":0}},"upstage/solar-10.7b-instruct":{"id":"upstage/solar-10.7b-instruct","name":"solar-10.7b-instruct","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-06-05","last_updated":"2025-04-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0,"output":0}},"baai/bge-m3":{"id":"baai/bge-m3","name":"BGE M3","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"bge","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-01-30","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":1024},"cost":{"input":0,"output":0}},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0,"output":0}},"google/gemma-3n-e4b-it":{"id":"google/gemma-3n-e4b-it","name":"Gemma 3n E4b It","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-06-03","last_updated":"2025-06-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"google/gemma-3n-e2b-it":{"id":"google/gemma-3n-e2b-it","name":"Gemma 3n E2b It","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-06-12","last_updated":"2025-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"google/google-paligemma":{"id":"google/google-paligemma","name":"paligemma","description":"Gemini multimodal model for text, image, audio, video, and document tasks","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-05-14","last_updated":"2024-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0}},"google/gemma-2-2b-it":{"id":"google/gemma-2-2b-it","name":"Gemma 2 2b It","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-07-16","last_updated":"2024-07-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"google/gemma-3-4b-it":{"id":"google/gemma-3-4b-it","name":"Gemma 3 4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0,"output":0}},"google/diffusiongemma-26b-a4b-it":{"id":"google/diffusiongemma-26b-a4b-it","name":"DiffusionGemma 26B A4B IT","description":"Gemma 4 based discrete text diffusion model for fast chat, reasoning, and tool use","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-09","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":250000,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/diffusiongemma-26b-a4b-it"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma-4-31B-IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0,"output":0}},"openai/whisper-large-v3":{"id":"openai/whisper-large-v3","name":"Whisper Large v3","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2023-09","release_date":"2023-09-01","last_updated":"2025-09-05","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":0,"output":4096},"cost":{"input":0,"output":0}},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0,"output":0}}}},"ovhcloud":{"id":"ovhcloud","env":["OVHCLOUD_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://oai.endpoints.kepler.ai.cloud.ovh.net/v1","name":"OVHcloud AI Endpoints","doc":"https://www.ovhcloud.com/en/public-cloud/ai-endpoints/catalog//","models":{"qwen3.5-9b":{"id":"qwen3.5-9b","name":"Qwen3.5-9B","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.12,"output":0.18},"canonical_model_id":"alibaba/qwen3.5-9b"},"mistral-nemo-instruct-2407":{"id":"mistral-nemo-instruct-2407","name":"Mistral-Nemo-Instruct-2407","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.14,"output":0.14},"canonical_model_id":"mistral/mistral-nemo"},"meta-llama-3_3-70b-instruct":{"id":"meta-llama-3_3-70b-instruct","name":"Meta-Llama-3_3-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.74,"output":0.74},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"gpt-oss-20b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.18},"canonical_model_id":"openai/gpt-oss-20b"},"mistral-small-3.2-24b-instruct-2506":{"id":"mistral-small-3.2-24b-instruct-2506","name":"Mistral-Small-3.2-24B-Instruct-2506","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.1,"output":0.31},"canonical_model_id":"mistral/mistral-small-3.2-24b-instruct"},"qwen3guard-gen-8b":{"id":"qwen3guard-gen-8b","name":"Qwen3Guard-Gen-8B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-22","last_updated":"2026-01-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0,"output":0}},"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8-27B","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.47,"output":3.19},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5-397B-A17B","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.71,"output":4.25},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen3guard-gen-0.6b":{"id":"qwen3guard-gen-0.6b","name":"Qwen3Guard-Gen-0.6B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-22","last_updated":"2026-01-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0,"output":0}},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"gpt-oss-120b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.09,"output":0.47},"canonical_model_id":"openai/gpt-oss-120b"},"qwen3.6-27b":{"id":"qwen3.6-27b","name":"Qwen3.6-27B","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.47,"output":3.19},"canonical_model_id":"alibaba/qwen3.6-27b"},"qwen2.5-vl-72b-instruct":{"id":"qwen2.5-vl-72b-instruct","name":"Qwen2.5-VL-72B-Instruct","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-01-26","last_updated":"2025-01-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":1.01,"output":1.01},"canonical_model_id":"alibaba/qwen2-5-vl-72b-instruct"}}},"jalapeno":{"id":"jalapeno","env":["JALAPENO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.jalapeno-cloud.ai/v1","name":"Jalapeno Cloud","doc":"https://www.jalapeno-cloud.ai/docs/","models":{"Qwen3-VL-235B-A22B-Instruct":{"id":"Qwen3-VL-235B-A22B-Instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":129024,"output":32768},"cost":{"input":0.3,"output":1.5},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":512000},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M3"},"Qwen3.5-35B-A3B":{"id":"Qwen3.5-35B-A3B","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"Kimi-K2.7-Code":{"id":"Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":271360,"output":262144},"cost":{"input":0.95,"output":4},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"GLM-5.1":{"id":"GLM-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1.38,"output":4.4},"canonical_model_id":"zhipuai/glm-5.1"},"Kimi-K3":{"id":"Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15},"canonical_model_id":"moonshotai/kimi-k3"},"Qwen3-Next-80B-A3B-Instruct":{"id":"Qwen3-Next-80B-A3B-Instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":129024,"output":32768},"cost":{"input":0.15,"output":1.5},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"Qwen3-VL-235B-A22B-Thinking":{"id":"Qwen3-VL-235B-A22B-Thinking","name":"Qwen3 VL 235B A22B Thinking","description":"Qwen vision-language thinking model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.98,"output":3.95},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-thinking"},"Qwen3.5-122B-A10B":{"id":"Qwen3.5-122B-A10B","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":3.2},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"Hy3":{"id":"Hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"input":192000,"output":128000},"cost":{"input":0.14,"output":0.58},"canonical_model_id":"tencent/hy3"},"GLM-5.2":{"id":"GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4},"canonical_model_id":"zhipuai/glm-5.2"},"Qwen3-Next-80B-A3B-Thinking":{"id":"Qwen3-Next-80B-A3B-Thinking","name":"Qwen3-Next 80B-A3B (Thinking)","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":1.5},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"DeepSeek-V4-Flash":{"id":"DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.14,"output":0.28},"canonical_model_id":"deepseek/deepseek-v4-flash"},"Qwen3.5-27B":{"id":"Qwen3.5-27B","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":2.4},"canonical_model_id":"alibaba/qwen3.5-27b"},"Qwen3.5-397B-A17B":{"id":"Qwen3.5-397B-A17B","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"DeepSeek-V4-Pro":{"id":"DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.6,"output":3.38},"canonical_model_id":"deepseek/deepseek-v4-pro"},"Kimi-K2.5":{"id":"Kimi-K2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":180224},"cost":{"input":0.6,"output":3},"canonical_model_id":"moonshotai/kimi-k2.5"}}},"abacus":{"id":"abacus","env":["ABACUS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://routellm.abacus.ai/v1","name":"Abacus","doc":"https://abacus.ai/help/api","models":{"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4},"canonical_model_id":"openai/o4-mini"},"gpt-4o-mini":{"id":"gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-4o-mini"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"grok-code-fast-1":{"id":"grok-code-fast-1","name":"Grok Code Fast 1","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-09-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":16384},"cost":{"input":0.2,"output":1.5}},"gpt-5.3-chat-latest":{"id":"gpt-5.3-chat-latest","name":"GPT-5.3 Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-01","last_updated":"2026-03-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.75,"output":14}},"claude-3-7-sonnet-20250219":{"id":"claude-3-7-sonnet-20250219","name":"Claude Sonnet 3.7","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-10-31","release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15}},"gemini-3.1-flash-image":{"id":"gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":1048576,"output":32768},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image"},"claude-opus-4-1-20250805":{"id":"claude-opus-4-1-20250805","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75}},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"gemini-3-pro-image-preview":{"id":"gemini-3-pro-image-preview","type":"image","name":"Nano Banana Pro Preview","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"google/gemini-3-pro-image-preview"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3-flash-preview"},"gemini-3.1-flash-image-preview":{"id":"gemini-3.1-flash-image-preview","type":"image","name":"Nano Banana 2 Preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":1048576,"output":32768},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image-preview"},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50},"canonical_model_id":"anthropic/claude-fable-5"},"grok-4-0709":{"id":"grok-4-0709","name":"Grok 4","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":16384},"cost":{"input":3,"output":15}},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"claude-opus-4-5-20251101":{"id":"claude-opus-4-5-20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"o3-pro":{"id":"o3-pro","name":"o3-pro","description":"High-effort o3 tier for difficult technical reasoning and careful answers","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":20,"output":40},"canonical_model_id":"openai/o3-pro"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.6-terra"},"grok-4-fast-non-reasoning":{"id":"grok-4-fast-non-reasoning","name":"Grok 4 Fast (Non-Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":16384},"cost":{"input":0.2,"output":0.5}},"o3-mini":{"id":"o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"google/gemini-2.5-pro"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-6"},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"gpt-5-codex":{"id":"gpt-5-codex","name":"GPT-5-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5-codex"},"claude-sonnet-4-5-20250929":{"id":"claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025},"canonical_model_id":"google/gemini-3.1-flash-lite"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"claude-sonnet-4-20250514":{"id":"claude-sonnet-4-20250514","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-05-14","last_updated":"2025-05-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15}},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":1,"output":6,"cache_read":0.1},"canonical_model_id":"openai/gpt-5.6-luna"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"google/gemini-3.1-pro-preview"},"gpt-5.1-chat-latest":{"id":"gpt-5.1-chat-latest","name":"GPT-5.1 Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.25,"output":10}},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"gemini-3-pro-image":{"id":"gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"google/gemini-3-pro-image"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"google/gemini-3.5-flash"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"o3":{"id":"o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"kimi-k2-turbo-preview":{"id":"kimi-k2-turbo-preview","name":"Kimi K2 Turbo Preview","description":"Fast Kimi model for responsive chat, coding help, and agent loops","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-08","last_updated":"2025-07-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":8192},"cost":{"input":0.15,"output":8}},"route-llm":{"id":"route-llm","name":"RouteLLM","description":"RouteLLM routes prompts to an appropriate Abacus-backed text-generation model","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-01-01","last_updated":"2026-07-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"cost":{"input":3,"output":15}},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"llama-3.3-70b-versatile":{"id":"llama-3.3-70b-versatile","name":"Llama 3.3 70B Versatile","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.59,"output":0.79}},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":32768},"cost":{"input":2,"output":6},"canonical_model_id":"xai/grok-4.5"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.6-sol"},"muse-spark-1.1":{"id":"muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-7"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-5.2-chat-latest":{"id":"gpt-5.2-chat-latest","name":"GPT-5.2 Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2026-01-01","last_updated":"2026-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.75,"output":14}},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":32768},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"qwen-2.5-coder-32b":{"id":"qwen-2.5-coder-32b","name":"Qwen 2.5 Coder 32B","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-11-11","last_updated":"2024-11-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.79,"output":0.79}},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2.5,"output":7.5},"canonical_model_id":"alibaba/qwen3.7-max"},"grok-4-1-fast-non-reasoning":{"id":"grok-4-1-fast-non-reasoning","name":"Grok 4.1 Fast (Non-Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-11-17","last_updated":"2025-11-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":16384},"cost":{"input":0.2,"output":0.5}},"gpt-5.1-codex-max":{"id":"gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10}},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-2.5-flash"},"mimo-v2-pro":{"id":"mimo-v2-pro","name":"MiMo-V2-Pro","description":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1,"output":3,"cache_read":0.2},"canonical_model_id":"xiaomi/mimo-v2-pro"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.6-flash"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.6,"output":3}},"gemini-3.1-flash-lite-preview":{"id":"gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-01","last_updated":"2026-03-01","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":1}},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-5.3-codex-xhigh":{"id":"gpt-5.3-codex-xhigh","name":"GPT-5.3 Codex XHigh","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14}},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"claude-opus-4-20250514":{"id":"claude-opus-4-20250514","name":"Claude Opus 4","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-05-14","last_updated":"2025-05-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75}},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen3-max":{"id":"qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"cost":{"input":1.2,"output":6}},"gemini-2.5-flash-image":{"id":"gemini-2.5-flash-image","type":"image","name":"Nano Banana","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":30},"canonical_model_id":"google/gemini-2.5-flash-image"},"gpt-4o-2024-11-20":{"id":"gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10}},"muse-spark-1.2":{"id":"muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.18},"canonical_model_id":"openai/gpt-5.3-codex"},"deepseek/deepseek-v3.1":{"id":"deepseek/deepseek-v3.1","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-01-20","last_updated":"2025-01-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.55,"output":1.66}},"deepseek-ai/DeepSeek-V3.2":{"id":"deepseek-ai/DeepSeek-V3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-06-15","last_updated":"2025-06-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.27,"output":0.4}},"deepseek-ai/DeepSeek-V4-Flash":{"id":"deepseek-ai/DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":0.14,"output":0.28,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":1.74,"output":3.48,"cache_read":0.15},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/DeepSeek-V3.1-Terminus":{"id":"deepseek-ai/DeepSeek-V3.1-Terminus","name":"DeepSeek V3.1 Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-06-01","last_updated":"2025-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.27,"output":1}},"deepseek-ai/DeepSeek-R1":{"id":"deepseek-ai/DeepSeek-R1","name":"DeepSeek R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-01-20","last_updated":"2025-01-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":3,"output":7}},"meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo":{"id":"meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo","name":"Llama 3.1 405B Instruct Turbo","description":"Compact Llama instruction model for fast chat and local deployment","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":3.5,"output":3.5}},"meta-llama/Meta-Llama-3.3-70B-Instruct":{"id":"meta-llama/Meta-Llama-3.3-70B-Instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.59,"output":0.79},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8":{"id":"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8","name":"Llama 4 Maverick 17B Instruct","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":8192},"cost":{"input":0.14,"output":0.59},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"meta-llama/Meta-Llama-3.1-8B-Instruct":{"id":"meta-llama/Meta-Llama-3.1-8B-Instruct","name":"Llama 3.1 8B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.02,"output":0.05}},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMaxAI/MiniMax-M2.7":{"id":"MiniMaxAI/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.7"},"thinkingmachines/Inkling":{"id":"thinkingmachines/Inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":3.74,"output":9.36,"cache_read":0.748},"canonical_model_id":"thinkingmachines/inkling"},"Qwen/QwQ-32B":{"id":"Qwen/QwQ-32B","name":"QwQ 32B","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2024-11-28","last_updated":"2024-11-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.4,"output":0.4}},"Qwen/Qwen2.5-72B-Instruct":{"id":"Qwen/Qwen2.5-72B-Instruct","name":"Qwen 2.5 72B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-09-19","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.11,"output":0.38}},"Qwen/Qwen3-Coder-480B-A35B-Instruct":{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct","name":"Qwen3-Coder 480B-A35B Instruct","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.29,"output":1.2},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"Qwen/Qwen3.6-27B":{"id":"Qwen/Qwen3.6-27B","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.32,"output":3.2},"canonical_model_id":"alibaba/qwen3.6-27b"},"Qwen/Qwen3-32B":{"id":"Qwen/Qwen3-32B","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.09,"output":0.29},"canonical_model_id":"alibaba/qwen3-32b"},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B A22B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-07-01","last_updated":"2025-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.13,"output":0.6}},"zai-org/GLM-4.7":{"id":"zai-org/GLM-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2},"canonical_model_id":"zhipuai/glm-4.7"},"zai-org/GLM-4.6":{"id":"zai-org/GLM-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.6,"output":2.2},"canonical_model_id":"zhipuai/glm-4.6"},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-5":{"id":"zai-org/GLM-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1,"output":3.2},"canonical_model_id":"zhipuai/glm-5"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-4.5":{"id":"zai-org/GLM-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":96000},"cost":{"input":0.6,"output":2.2},"canonical_model_id":"zhipuai/glm-4.5"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.08,"output":0.44},"canonical_model_id":"openai/gpt-oss-120b"}}},"trustedrouter":{"id":"trustedrouter","env":["TRUSTEDROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.trustedrouter.com/v1","name":"TrustedRouter","doc":"https://trustedrouter.com/docs","models":{"trustedrouter/synth":{"id":"trustedrouter/synth","name":"Synth","description":"TrustedRouter synthesis orchestration alias that combines multiple model responses into one answer.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-20","last_updated":"2026-06-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072}},"trustedrouter/fast":{"id":"trustedrouter/fast","name":"Fast","description":"TrustedRouter speed routing alias that prefers low-latency healthy model endpoints.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072}},"trustedrouter/auto":{"id":"trustedrouter/auto","name":"Auto","description":"TrustedRouter automatic routing alias that chooses a healthy supported model endpoint for the request.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-01","last_updated":"2026-06-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072}},"trustedrouter/zdr":{"id":"trustedrouter/zdr","name":"Zero Data Retention","description":"TrustedRouter privacy routing alias that prefers zero data retention model endpoints.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072}},"trustedrouter/cheap":{"id":"trustedrouter/cheap","name":"Cheap","description":"TrustedRouter low-cost routing alias that prefers inexpensive healthy model endpoints.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-01","last_updated":"2026-06-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072}},"trustedrouter/e2e":{"id":"trustedrouter/e2e","name":"End-to-End Encrypted","description":"TrustedRouter privacy routing alias for end-to-end encrypted provider routes where available.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072}},"trustedrouter/synth-code":{"id":"trustedrouter/synth-code","name":"Synth Code","description":"TrustedRouter code synthesis orchestration alias that combines multiple model responses into one answer.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-20","last_updated":"2026-06-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072}}}},"xiaomi-token-plan-sgp":{"id":"xiaomi-token-plan-sgp","env":["XIAOMI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://token-plan-sgp.xiaomimimo.com/v1","name":"Xiaomi Token Plan (Singapore)","doc":"https://platform.xiaomimimo.com/#/docs","models":{"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"mimo-v2-pro":{"id":"mimo-v2-pro","name":"MiMo-V2-Pro","description":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2-pro"},"mimo-v2.5-tts":{"id":"mimo-v2.5-tts","name":"MiMo-V2.5-TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0}},"mimo-v2-tts":{"id":"mimo-v2-tts","name":"MiMo-V2-TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0}},"mimo-v2.5-tts-voiceclone":{"id":"mimo-v2.5-tts-voiceclone","name":"MiMo-V2.5-TTS-VoiceClone","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0}},"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"mimo-v2.5-tts-voicedesign":{"id":"mimo-v2.5-tts-voicedesign","name":"MiMo-V2.5-TTS-VoiceDesign","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0}},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5-pro"}}},"vancine":{"id":"vancine","env":["VANCINE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://vancine.com/v1","name":"Vancine","doc":"https://vancine.com/docs","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.12,"output":0.4,"cache_read":0.024},"canonical_model_id":"zhipuai/glm-5.3-flash"},"hy4-preview":{"id":"hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":64000},"cost":{"input":0.67,"output":2,"cache_read":0.034},"canonical_model_id":"tencent/hy4-preview"},"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.24,"output":0.96,"cache_read":0.048},"canonical_model_id":"minimax/MiniMax-M3"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":2.4,"output":12,"cache_read":0.24},"canonical_model_id":"moonshotai/kimi-k3"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.12,"output":0.38,"cache_read":0.013},"canonical_model_id":"alibaba/qwen3.8-flash"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.12,"output":3.52,"cache_read":0.208},"canonical_model_id":"zhipuai/glm-5.3"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.24,"output":0.96,"cache_read":0.0048},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.6,"output":4.8,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.8-max"}}},"daoxe":{"id":"daoxe","env":["DAOXE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://daoxe.com/v1","name":"DaoXE","doc":"https://daoxe.com/pricing","models":{"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"google/gemini-3.1-pro-preview"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.5"},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":5},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"}}},"zenifra":{"id":"zenifra","env":["ZENIFRA_AI_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://ai.zenifra.com/v1","name":"Zenifra","doc":"https://docs.zenifra.com","models":{"alibaba/qwen3.6-35b-a3b":{"id":"alibaba/qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"provider":{"shape":"completions"},"cost":{"input":0.19,"output":0.48},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"}}},"qihang-ai":{"id":"qihang-ai","env":["QIHANG_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.qhaigc.net/v1","name":"QiHang","doc":"https://www.qhaigc.net/docs","models":{"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.07,"output":0.43,"tiers":[{"input":0.07,"output":0.43,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":0.07,"output":0.43}}},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2}},"claude-opus-4-5-20251101":{"id":"claude-opus-4-5-20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":0.71,"output":3.57}},"claude-sonnet-4-5-20250929":{"id":"claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":0.43,"output":2.14}},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5-Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":0.04,"output":0.29}},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.14,"output":1.14}},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-10-01","last_updated":"2025-10-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":0.14,"output":0.71}},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.09,"output":0.71,"tiers":[{"input":0.09,"output":0.71,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":0.09,"output":0.71}}},"gemini-3-pro-preview":{"id":"gemini-3-pro-preview","name":"Gemini 3 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65000},"cost":{"input":0.57,"output":3.43}}}},"ollama-cloud":{"id":"ollama-cloud","env":["OLLAMA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://ollama.com/v1","name":"Ollama Cloud","doc":"https://docs.ollama.com/cloud","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"nemotron-3-nano:30b":{"id":"nemotron-3-nano:30b","name":"nemotron-3-nano:30b","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.06,"output":0.24},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"qwen3.5:397b":{"id":"qwen3.5:397b","name":"qwen3.5:397b","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"release_date":"2026-02-15","last_updated":"2026-02-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6}},"minimax-m2.7":{"id":"minimax-m2.7","name":"minimax-m2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":196608},"cost":{"input":0.3,"output":1.2,"cache_read":0.06}},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"deepseek-v4-pro","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.66,"output":1.98,"cache_read":0.022}},"kimi-k2.6":{"id":"kimi-k2.6","name":"kimi-k2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16}},"kimi-k3":{"id":"kimi-k3","name":"kimi-k3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"nemotron-3-super":{"id":"nemotron-3-super","name":"nemotron-3-super","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.015,"output":0.6,"cache_read":0.015},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"gemma4:31b":{"id":"gemma4:31b","name":"gemma4:31b","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.14,"output":0.4,"cache_read":0.05}},"nemotron-3-ultra":{"id":"nemotron-3-ultra","name":"nemotron-3-ultra","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"cost":{"input":0.1,"output":3,"cache_read":0.1},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"gpt-oss:20b":{"id":"gpt-oss:20b","name":"gpt-oss:20b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"release_date":"2025-08-05","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.07,"output":0.3,"cache_read":0.035}},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"gpt-oss:120b":{"id":"gpt-oss:120b","name":"gpt-oss:120b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"release_date":"2025-08-05","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.014}},"glm-5.1":{"id":"glm-5.1","name":"glm-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"release_date":"2026-03-27","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2}},"deepseek-v4.1-flash:1017":{"id":"deepseek-v4.1-flash:1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"deepseek-v4-flash","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.22,"output":0.66,"cache_read":0.007}},"minimax-m3.1-flash":{"id":"minimax-m3.1-flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":512000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"kimi-k2.7-code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"deepseek-v4-pro:0813":{"id":"deepseek-v4-pro:0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"minimax-m3":{"id":"minimax-m3","name":"minimax-m3","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax-m3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-31","last_updated":"2026-05-31","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.12}},"muse-spark-open":{"id":"muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"mistral-large-4":{"id":"mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":262144},"cost":{"input":1.36,"output":4.18,"cache_read":0.14},"canonical_model_id":"mistral/mistral-large-4"},"deepseek-v4-flash:0731":{"id":"deepseek-v4-flash:0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.22,"output":0.66,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":976000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"kimi-k2.5":{"id":"kimi-k2.5","name":"kimi-k2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144}},"mistral-large-3:675b":{"id":"mistral-large-3:675b","name":"mistral-large-3:675b","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"release_date":"2025-12-02","last_updated":"2026-01-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5}},"minimax-m2.5":{"id":"minimax-m2.5","name":"minimax-m2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"knowledge":"2025-01","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072}}}},"aiand":{"id":"aiand","env":["AIAND_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.aiand.com/v1","name":"ai&","doc":"https://docs.aiand.com/","models":{"deepseek-ai/deepseek-v4-pro":{"id":"deepseek-ai/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1,"output":2.5,"cache_read":0.25},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/deepseek-v4-flash":{"id":"deepseek-ai/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.15,"output":0.25,"cache_read":0.08},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-ai/deepseek-v4.1-flash":{"id":"deepseek-ai/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.3,"output":0.6,"cache_read":0.02},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.4,"output":3,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.32,"output":3.2,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.6-27b"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":12.5,"cache_read":0.5},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.75,"output":3.5,"cache_read":0.2},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"zai-org/glm-5.3-flash":{"id":"zai-org/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048550,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai-org/glm-5.3":{"id":"zai-org/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1,"output":4,"cache_read":0.3},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/glm-5.2":{"id":"zai-org/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1,"output":4,"cache_read":0.3},"canonical_model_id":"zhipuai/glm-5.2"},"motif-technologies/motif-3":{"id":"motif-technologies/motif-3","name":"Motif 3","description":"Motif 3 is a large-scale, decoder-only Mixture-of-Experts (MoE) language model with 314 billion total parameters and 13.2 billion parameters activated per token.","family":"motif","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-08-07","last_updated":"2026-08-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":2,"cache_read":0.2},"canonical_model_id":"motif-technologies/motif-3"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"google/gemma-4-31b-it"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.08},"canonical_model_id":"openai/gpt-oss-120b"}}},"chutes":{"id":"chutes","env":["CHUTES_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://llm.chutes.ai/v1","name":"Chutes","doc":"https://llm.chutes.ai/v1/models","models":{"Nemotron-3-Nano-Omni-30B-TEE":{"id":"Nemotron-3-Nano-Omni-30B-TEE","name":"Nemotron 3 Nano Omni 30B TEE","description":"Omni-modal model for text, vision, audio, and multimodal agent tasks","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":0},"cost":{"input":0.0245,"output":0.0978,"cache_read":0.0024499999999999995}},"unsloth/Mistral-Nemo-Instruct-2407-TEE":{"id":"unsloth/Mistral-Nemo-Instruct-2407-TEE","name":"Mistral Nemo Instruct 2407 TEE","description":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.0245,"output":0.0978,"cache_read":0.0024499999999999995},"canonical_model_id":"mistral/mistral-nemo"},"deepseek-ai/DeepSeek-V3.2-TEE":{"id":"deepseek-ai/DeepSeek-V3.2-TEE","name":"DeepSeek V3.2 TEE","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":1,"output":1,"cache_read":0.09999999999999998},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek-ai/DeepSeek-V4-Flash-0731-TEE":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731-TEE","name":"DeepSeek V4 Flash 0731 TEE","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.44,"output":1.32,"cache_read":0.04399999999999999},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"moonshotai/Kimi-K3-TEE":{"id":"moonshotai/Kimi-K3-TEE","name":"Kimi K3 TEE","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65535},"cost":{"input":1.85,"output":9.5,"cache_read":0.18499999999999997},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.6-TEE":{"id":"moonshotai/Kimi-K2.6-TEE","name":"Kimi K2.6 TEE","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65535},"cost":{"input":0.45,"output":1.95,"cache_read":0.04499999999999999},"canonical_model_id":"moonshotai/kimi-k2.6"},"Qwen/Qwen3.8-27B-TEE":{"id":"Qwen/Qwen3.8-27B-TEE","name":"Qwen3.8 27B TEE","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.039,"output":1.25,"cache_read":0.003899999999999999},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3-235B-A22B-Thinking-2507-TEE":{"id":"Qwen/Qwen3-235B-A22B-Thinking-2507-TEE","name":"Qwen3 235B A22B Thinking 2507 TEE","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07","last_updated":"2026-06-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.2989,"output":1.1957,"cache_read":0.029889999999999993}},"Qwen/Qwen3-32B-TEE":{"id":"Qwen/Qwen3-32B-TEE","name":"Qwen3 32B TEE","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":40960},"cost":{"input":0.104,"output":0.416,"cache_read":0.010399999999999998},"canonical_model_id":"alibaba/qwen3-32b"},"Qwen/Qwen3.5-397B-A17B-TEE":{"id":"Qwen/Qwen3.5-397B-A17B-TEE","name":"Qwen3.5 397B A17B TEE","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.45,"output":3,"cache_read":0.04499999999999999},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"Qwen/Qwen3.6-27B-TEE":{"id":"Qwen/Qwen3.6-27B-TEE","name":"Qwen3.6 27B TEE","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2,"cache_read":0.024999999999999994},"canonical_model_id":"alibaba/qwen3.6-27b"},"zai-org/GLM-5.2-TEE":{"id":"zai-org/GLM-5.2-TEE","name":"GLM 5.2 TEE","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65535},"cost":{"input":1.25,"output":3.95,"cache_read":0.12499999999999997},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.1-TEE":{"id":"zai-org/GLM-5.1-TEE","name":"GLM 5.1 TEE","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":65535},"cost":{"input":0.98,"output":3.08,"cache_read":0.09799999999999998},"canonical_model_id":"zhipuai/glm-5.1"},"google/gemma-4-31B-turbo-TEE":{"id":"google/gemma-4-31B-turbo-TEE","name":"gemma 4 31B turbo TEE","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.12,"output":0.37,"cache_read":0.011999999999999997},"canonical_model_id":"google/gemma-4-31b-it"}}},"meta":{"id":"meta","env":["META_MODEL_API_KEY"],"npm":"@ai-sdk/openai","api":"https://api.meta.ai/v1","name":"Meta","doc":"https://dev.meta.ai/docs","models":{"muse-spark-1.3-contributor":{"id":"muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.002},"canonical_model_id":"meta/muse-spark-1.3"},"muse-spark-1.2-contributor":{"id":"muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.002},"canonical_model_id":"meta/muse-spark-1.2"},"muse-spark-1.3":{"id":"muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"muse-spark-1.1":{"id":"muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"muse-wonder-1.0":{"id":"muse-wonder-1.0","name":"Muse Wonder","description":"Muse Wonder is a frontier multimodal reasoning model from Meta for personal and professional agents, long-horizon coding, and computer use. It combines Active Memory, which carries relevant context across conversations, with Active Focus, which keeps it working through blockers toward a goal.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"muse-spark-1.2":{"id":"muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"}}},"sambanova":{"id":"sambanova","env":["SAMBANOVA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.sambanova.ai/v1","name":"SambaNova","doc":"https://docs.sambanova.ai/cloud/docs/get-started/supported-models","models":{"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.6,"output":2.4,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"DeepSeek-V3.1":{"id":"DeepSeek-V3.1","name":"DeepSeek-V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":7168},"cost":{"input":3,"output":4.5},"canonical_model_id":"deepseek/deepseek-v3.1"},"Meta-Llama-3.3-70B-Instruct":{"id":"Meta-Llama-3.3-70B-Instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":3072},"cost":{"input":0.6,"output":1.2},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"DeepSeek-V3.2":{"id":"DeepSeek-V3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":7168},"cost":{"input":3,"output":4.5},"canonical_model_id":"deepseek/deepseek-v3.2"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.22,"output":0.59},"canonical_model_id":"openai/gpt-oss-120b"},"gemma-4-31B-it":{"id":"gemma-4-31B-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.38,"output":1.15},"canonical_model_id":"google/gemma-4-31b-it"}}},"tencent-tokenhub":{"id":"tencent-tokenhub","env":["TENCENT_TOKENHUB_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://tokenhub.tencentmaas.com/v1","name":"Tencent TokenHub","doc":"https://cloud.tencent.com/document/product/1823/130050","models":{"hy4-preview":{"id":"hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"hy3-preview":{"id":"hy3-preview","name":"Hy3 preview","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"tencent/hy3-preview"},"hy3":{"id":"hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"tencent/hy3"}}},"digitalocean":{"id":"digitalocean","env":["DIGITALOCEAN_ACCESS_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.do-ai.run/v1","name":"DigitalOcean","doc":"https://docs.digitalocean.com/products/gradient-ai-platform/details/models/","models":{"openai-gpt-5.4-pro":{"id":"openai-gpt-5.4-pro","name":"OpenAI GPT-5.4 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}}},"openai-gpt-5.3-codex":{"id":"openai-gpt-5.3-codex","name":"OpenAI GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"openai-gpt-oss-120b":{"id":"openai-gpt-oss-120b","name":"OpenAI GPT-oss-120b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.1,"output":0.7,"cache_read":0.02}},"anthropic-claude-5.5-haiku":{"id":"anthropic-claude-5.5-haiku","name":"Anthropic Claude 5.5 Haiku","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"anthropic/claude-haiku-5-5"},"nvidia-nemotron-3-super-120b":{"id":"nvidia-nemotron-3-super-120b","name":"NVIDIA Nemotron 3 Super 120B (Public Preview)","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":0.3,"output":0.65,"cache_read":0.06},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"openai-gpt-image-2":{"id":"openai-gpt-image-2","name":"OpenAI GPT Image 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-24","last_updated":"2025-04-24","modalities":{"input":["text","image"],"output":["image","text"]},"open_weights":false,"limit":{"context":0,"output":16384},"cost":{"input":8,"output":30}},"mistral-7b-instruct-v0.3":{"id":"mistral-7b-instruct-v0.3","name":"Mistral 7B Instruct v0.3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-05-22","last_updated":"2024-05-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768}},"openai-gpt-5.2":{"id":"openai-gpt-5.2","name":"OpenAI GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"openai-gpt-4o":{"id":"openai-gpt-4o","name":"OpenAI GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25}},"arcee-trinity-large-thinking":{"id":"arcee-trinity-large-thinking","name":"Arcee Trinity Large Thinking (Public Preview)","description":"Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use","family":"trinity","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.25,"output":0.9,"cache_read":0.06},"canonical_model_id":"arcee-ai/trinity-large-thinking"},"openai-gpt-5.2-pro":{"id":"openai-gpt-5.2-pro","name":"OpenAI GPT-5.2 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":21,"output":168}},"openai-gpt-5.6-terra":{"id":"openai-gpt-5.6-terra","name":"OpenAI GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"openai/gpt-5.6-terra"},"nemotron-3.5-super-vl":{"id":"nemotron-3.5-super-vl","name":"NVIDIA Nemotron 3.5 Super VL","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.3,"output":0.65,"cache_read":0.06}},"anthropic-claude-3.5-sonnet":{"id":"anthropic-claude-3.5-sonnet","name":"Claude 3.5 Sonnet","description":"Legacy model retained for compatibility with older integrations","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-06-20","last_updated":"2024-10-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"all-mini-lm-l6-v2":{"id":"all-mini-lm-l6-v2","name":"All-MiniLM-L6-v2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2021-08-30","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256,"output":384},"cost":{"input":0.009,"output":0}},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"Deepseek V4 Pro","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.74,"output":3.48,"cache_read":0.348}},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo V2.5 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.8,"output":3,"cache_read":0.16},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"stable-diffusion-3.5-large":{"id":"stable-diffusion-3.5-large","name":"Stable Diffusion 3.5 Large","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"stable-diffusion","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-10-22","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["image"]},"open_weights":true,"limit":{"context":256,"output":1},"cost":{"input":0.08,"output":0}},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"qwen3-tts-voicedesign":{"id":"qwen3-tts-voicedesign","name":"Qwen3 TTS VoiceDesign","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":32768,"output":1}},"openai-gpt-5-mini":{"id":"openai-gpt-5-mini","name":"OpenAI GPT-5 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"openai-gpt-6-luna":{"id":"openai-gpt-6-luna","name":"OpenAI GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02}},"canonical_model_id":"openai/gpt-6-luna"},"openai-gpt-5.4":{"id":"openai-gpt-5.4","name":"OpenAI GPT-5.4","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}}},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19}},"qwen3-embedding-0.6b":{"id":"qwen3-embedding-0.6b","name":"Qwen3 Embedding 0.6B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-03","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"output":1024},"status":"beta","cost":{"input":0.04,"output":0}},"llama3.3-70b-instruct":{"id":"llama3.3-70b-instruct","name":"Llama 3.3 Instruct (70B)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.65,"output":0.65}},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"anthropic-claude-opus-4.8":{"id":"anthropic-claude-opus-4.8","name":"Anthropic Claude Opus 4.8","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-05-28","last_updated":"2026-05-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"openai-o3-mini":{"id":"openai-o3-mini","name":"OpenAI o3 mini","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55}},"qwen3-coder-flash":{"id":"qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.45,"output":1.7,"cache_read":0.09}},"deepseek-v3":{"id":"deepseek-v3","name":"DeepSeek V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-12-26","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":131072}},"anthropic-claude-opus-4.6":{"id":"anthropic-claude-opus-4.6","name":"Anthropic Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}}},"anthropic-claude-fable-5":{"id":"anthropic-claude-fable-5","name":"Anthropic Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"release_date":"2026-06-09","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5}},"anthropic-claude-4.6-sonnet":{"id":"anthropic-claude-4.6-sonnet","name":"Anthropic Claude Sonnet 4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}}},"openai-gpt-6-sol":{"id":"openai-gpt-6-sol","name":"OpenAI GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"tiers":[{"input":4,"output":15,"cache_read":0.4,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4}},"canonical_model_id":"openai/gpt-6-sol"},"anthropic-claude-3-opus":{"id":"anthropic-claude-3-opus","name":"Claude 3 Opus","description":"Legacy model retained for compatibility with older integrations","family":"claude-opus","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-08","release_date":"2024-02-29","last_updated":"2024-02-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":4096},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"llama-4-maverick":{"id":"llama-4-maverick","name":"Llama 4 Maverick","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.25,"output":0.87,"cache_read":0.05}},"anthropic-claude-sonnet-4":{"id":"anthropic-claude-sonnet-4","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.3,"cache_write":3.75,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.3,"cache_write":3.75}}},"mistral-nemo-instruct-2407":{"id":"mistral-nemo-instruct-2407","name":"Mistral Nemo Instruct","description":"Legacy model retained for compatibility with older integrations","family":"mistral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":0.3,"output":0.3}},"openai-o1":{"id":"openai-o1","name":"OpenAI o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5}},"anthropic-claude-4.1-opus":{"id":"anthropic-claude-4.1-opus","name":"Anthropic Claude 4.1 Opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"openai-gpt-4o-mini":{"id":"openai-gpt-4o-mini","name":"OpenAI GPT-4o mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075}},"anthropic-claude-4.5-sonnet":{"id":"anthropic-claude-4.5-sonnet","name":"Anthropic Claude 4.5 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}}},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"glm-5.3":{"id":"glm-5.3","name":"GLM5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"anthropic-claude-3.5-haiku":{"id":"anthropic-claude-3.5-haiku","name":"Claude 3.5 Haiku","description":"Legacy model retained for compatibility with older integrations","family":"claude-haiku","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-11-05","last_updated":"2024-11-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"status":"deprecated","cost":{"input":0.8,"output":4,"cache_read":0.08,"cache_write":1}},"openai-gpt-4.1":{"id":"openai-gpt-4.1","name":"OpenAI GPT-4.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5}},"openai-gpt-oss-20b":{"id":"openai-gpt-oss-20b","name":"OpenAI GPT-oss-20b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.05,"output":0.45,"cache_read":0.01}},"openai-gpt-5-nano":{"id":"openai-gpt-5-nano","name":"OpenAI GPT-5 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005}},"openai-o3":{"id":"openai-o3","name":"OpenAI o3","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5}},"anthropic-claude-opus-5.5":{"id":"anthropic-claude-opus-5.5","name":"Anthropic Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.4,"cache_write":10,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.4,"cache_write":10}},"canonical_model_id":"anthropic/claude-opus-5-5"},"mistral-3-14B":{"id":"mistral-3-14B","name":"Ministral 3 14B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-15","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"cost":{"input":0.2,"output":0.2,"cache_read":0.04}},"qwen-2.5-14b-instruct":{"id":"qwen-2.5-14b-instruct","name":"Qwen 2.5 14B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2024-09-19","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072}},"multi-qa-mpnet-base-dot-v1":{"id":"multi-qa-mpnet-base-dot-v1","name":"Multi-QA-mpnet-base-dot-v1","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2021-08-30","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":768},"cost":{"input":0.009,"output":0}},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":1.3,"output":4.3,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"nemotron-3-nano-omni":{"id":"nemotron-3-nano-omni","name":"Nemotron 3 Nano Omni","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.5,"output":0.9,"cache_read":0.1},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"openai-gpt-image-1.5":{"id":"openai-gpt-image-1.5","name":"OpenAI GPT Image 1.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["image","text"]},"open_weights":false,"limit":{"context":0,"output":16384},"cost":{"input":5,"output":10,"cache_read":1}},"alibaba-qwen3-32b":{"id":"alibaba-qwen3-32b","name":"Qwen3 32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-30","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.25,"output":0.55}},"llama3-8b-instruct":{"id":"llama3-8b-instruct","name":"Llama 3.1 Instruct (8B)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.198,"output":0.198}},"deepseek-4-flash":{"id":"deepseek-4-flash","name":"Deepseek V4 Flash","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-05-27","last_updated":"2026-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.028}},"anthropic-claude-sonnet-5.5":{"id":"anthropic-claude-sonnet-5.5","name":"Anthropic Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"openai-gpt-5":{"id":"openai-gpt-5","name":"OpenAI GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"glm-5":{"id":"glm-5","name":"GLM 5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-11","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":64000},"cost":{"input":1,"output":3.2,"cache_read":0.2}},"openai-gpt-5.1-codex-max":{"id":"openai-gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"deepseek-r1-distill-llama-70b":{"id":"deepseek-r1-distill-llama-70b","name":"DeepSeek R1 Distill Llama 70B","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-01-30","last_updated":"2025-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32678,"output":8192},"cost":{"input":0.99,"output":0.99}},"ministral-3-8b-instruct-2512":{"id":"ministral-3-8b-instruct-2512","name":"Ministral 3 8B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144}},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen 3.5 397B A17B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-04-30","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.55,"output":3.5,"cache_read":0.111}},"openai-gpt-5.4-nano":{"id":"openai-gpt-5.4-nano","name":"OpenAI GPT-5.4 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02}},"gte-large-en-v1.5":{"id":"gte-large-en-v1.5","name":"GTE Large (v1.5)","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-03-27","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":1024},"cost":{"input":0.09,"output":0}},"anthropic-claude-opus-5":{"id":"anthropic-claude-opus-5","name":"Anthropic Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"e5-large-v2":{"id":"e5-large-v2","name":"E5 Large v2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-05-19","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":1024},"cost":{"input":0.02,"output":0}},"openai-gpt-5.4-mini":{"id":"openai-gpt-5.4-mini","name":"OpenAI GPT-5.4 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075}},"wan2-2-t2v-a14b":{"id":"wan2-2-t2v-a14b","name":"Wan2.2-T2V-A14B","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-07-28","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["video"]},"open_weights":true,"limit":{"context":100,"output":1},"cost":{"input":0.6,"output":0}},"anthropic-claude-fable-5.1":{"id":"anthropic-claude-fable-5.1","name":"Anthropic Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic-claude-5-sonnet":{"id":"anthropic-claude-5-sonnet","name":"Anthropic Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic-claude-opus-4.7":{"id":"anthropic-claude-opus-4.7","name":"Anthropic Claude Opus 4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"nemotron-nano-12b-v2-vl":{"id":"nemotron-nano-12b-v2-vl","name":"Nemotron-nano 12b v2-vl","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"temperature":true,"release_date":"2025-10-28","last_updated":"2025-10-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.2,"output":0.6,"cache_read":0.04},"canonical_model_id":"nvidia/nemotron-nano-12b-v2-vl"},"nemotron-3-ultra-550b":{"id":"nemotron-3-ultra-550b","name":"Nemotron 3 Ultra","description":"Flagship Nemotron model for high-throughput reasoning and complex agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-06-04","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":0.9,"output":1.7,"cache_read":0.18}},"bge-m3":{"id":"bge-m3","name":"BGE M3","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"bge","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-01-30","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":1024},"cost":{"input":0.02,"output":0}},"gemma-4-31B-it":{"id":"gemma-4-31B-it","name":"Gemma 4","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-30","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.18,"output":0.5,"cache_read":0.036}},"openai-gpt-image-1":{"id":"openai-gpt-image-1","name":"GPT Image 1","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-24","last_updated":"2025-04-24","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":40,"cache_read":1.25}},"deepseek-3.2":{"id":"deepseek-3.2","name":"Deepseek 3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2025-12-02","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.5,"output":1.6,"cache_read":0.15}},"nemotron-3-nano-30b":{"id":"nemotron-3-nano-30b","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.21},"canonical_model_id":"zhipuai/glm-5.2"},"anthropic-claude-haiku-4.5":{"id":"anthropic-claude-haiku-4.5","name":"Anthropic Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25}},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"openai-gpt-5.5":{"id":"openai-gpt-5.5","name":"OpenAI GPT-5.5","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}}},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":2.7,"cache_read":0.203}},"openai-gpt-6-1-sol":{"id":"openai-gpt-6-1-sol","name":"OpenAI GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"tiers":[{"input":4,"output":15,"cache_read":0.2,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai-gpt-6-astra":{"id":"openai-gpt-6-astra","name":"OpenAI GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"tiers":[{"input":20,"output":75,"cache_read":2,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2}},"canonical_model_id":"openai/gpt-6-astra"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8-Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":262144},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.8-max"},"anthropic-claude-3.7-sonnet":{"id":"anthropic-claude-3.7-sonnet","name":"Claude 3.7 Sonnet","description":"Legacy model retained for compatibility with older integrations","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-02-24","last_updated":"2025-02-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"minimax-m2.5":{"id":"minimax-m2.5","name":"MiniMax M2.5 (Public Preview)","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax-m2.5","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08","release_date":"2026-02-12","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.3,"output":1.2,"cache_read":0.06}},"bge-reranker-v2-m3":{"id":"bge-reranker-v2-m3","name":"BGE Reranker v2 M3","description":"Reranking model for improving retrieval quality in search and recommendation systems","family":"bge","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-03-12","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":1},"cost":{"input":0.01,"output":0}},"anthropic-claude-opus-4":{"id":"anthropic-claude-opus-4","name":"Claude Opus 4","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"openai-gpt-5.6-luna":{"id":"openai-gpt-5.6-luna","name":"OpenAI GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai-gpt-5.6-sol":{"id":"openai-gpt-5.6-sol","name":"OpenAI GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"tiers":[{"input":8,"output":30,"cache_read":0.8,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8}},"canonical_model_id":"openai/gpt-5.6-sol"},"anthropic-claude-opus-4.5":{"id":"anthropic-claude-opus-4.5","name":"Anthropic Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"fal-ai/fast-sdxl":{"id":"fal-ai/fast-sdxl","name":"Fast SDXL","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"stable-diffusion","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-07-26","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["image"]},"open_weights":true,"limit":{"context":0,"output":0}},"fal-ai/elevenlabs/tts/multilingual-v2":{"id":"fal-ai/elevenlabs/tts/multilingual-v2","name":"ElevenLabs Multilingual TTS v2","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"elevenlabs","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-08-22","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"fal-ai/stable-audio-25/text-to-audio":{"id":"fal-ai/stable-audio-25/text-to-audio","name":"Stable Audio 2.5 (Text-to-Audio)","description":"Speech generation model for controllable voice, narration, and audio delivery","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-08","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"fal-ai/flux/schnell":{"id":"fal-ai/flux/schnell","name":"FLUX.1 [schnell]","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-08-01","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["image"]},"open_weights":true,"limit":{"context":0,"output":0}}}},"aixy":{"id":"aixy","env":["AIXY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.aixy-gateway.com/v1","name":"Aixy","doc":"https://docs.aixy-gateway.com/integrations/overview","models":{"anthropic/claude-fable-5-1":{"id":"anthropic/claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.aixy-gateway.com/v1"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-sonnet-5-5":{"id":"anthropic/claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.aixy-gateway.com/v1"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-haiku-5-5":{"id":"anthropic/claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.aixy-gateway.com/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-5-5":{"id":"anthropic/claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.aixy-gateway.com/v1"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gemini/gemini-3.8-flash":{"id":"gemini/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.8-flash"},"gemini/gemini-4-argon":{"id":"gemini/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":32768},"cost":{"input":2,"output":10,"cache_read":0.1,"input_audio":2},"canonical_model_id":"google/gemini-4-argon"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.aixy-gateway.com/v1","shape":"responses"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.aixy-gateway.com/v1","shape":"responses"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.aixy-gateway.com/v1","shape":"responses"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.aixy-gateway.com/v1","shape":"responses"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.aixy-gateway.com/v1","shape":"responses"},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.aixy-gateway.com/v1","shape":"responses"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.aixy-gateway.com/v1","shape":"responses"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"}}},"wafer.ai":{"id":"wafer.ai","env":["WAFER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://pass.wafer.ai/v1","name":"Wafer","doc":"https://docs.wafer.ai/wafer-pass","models":{"Kimi-K2.6":{"id":"Kimi-K2.6","name":"Kimi K2.6","description":"Kimi K2.6 sparse MoE model with a 262K context window. Available serverless and not included in standard Wafer Pass. Non-ZDR only: requests with `Wafer-ZDR: required` are rejected.","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":1.14,"output":4.8,"cache_read":0.19,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.33,"output":1.32,"cache_read":0.07,"cache_write":0,"tiers":[{"input":0.66,"output":2.64,"cache_read":0.13,"cache_write":0,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":0.66,"output":2.64,"cache_read":0.13,"cache_write":0}},"canonical_model_id":"minimax/MiniMax-M3"},"glm5.2-fast":{"id":"glm5.2-fast","name":"GLM5.2-Fast","description":"The same model served for high TPS.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":10.25,"cache_read":0.5,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"GLM-5.1":{"id":"GLM-5.1","name":"GLM-5.1","description":"General Language Model 5.1 — high-quality bilingual (EN/ZH) generation with strong coding and reasoning capabilities.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-07","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.1,"cache_write":0}},"GLM-5.2":{"id":"GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.2,"output":4.1,"cache_read":0.2,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"}}},"watsonx":{"id":"watsonx","env":["WATSONX_AI_APIKEY","WATSONX_AI_PROJECT_ID"],"npm":"watsonx-ai-provider","name":"watsonx.ai","doc":"https://www.ibm.com/docs/en/watsonx/saas?topic=solutions-supported-foundation-models","models":{"meta-llama/llama-3-3-70b-instruct":{"id":"meta-llama/llama-3-3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.7526,"output":0.7526},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/llama-4-maverick-17b-128e-instruct-fp8":{"id":"meta-llama/llama-4-maverick-17b-128e-instruct-fp8","name":"Llama 4 Maverick 17B 128E Instruct FP8","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.371,"output":1.484},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"mistralai/mistral-small-3-1-24b-instruct-2503":{"id":"mistralai/mistral-small-3-1-24b-instruct-2503","name":"Mistral Small 3.1 24B","description":"Efficient multimodal model for instruction following, coding, reasoning, and function calling","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.106,"output":0.318},"canonical_model_id":"mistral/mistral-small-3-1-24b-instruct-2503"},"ibm/granite-4-h-small":{"id":"ibm/granite-4-h-small","name":"Granite-4.0-H-Small","description":"Open-weight hybrid model for enterprise chat, coding, retrieval-augmented generation, and tool-calling workloads","family":"granite","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-02","last_updated":"2025-10-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.0636,"output":0.265},"canonical_model_id":"ibm/granite-4-h-small"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.159,"output":0.636},"canonical_model_id":"openai/gpt-oss-120b"}}},"impossibl":{"id":"impossibl","env":["IMPOSSIBL_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.impossibl.com/v1","name":"Impossibl","doc":"https://impossibl.com/docs/models","models":{"anthropic/claude-haiku-4-5":{"id":"anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024},{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-4-8":{"id":"anthropic/claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-sonnet-4-5":{"id":"anthropic/claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024},{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-opus-4-5":{"id":"anthropic/claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"xai/grok-4.3":{"id":"xai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"xai/grok-4.20-0309-reasoning":{"id":"xai/grok-4.20-0309-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"xai/grok-build-0.1":{"id":"xai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2},"canonical_model_id":"xai/grok-build-0.1"},"xai/grok-4.5":{"id":"xai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3},"canonical_model_id":"xai/grok-4.5"},"xai/grok-4.20-0309-non-reasoning":{"id":"xai/grok-4.20-0309-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.74,"output":3.48,"cache_read":0.145},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.19,"output":0.51,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.05,"tiers":[{"input":1,"output":4,"cache_read":0.2,"tier":{"type":"context","size":262144}}],"context_over_200k":{"input":1,"output":4,"cache_read":0.2}},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3.8-max-preview":{"id":"qwen/qwen3.8-max-preview","name":"Qwen3.8 Max Preview","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-19","last_updated":"2026-07-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.5,"output":7.5},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.24,"tier":{"type":"context","size":262144}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.24}},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.5},"canonical_model_id":"alibaba/qwen3.7-max"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"fireworks/gpt-oss-20b":{"id":"fireworks/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.07,"output":0.3,"cache_read":0.035},"canonical_model_id":"openai/gpt-oss-20b"},"fireworks/gpt-oss-120b":{"id":"fireworks/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"openai/gpt-oss-120b"},"fireworks/glm-5.2":{"id":"fireworks/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":1.87,"output":4.68,"cache_read":0.374},"canonical_model_id":"thinkingmachines/inkling"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.003},"canonical_model_id":"xiaomi/mimo-v2.5"},"groq/gpt-oss-20b":{"id":"groq/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-oss-20b"},"groq/gpt-oss-120b":{"id":"groq/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-oss-120b"},"zai/glm-4.5-air":{"id":"zai/glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.2,"output":1.1,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-4.5-air"},"zai/glm-4.7":{"id":"zai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"zai/glm-5.1":{"id":"zai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai/glm-5":{"id":"zai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"zai/glm-5-turbo":{"id":"zai/glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"zai/glm-4.6":{"id":"zai/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"zai/glm-5.2":{"id":"zai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai/glm-4.5":{"id":"zai/glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.6-flash"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"cache_read":3,"tiers":[{"input":60,"output":270,"cache_read":3,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270,"cache_read":3}},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5-turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5-codex":{"id":"openai/gpt-5-codex","name":"GPT-5-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5-codex"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":45,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"cache_read":3,"tiers":[{"input":60,"output":270,"cache_read":3,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270,"cache_read":3}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"cerebras/gpt-oss-120b":{"id":"cerebras/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.35,"output":0.75},"canonical_model_id":"openai/gpt-oss-120b"}}},"crusoe":{"id":"crusoe","env":["CRUSOE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.inference.crusoecloud.com/v1","name":"Crusoe","doc":"https://docs.crusoecloud.com/managed-inference/overview","models":{"nvidia/Nemotron-3-Nano-Omni-Reasoning-30B-A3B":{"id":"nvidia/Nemotron-3-Nano-Omni-Reasoning-30B-A3B","name":"Nemotron 3 Nano Omni 30B A3B Reasoning","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.3,"output":1.83,"cache_read":0.3,"input_audio":0.5},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"nvidia/NVIDIA-Nemotron-3-Super-120B-A12B":{"id":"nvidia/NVIDIA-Nemotron-3-Super-120B-A12B","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":2.4,"cache_read":0.15},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B":{"id":"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.05,"output":0.2,"cache_read":0.03},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"deepseek-ai/DeepSeek-V3-0324":{"id":"deepseek-ai/DeepSeek-V3-0324","name":"DeepSeek V3 0324","description":"March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.5,"output":1.5,"cache_read":0.25},"canonical_model_id":"deepseek/deepseek-v3-0324"},"meta-llama/Llama-3.3-70B-Instruct":{"id":"meta-llama/Llama-3.3-70B-Instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.25,"output":0.75,"cache_read":0.13},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.7,"output":3.5,"cache_read":0.35},"canonical_model_id":"moonshotai/kimi-k2.6"},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B-A22B Instruct 2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.22,"output":0.8,"cache_read":0.11},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"zai/GLM-5.1":{"id":"zai/GLM-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4.4,"cache_read":0.25},"canonical_model_id":"zhipuai/glm-5.1"},"zai/GLM-5.2":{"id":"zai/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.14,"output":0.4,"cache_read":0.14},"canonical_model_id":"google/gemma-4-31b-it"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.05,"output":0.2,"cache_read":0.05},"canonical_model_id":"openai/gpt-oss-120b"}}},"arcee":{"id":"arcee","env":["ARCEE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.arcee.ai/api/v1","name":"Arcee","doc":"https://docs.arcee.ai","models":{"trinity-large-thinking":{"id":"trinity-large-thinking","name":"Trinity Large Thinking","description":"Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use","family":"trinity","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-01","last_updated":"2026-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"beta","cost":{"input":0.25,"output":0.8,"cache_read":0.06},"canonical_model_id":"arcee-ai/trinity-large-thinking"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":384000},"status":"beta","cost":{"input":1.74,"output":3.48,"cache_read":0.2},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"status":"beta","cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-flash-latest":{"id":"deepseek/deepseek-v4-flash-latest","name":"DeepSeek V4 Flash Latest","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"status":"beta","cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"beta","cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"beta","cost":{"input":0.5,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"zai-org/glm-5.2":{"id":"zai-org/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"status":"beta","cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"}}},"tencent-token-plan":{"id":"tencent-token-plan","env":["TENCENT_TOKEN_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.lkeap.cloud.tencent.com/plan/v3","name":"Tencent Token Plan","doc":"https://cloud.tencent.com/document/product/1823/130060","models":{"hy4-preview":{"id":"hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"hy3":{"id":"hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"tencent/hy3"}}},"drun":{"id":"drun","env":["DRUN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://chat.d.run/v1","name":"D.Run (China)","doc":"https://www.d.run","models":{"public/deepseek-v3":{"id":"public/deepseek-v3","name":"DeepSeek V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.28,"output":1.1}},"public/deepseek-r1":{"id":"public/deepseek-r1","name":"DeepSeek R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2025-01-20","last_updated":"2025-01-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32000},"cost":{"input":0.55,"output":2.2}},"public/minimax-m25":{"id":"public/minimax-m25","name":"MiniMax M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2025-03-01","last_updated":"2025-03-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":204800,"output":131072},"cost":{"input":0.29,"output":1.16}}}},"snowflake-cortex":{"id":"snowflake-cortex","env":["SNOWFLAKE_ACCOUNT","SNOWFLAKE_CORTEX_PAT"],"npm":"@ai-sdk/openai-compatible","api":"https://${SNOWFLAKE_ACCOUNT}.snowflakecomputing.com/api/v2/cortex/v1","name":"Snowflake Cortex","doc":"https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-rest-api","models":{"openai-gpt-5.2":{"id":"openai-gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"canonical_model_id":"openai/gpt-5.2"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":16384},"canonical_model_id":"anthropic/claude-haiku-4-5"},"openai-gpt-5.6-terra":{"id":"openai-gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"status":"beta","canonical_model_id":"openai/gpt-5.6-terra"},"openai-gpt-5.1":{"id":"openai-gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"input":272000,"output":8192},"canonical_model_id":"openai/gpt-5.1"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"canonical_model_id":"anthropic/claude-fable-5"},"openai-gpt-5-mini":{"id":"openai-gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"input":272000,"output":8192},"status":"beta","canonical_model_id":"openai/gpt-5-mini"},"openai-gpt-5.4":{"id":"openai-gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"status":"beta","experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"canonical_model_id":"openai/gpt-5.4"},"mistral-large2":{"id":"mistral-large2","name":"Mistral Large (latest)","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"canonical_model_id":"mistral/mistral-large-latest"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"canonical_model_id":"anthropic/claude-opus-4-6"},"gemini-3.1-pro":{"id":"gemini-3.1-pro","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-3.1-pro-preview"},"deepseek-r1":{"id":"deepseek-r1","name":"DeepSeek-R1","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"canonical_model_id":"deepseek/deepseek-r1"},"openai-gpt-4.1":{"id":"openai-gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000},"canonical_model_id":"openai/gpt-4.1"},"openai-gpt-5-nano":{"id":"openai-gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"input":272000,"output":8192},"status":"beta","canonical_model_id":"openai/gpt-5-nano"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"canonical_model_id":"anthropic/claude-opus-4-8"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":16384},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"openai-gpt-5":{"id":"openai-gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"input":272000,"output":8192},"status":"beta","canonical_model_id":"openai/gpt-5"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":16384},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"beta","experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"canonical_model_id":"anthropic/claude-opus-4-7"},"snowflake-llama3.3-70b":{"id":"snowflake-llama3.3-70b","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"canonical_model_id":"anthropic/claude-opus-5"},"openai-gpt-5.5":{"id":"openai-gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"status":"beta","canonical_model_id":"openai/gpt-5.5"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"canonical_model_id":"anthropic/claude-sonnet-5"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"canonical_model_id":"anthropic/claude-opus-4-5"},"openai-gpt-5.6-luna":{"id":"openai-gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"status":"beta","canonical_model_id":"openai/gpt-5.6-luna"},"openai-gpt-5.6-sol":{"id":"openai-gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"status":"beta","canonical_model_id":"openai/gpt-5.6-sol"}}},"alibaba-token-plan-cn":{"id":"alibaba-token-plan-cn","env":["ALIBABA_TOKEN_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://token-plan.cn-beijing.maas.aliyuncs.com/compatible-mode/v1","name":"Alibaba Token Plan (China)","doc":"https://www.alibabacloud.com/help/zh/model-studio/token-plan-overview","models":{"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":131072}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":131072}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen-image-2.0-pro":{"id":"qwen-image-2.0-pro","type":"image","name":"Qwen Image 2.0 Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"qwen3.8-max-preview":{"id":"qwen3.8-max-preview","name":"Qwen3.8 Max Preview","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-19","last_updated":"2026-07-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.8-flash"},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"input":196601,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"minimax/MiniMax-M2.5"},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.7-plus"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"happyhorse-1.1-r2v":{"id":"happyhorse-1.1-r2v","type":"video","name":"HappyHorse 1.1 Reference-to-Video","description":"Video model for reference-guided video generation","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-17","last_updated":"2026-07-17","modalities":{"input":["image","text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"wan2.7-image":{"id":"wan2.7-image","type":"image","name":"Wan2.7 Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"wan2.7-image-pro":{"id":"wan2.7-image-pro","type":"image","name":"Wan2.7 Image Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"glm-5":{"id":"glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":16384},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-03","last_updated":"2025-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"status":"deprecated","cost":{"input":0,"output":0}},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"happyhorse-1.1-t2v":{"id":"happyhorse-1.1-t2v","type":"video","name":"HappyHorse 1.1 Text-to-Video","description":"Video model for prompt-driven text-to-video generation","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-17","last_updated":"2026-07-17","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.7-max"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.5"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen-image-2.0":{"id":"qwen-image-2.0","type":"image","name":"Qwen Image 2.0","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"happyhorse-1.1-i2v":{"id":"happyhorse-1.1-i2v","type":"video","name":"HappyHorse 1.1 Image-to-Video","description":"Video model for image-to-video generation","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-17","last_updated":"2026-07-17","modalities":{"input":["image","text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}}}},"runinfra":{"id":"runinfra","env":["RUNINFRA_GATEWAY_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.runinfra.ai/v1","name":"RunInfra","doc":"https://runinfra.ai/docs","models":{"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16":{"id":"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16","name":"Nemotron 3.5 Lightning 30B A3B","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.05,"output":0.15,"cache_read":0.01},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"Inferact/Qwen3.8-2.4T-A95B-NVFP4":{"id":"Inferact/Qwen3.8-2.4T-A95B-NVFP4","name":"Qwen3.8 2.4T A95B (NVFP4)","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.13,"output":0.27,"cache_read":0.01},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.6,"output":1.9,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"ornith-ai/Ornith-1.5-35B-A3B":{"id":"ornith-ai/Ornith-1.5-35B-A3B","name":"Ornith 1.5 35B A3B","description":"Mixture-of-experts coding-reasoning model for agentic software tasks, tool use, and image understanding","family":"ornith","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-18","last_updated":"2026-08-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"deepreinforce/ornith-1.5-35b-a3b"},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"alibaba/qwen3.8-27b"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":1.3,"output":4.2,"cache_read":0.13},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-5.3-flash"}}},"vispark":{"id":"vispark","env":["VISPARK_LAB_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.lab.vispark.in/v1","name":"Vispark","doc":"https://lab.vispark.in/#vision","models":{"vispark/vision-small":{"id":"vispark/vision-small","name":"Vision Small","description":"Fast, low-cost multimodal model for understanding text, images, audio, video, and PDFs, with tool calling and a 1M-token context window.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-05-15","last_updated":"2026-09","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.05,"output":3.16}},"vispark/vision-large":{"id":"vispark/vision-large","name":"Vision Large","description":"Most capable Vision model for complex reasoning, detailed media analysis, and structured output over a 1M-token context window.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-05-15","last_updated":"2026-09","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":7.37,"output":22.11}},"vispark/vision-medium":{"id":"vispark/vision-medium","name":"Vision Medium","description":"Balanced multimodal model pairing a 1M-token context window with deeper reasoning for analysis, content creation, and tool use across text, image, audio, video, and PDF inputs.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-05-15","last_updated":"2026-09","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":4.21,"output":12.63}}}},"hetzner":{"id":"hetzner","env":["HETZNER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.hetzner.com/api/v1","name":"Hetzner","doc":"https://experiments.hetzner.com/docs/inference","models":{"Qwen3.8-27B":{"id":"Qwen3.8-27B","name":"Qwen3.8-27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"beta","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3.6-35B-A3B-FP8":{"id":"Qwen/Qwen3.6-35B-A3B-FP8","name":"Qwen3.6 35B A3B FP8","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"beta","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"}}},"moonshotai":{"id":"moonshotai","env":["MOONSHOT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.moonshot.ai/v1","name":"Moonshot AI","doc":"https://platform.moonshot.ai/docs/api/chat","models":{"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3},"canonical_model_id":"moonshotai/kimi-k3"},"kimi-k2.7-code-highspeed":{"id":"kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code HighSpeed","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"}}},"berget":{"id":"berget","env":["BERGET_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.berget.ai/v1","name":"Berget.AI","doc":"https://api.berget.ai","models":{"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-27","last_updated":"2026-07-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":327680,"output":32768},"cost":{"input":3,"output":15},"canonical_model_id":"moonshotai/kimi-k3"},"mistralai/Mistral-Small-3.2-24B-Instruct-2506":{"id":"mistralai/Mistral-Small-3.2-24B-Instruct-2506","name":"Mistral Small 3.2 24B Instruct 2506","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-10-01","last_updated":"2025-10-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":8192},"cost":{"input":0.33,"output":0.33}},"Qwen/Qwen3.8-27B-FP8":{"id":"Qwen/Qwen3.8-27B-FP8","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.46,"output":3.48},"canonical_model_id":"alibaba/qwen3.8-27b"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":32768},"cost":{"input":1.54,"output":4.84},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-09-30","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":16384},"cost":{"input":0.29,"output":0.58},"canonical_model_id":"zhipuai/glm-5.3-flash"},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B Instruct","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["audio","image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.275,"output":0.55}}}},"llmtech":{"id":"llmtech","env":["LLMTECH_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.llmtech.eu/v1","name":"LLM Tech","doc":"https://llmtech.eu/models/qwen3.8-27b","models":{"nvidia/Qwen3.8-27B-NVFP4":{"id":"nvidia/Qwen3.8-27B-NVFP4","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.25,"output":2.09,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.8-27b"}}},"alibaba-cn":{"id":"alibaba-cn","env":["DASHSCOPE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://dashscope.aliyuncs.com/compatible-mode/v1","name":"Alibaba (China)","doc":"https://www.alibabacloud.com/help/en/model-studio/models","models":{"qwen3.7-flash":{"id":"qwen3.7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":131072},"cost":{"input":0.02962,"output":0.1185,"cache_read":0.002962,"cache_write":0.03703,"tiers":[{"input":0.08887,"output":0.35549,"cache_read":0.008887,"cache_write":0.11109,"tier":{"type":"context","size":32000}},{"input":0.17774,"output":0.71098,"cache_read":0.017774,"cache_write":0.22218,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen3-vl-235b-a22b":{"id":"qwen3-vl-235b-a22b","name":"Qwen3-VL 235B-A22B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.286705,"output":1.14682,"reasoning":2.867051}},"qwen-math-plus":{"id":"qwen-math-plus","name":"Qwen Math Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-08-16","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4096,"output":3072},"status":"deprecated","cost":{"input":0.574,"output":1.721}},"qwen2-5-omni-7b":{"id":"qwen2-5-omni-7b","name":"Qwen2.5-Omni 7B","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-12","last_updated":"2024-12","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":true,"limit":{"context":32768,"output":2048},"status":"deprecated","cost":{"input":0.087,"output":0.345,"input_audio":5.448}},"moonshot-kimi-k2-instruct":{"id":"moonshot-kimi-k2-instruct","name":"Moonshot Kimi K2 Instruct","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.574,"output":2.294}},"qwen-plus":{"id":"qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.115,"output":0.287,"reasoning":1.147,"cache_read":0.012,"cache_write":0.144,"tiers":[{"input":0.345,"output":2.868,"reasoning":3.441,"tier":{"type":"context","size":128000}},{"input":0.689,"output":6.881,"reasoning":9.175,"tier":{"type":"context","size":256000}}]}},"qwen-flash":{"id":"qwen-flash","name":"Qwen Flash","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.022,"output":0.216,"tiers":[{"input":0.087,"output":0.861,"tier":{"type":"context","size":128000}},{"input":0.173,"output":1.721,"tier":{"type":"context","size":256000}}]}},"qwen3-coder-plus":{"id":"qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Hosted Qwen coder for software agents, repo edits, and long-context code","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2026-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.574,"output":2.296,"tiers":[{"input":0.861,"output":3.444,"tier":{"type":"context","size":32000}},{"input":1.435,"output":5.74,"tier":{"type":"context","size":128000}},{"input":2.87,"output":28.7,"tier":{"type":"context","size":256000}}]}},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.65,"output":3.301},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-r1-distill-qwen-32b":{"id":"deepseek-r1-distill-qwen-32b","name":"DeepSeek R1 Distill Qwen 32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":16384},"status":"deprecated","cost":{"input":0.287,"output":0.861}},"qwen3-8b":{"id":"qwen3-8b","name":"Qwen3 8B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":38912}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.072,"output":0.287,"reasoning":0.717}},"qwen3.5-plus":{"id":"qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-09-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.115,"output":0.688,"reasoning":0.688,"tiers":[{"input":0.287,"output":1.72,"reasoning":1.72,"tier":{"type":"context","size":128000}},{"input":0.573,"output":3.44,"reasoning":3.44,"tier":{"type":"context","size":256000}}]}},"kimi-k2.6":{"id":"kimi-k2.6","name":"Moonshot Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.929,"output":3.858}},"qwen3.8-omni-flash":{"id":"qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.113,"output":0.382,"cache_read":0.014},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"qwen-vl-plus":{"id":"qwen-vl-plus","name":"Qwen-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-08-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.115,"output":0.287}},"qwen-max":{"id":"qwen-max","name":"Qwen Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.345,"output":1.377}},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":2.827,"output":14.133,"cache_read":0.283},"canonical_model_id":"moonshotai/kimi-k3"},"deepseek-r1-0528":{"id":"deepseek-r1-0528","name":"DeepSeek R1 0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.574,"output":2.294}},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.276,"output":1.651,"cache_read":0.0276,"cache_write":0.345,"tiers":[{"input":1.101,"output":6.602,"cache_read":0.1101,"cache_write":1.37625,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.101,"output":6.602,"cache_read":0.1101,"cache_write":1.37625}}},"deepseek-r1-distill-qwen-1-5b":{"id":"deepseek-r1-distill-qwen-1-5b","name":"DeepSeek R1 Distill Qwen 1.5B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":16384},"cost":{"input":0,"output":0}},"deepseek-v3-2-exp":{"id":"deepseek-v3-2-exp","name":"DeepSeek V3.2 Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":65536},"status":"deprecated","cost":{"input":0.287,"output":0.431}},"qwq-plus":{"id":"qwq-plus","name":"QwQ Plus","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-03-05","last_updated":"2025-03-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.23,"output":0.574}},"qwen3-asr-flash":{"id":"qwen3-asr-flash","type":"transcription","name":"Qwen3-ASR Flash","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-04","release_date":"2025-09-08","last_updated":"2025-09-08","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":53248,"output":4096},"cost":{"input":0.032,"output":0.032}},"qwen3-omni-flash":{"id":"qwen3-omni-flash","name":"Qwen3-Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":false,"limit":{"context":65536,"output":16384},"cost":{"input":0.258,"output":0.989,"input_audio":2.265,"output_audio":8.974}},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":131072}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.165,"output":0.99,"cache_read":0.0165,"cache_write":0.20625,"tiers":[{"input":0.66,"output":3.961,"cache_read":0.066,"cache_write":0.825,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.66,"output":3.961,"cache_read":0.066,"cache_write":0.825}},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen3.5-flash":{"id":"qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-23","last_updated":"2026-09-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.029,"output":0.287,"reasoning":0.287,"tiers":[{"input":0.115,"output":1.147,"reasoning":1.147,"tier":{"type":"context","size":128000}},{"input":0.172,"output":1.72,"reasoning":1.72,"tier":{"type":"context","size":256000}}]}},"qwen2-5-coder-32b-instruct":{"id":"qwen2-5-coder-32b-instruct","name":"Qwen2.5-Coder 32B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-11","last_updated":"2024-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.287,"output":0.861}},"qwen3-235b-a22b":{"id":"qwen3-235b-a22b","name":"Qwen3 235B-A22B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":38912}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.287,"output":1.147,"reasoning":2.868}},"qwen3-coder-flash":{"id":"qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.144,"output":0.574,"tiers":[{"input":0.216,"output":0.861,"tier":{"type":"context","size":32000}},{"input":0.359,"output":1.434,"tier":{"type":"context","size":128000}},{"input":0.717,"output":3.584,"tier":{"type":"context","size":256000}}]}},"deepseek-v3":{"id":"deepseek-v3","name":"DeepSeek V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-12-01","last_updated":"2024-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":8192},"status":"deprecated","cost":{"input":0.287,"output":1.147}},"qwen2-5-vl-72b-instruct":{"id":"qwen2-5-vl-72b-instruct","name":"Qwen2.5-VL 72B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":2.294,"output":6.881}},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.11875,"output":0.40073,"cache_read":0.01187,"cache_write":0.14844},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen-omni-turbo-realtime":{"id":"qwen-omni-turbo-realtime","type":"realtime","name":"Qwen-Omni Turbo Realtime","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-05-08","last_updated":"2025-05-08","modalities":{"input":["text","image","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":32768,"output":2048},"status":"deprecated","cost":{"input":0.23,"output":0.918,"input_audio":3.584,"output_audio":7.168}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2}},"qwq-32b":{"id":"qwq-32b","name":"QwQ 32B","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-12","last_updated":"2024-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.287,"output":0.861}},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.276,"output":1.101,"cache_read":0.0276,"cache_write":0.345,"tiers":[{"input":0.826,"output":3.301,"cache_read":0.0826,"cache_write":1.0325,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.826,"output":3.301,"cache_read":0.0826,"cache_write":1.0325}},"canonical_model_id":"alibaba/qwen3.7-plus"},"deepseek-r1":{"id":"deepseek-r1","name":"DeepSeek R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.574,"output":2.294}},"qwen-vl-max":{"id":"qwen-vl-max","name":"Qwen-VL Max","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-08","last_updated":"2025-08-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.23,"output":0.574}},"deepseek-r1-distill-qwen-7b":{"id":"deepseek-r1-distill-qwen-7b","name":"DeepSeek R1 Distill Qwen 7B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":16384},"status":"deprecated","cost":{"input":0.072,"output":0.144}},"qwen-turbo":{"id":"qwen-turbo","name":"Qwen Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":38912}],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-11-01","last_updated":"2025-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":16384},"status":"deprecated","cost":{"input":0.044,"output":0.087,"reasoning":0.431}},"qwen3-next-80b-a3b-instruct":{"id":"qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.144,"output":0.574}},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.1,"output":3.851,"cache_read":0.275,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"qvq-max":{"id":"qvq-max","name":"QVQ Max","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qvq","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-03-25","last_updated":"2025-03-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":1.147,"output":4.588}},"qwen2-5-coder-7b-instruct":{"id":"qwen2-5-coder-7b-instruct","name":"Qwen2.5-Coder 7B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-11","last_updated":"2024-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.144,"output":0.287}},"qwen3-next-80b-a3b-thinking":{"id":"qwen3-next-80b-a3b-thinking","name":"Qwen3-Next 80B-A3B (Thinking)","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.144,"output":1.434}},"qwen3-vl-30b-a3b":{"id":"qwen3-vl-30b-a3b","name":"Qwen3-VL 30B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.108,"output":0.431,"reasoning":1.076}},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":131072}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-14","last_updated":"2026-04-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":128000},"cost":{"input":0.825,"output":3.301,"cache_read":0.17,"tiers":[{"input":1.1,"output":3.851,"tier":{"type":"context","size":32000}}]}},"qwen-plus-character":{"id":"qwen-plus-character","name":"Qwen Plus Character","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-01","last_updated":"2024-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":4096},"cost":{"input":0.115,"output":0.287}},"qwen-math-turbo":{"id":"qwen-math-turbo","name":"Qwen Math Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09-19","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4096,"output":3072},"status":"deprecated","cost":{"input":0.287,"output":0.861}},"qwen-omni-turbo":{"id":"qwen-omni-turbo","name":"Qwen-Omni Turbo","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-01-19","last_updated":"2025-03-26","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":false,"limit":{"context":32768,"output":2048},"status":"deprecated","cost":{"input":0.058,"output":0.23,"input_audio":3.584,"output_audio":7.168}},"deepseek-r1-distill-llama-8b":{"id":"deepseek-r1-distill-llama-8b","name":"DeepSeek R1 Distill Llama 8B","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":16384},"status":"deprecated","cost":{"input":0,"output":0}},"qwen3-coder-480b-a35b-instruct":{"id":"qwen3-coder-480b-a35b-instruct","name":"Qwen3-Coder 480B-A35B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.861,"output":3.441,"tiers":[{"input":1.291,"output":5.161,"tier":{"type":"context","size":32000}},{"input":2.151,"output":8.602,"tier":{"type":"context","size":128000}}]}},"qwen2-5-math-7b-instruct":{"id":"qwen2-5-math-7b-instruct","name":"Qwen2.5-Math 7B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":3072},"status":"deprecated","cost":{"input":0.144,"output":0.287}},"tongyi-intent-detect-v3":{"id":"tongyi-intent-detect-v3","name":"Tongyi Intent Detect V3","description":"General-purpose chat model for instruction following, writing, and analysis","family":"yi","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-01","last_updated":"2024-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1024},"cost":{"input":0.058,"output":0.144}},"qwen2-5-72b-instruct":{"id":"qwen2-5-72b-instruct","name":"Qwen2.5 72B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.574,"output":1.721}},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"qwen3-32b":{"id":"qwen3-32b","name":"Qwen3 32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":38912}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.287,"output":1.147,"reasoning":2.868}},"qwen3-omni-flash-realtime":{"id":"qwen3-omni-flash-realtime","type":"realtime","name":"Qwen3-Omni Flash Realtime","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":65536,"output":16384},"status":"deprecated","cost":{"input":0.23,"output":0.918,"input_audio":3.584,"output_audio":7.168}},"qwen2-5-14b-instruct":{"id":"qwen2-5-14b-instruct","name":"Qwen2.5 14B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.144,"output":0.431}},"qwen2-5-math-72b-instruct":{"id":"qwen2-5-math-72b-instruct","name":"Qwen2.5-Math 72B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":3072},"status":"deprecated","cost":{"input":0.574,"output":1.721}},"glm-5":{"id":"glm-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":16384},"cost":{"input":0.573,"output":2.58,"tiers":[{"input":0.86,"output":3.154,"tier":{"type":"context","size":32000}}]}},"deepseek-r1-distill-llama-70b":{"id":"deepseek-r1-distill-llama-70b","name":"DeepSeek R1 Distill Llama 70B","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":16384},"cost":{"input":0.287,"output":0.861}},"qwen3-coder-30b-a3b-instruct":{"id":"qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.216,"output":0.861,"tiers":[{"input":0.323,"output":1.291,"tier":{"type":"context","size":32000}},{"input":0.538,"output":2.151,"tier":{"type":"context","size":128000}}]}},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.172,"output":1.032,"reasoning":1.032,"tiers":[{"input":0.43,"output":2.58,"reasoning":2.58,"tier":{"type":"context","size":128000}}]}},"qwen-long":{"id":"qwen-long","name":"Qwen Long","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-01-25","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":10000000,"output":8192},"cost":{"input":0.072,"output":0.287}},"qwen2-5-vl-7b-instruct":{"id":"qwen2-5-vl-7b-instruct","name":"Qwen2.5-VL 7B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.287,"output":0.717}},"qwen-vl-ocr":{"id":"qwen-vl-ocr","name":"Qwen-VL OCR","description":"OCR model for extracting structured text from documents and screenshots","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-10-28","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":34096,"output":4096},"cost":{"input":0.043,"output":0.072}},"qwen3.6-max-preview":{"id":"qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":131072}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-20","last_updated":"2026-04-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":245800,"output":65536},"status":"deprecated","cost":{"input":1.32,"output":7.9,"cache_read":0.132}},"qwen-mt-turbo":{"id":"qwen-mt-turbo","name":"Qwen-MT Turbo","description":"Translation model for multilingual conversion, localization, and cross-language workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-01","last_updated":"2025-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":8192},"status":"deprecated","cost":{"input":0.101,"output":0.28}},"qwen-doc-turbo":{"id":"qwen-doc-turbo","name":"Qwen Doc Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01","last_updated":"2024-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.087,"output":0.144}},"qwen2-5-32b-instruct":{"id":"qwen2-5-32b-instruct","name":"Qwen2.5 32B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.287,"output":0.861}},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.65,"output":4.951,"cache_read":0.33,"cache_write":2.0625},"canonical_model_id":"alibaba/qwen3.7-max"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.1,"output":3.851,"cache_read":0.275,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.29754,"output":1.19015,"cache_read":0.01488},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen-mt-plus":{"id":"qwen-mt-plus","name":"Qwen-MT Plus","description":"Translation model for multilingual conversion, localization, and cross-language workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-01","last_updated":"2025-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":8192},"cost":{"input":0.259,"output":0.775}},"kimi-k2.5":{"id":"kimi-k2.5","name":"Moonshot Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.574,"output":3.011}},"qwen3-vl-plus":{"id":"qwen3-vl-plus","name":"Qwen3-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.143353,"output":1.433525,"tiers":[{"input":0.215,"output":2.15,"tier":{"type":"context","size":32000}},{"input":0.43,"output":4.301,"tier":{"type":"context","size":128000}}]}},"qwen2-5-7b-instruct":{"id":"qwen2-5-7b-instruct","name":"Qwen2.5 7B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.072,"output":0.144}},"qwen-deep-research":{"id":"qwen-deep-research","name":"Qwen Deep Research","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01","last_updated":"2024-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":7.742,"output":23.367}},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.77744,"output":5.33231,"cache_read":0.22218,"cache_write":2.22179},"canonical_model_id":"alibaba/qwen3.8-max"},"deepseek-v3-1":{"id":"deepseek-v3-1","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":65536},"status":"deprecated","cost":{"input":0.574,"output":1.721}},"qwen3-max":{"id":"qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-23","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.359,"output":1.434,"reasoning":1.434,"tiers":[{"input":0.574,"output":2.294,"reasoning":2.294,"tier":{"type":"context","size":32000}},{"input":1.004,"output":4.014,"reasoning":4.014,"tier":{"type":"context","size":128000}}]}},"deepseek-r1-distill-qwen-14b":{"id":"deepseek-r1-distill-qwen-14b","name":"DeepSeek R1 Distill Qwen 14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":16384},"status":"deprecated","cost":{"input":0.144,"output":0.431}},"qwen3-14b":{"id":"qwen3-14b","name":"Qwen3 14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":38912}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.144,"output":0.574,"reasoning":1.434}},"kimi-k2-thinking":{"id":"kimi-k2-thinking","name":"Moonshot Kimi K2 Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"status":"deprecated","cost":{"input":0.574,"output":2.294}},"siliconflow/deepseek-r1-0528":{"id":"siliconflow/deepseek-r1-0528","name":"siliconflow/deepseek-r1-0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":32768},"cost":{"input":0.5,"output":2.18}},"siliconflow/deepseek-v3-0324":{"id":"siliconflow/deepseek-v3-0324","name":"siliconflow/deepseek-v3-0324","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-12-26","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":163840},"cost":{"input":0.25,"output":1}},"siliconflow/deepseek-v3.1-terminus":{"id":"siliconflow/deepseek-v3.1-terminus","name":"siliconflow/deepseek-v3.1-terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-29","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":65536},"cost":{"input":0.27,"output":1}},"siliconflow/deepseek-v3.2":{"id":"siliconflow/deepseek-v3.2","name":"siliconflow/deepseek-v3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-03","last_updated":"2025-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":65536},"cost":{"input":0.27,"output":0.42}},"kimi/kimi-k2.5":{"id":"kimi/kimi-k2.5","name":"kimi/kimi-k2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3,"cache_read":0.1}},"MiniMax/MiniMax-M2.7":{"id":"MiniMax/MiniMax-M2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375}}}},"friendli":{"id":"friendli","env":["FRIENDLI_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"https://api.friendli.ai/serverless/v1","name":"Friendli","doc":"https://friendli.ai/docs/guides/serverless_endpoints/introduction","models":{"deepseek-ai/DeepSeek-V3.2":{"id":"deepseek-ai/DeepSeek-V3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":64000},"cost":{"input":0.5,"output":1.5,"cache_read":0.25},"canonical_model_id":"deepseek/deepseek-v3.2"},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.5"},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.26,"output":3.96,"cache_read":0.234},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"}}},"hpc-ai":{"id":"hpc-ai","env":["HPC_AI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.hpc-ai.com/inference/v1","name":"HPC-AI","doc":"https://www.hpc-ai.com/doc/docs/quickstart/","models":{"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-7"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1002000,"output":128000},"cost":{"input":1.74,"output":3.48,"cache_read":0.145},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196000,"output":195000},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"zai-org/glm-5.1":{"id":"zai-org/glm-5.1","name":"GLM 5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202000,"output":202000},"cost":{"input":0.615,"output":2.46,"cache_read":0.133}},"zai-org/glm-5.2":{"id":"zai-org/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"}}},"llmgateway":{"id":"llmgateway","env":["LLMGATEWAY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.llmgateway.io/v1","name":"DevPass (LLM Gateway)","doc":"https://llmgateway.io/docs","models":{"custom":{"id":"custom","name":"Custom Model","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0,"output":0}},"glm-4.6v":{"id":"glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.05},"canonical_model_id":"zhipuai/glm-4.6v"},"qwen3-235b-a22b-thinking-2507":{"id":"qwen3-235b-a22b-thinking-2507","name":"Qwen3 235B A22B Thinking (2507)","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-08","last_updated":"2025-07-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":8192},"cost":{"input":0.3,"output":3}},"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"qwen3.7-flash":{"id":"qwen3.7-flash","name":"Qwen3.7 Flash","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-27","last_updated":"2026-07-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":1000000},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.0375}},"gpt-4o-mini":{"id":"gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"seed-1-6-250615":{"id":"seed-1-6-250615","name":"Seed 1.6 (250615)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-06-25","last_updated":"2025-06-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.088,"output":0.25,"cache_read":0.025},"canonical_model_id":"zhipuai/glm-5.3-flash"},"muse-spark-1.3-contributor":{"id":"muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"seed-2-0-code-preview-260328":{"id":"seed-2-0-code-preview-260328","name":"Seed 2.0 Code Preview (260328)","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-28","last_updated":"2026-03-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":3,"cache_read":0.1}},"gpt-4":{"id":"gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":30,"output":60},"canonical_model_id":"openai/gpt-4"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"qwen3-vl-30b-a3b-instruct":{"id":"qwen3-vl-30b-a3b-instruct","name":"Qwen3 VL 30B A3B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-02","last_updated":"2025-10-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.15,"output":0.6}},"qwen3.5-9b":{"id":"qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.1,"output":0.15},"canonical_model_id":"alibaba/qwen3.5-9b"},"inkling":{"id":"inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":1048576},"cost":{"input":0.95,"output":4.05,"cache_read":0.16},"canonical_model_id":"thinkingmachines/inkling"},"minimax-m2.1-lightning":{"id":"minimax-m2.1-lightning","name":"MiniMax M2.1 Lightning","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.12,"output":0.48}},"gemini-pro-latest":{"id":"gemini-pro-latest","name":"Gemini Pro Latest","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-27","last_updated":"2026-02-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2}},"kimi-k3-fast":{"id":"kimi-k3-fast","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1040384,"output":131072},"cost":{"input":4.5,"output":22.5,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"fugu-max":{"id":"fugu-max","name":"Fugu Max","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":2,"output":6,"cache_read":0.25}},"auto":{"id":"auto","name":"Auto Route","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0,"output":0}},"hy4-preview":{"id":"hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.08,"output":0.32,"cache_read":0.017,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"grok-4-5":{"id":"grok-4-5","name":"Grok 4.5","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3}},"gpt-live-transcribe":{"id":"gpt-live-transcribe","type":"transcription","name":"GPT-Live-Transcribe","description":"Low-latency streaming speech-to-text model with keyword and language hints","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-07-28","last_updated":"2026-07-28","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-live-transcribe"},"qwen3-vl-235b-a22b-instruct":{"id":"qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.2,"output":0.88,"cache_read":0.11}},"claude-opus-4-1-20250805":{"id":"claude-opus-4-1-20250805","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":31999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"qwen-plus":{"id":"qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.4,"output":1.2,"reasoning":4,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen-plus"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"llama-3-70b-instruct":{"id":"llama-3-70b-instruct","name":"Llama 3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2024-04-18","last_updated":"2024-04-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8000},"cost":{"input":0.51,"output":0.74}},"qwen-flash":{"id":"qwen-flash","name":"Qwen Flash","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.05,"output":0.4,"cache_read":0.01,"cache_write":0.0625},"canonical_model_id":"alibaba/qwen-flash"},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"glm-5.3-fast":{"id":"glm-5.3-fast","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2.8,"output":8.8,"cache_read":0.56},"canonical_model_id":"zhipuai/glm-5.3"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3-flash-preview"},"qwen3.5-122b-a10b":{"id":"qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":3.2},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"gpt-5.4-pro":{"id":"gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.4-pro"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":384000},"cost":{"input":0.435,"output":0.87,"cache_read":0.003625},"canonical_model_id":"deepseek/deepseek-v4-pro"},"minimax-m2":{"id":"minimax-m2","name":"MiniMax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.2,"output":1,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2"},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"gpt-5-pro":{"id":"gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036,"tiers":[{"input":2,"output":6,"cache_read":0.4,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.4}},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"seed-1-6-flash-250715":{"id":"seed-1-6-flash-250715","name":"Seed 1.6 Flash (250715)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.07,"output":0.3,"cache_read":0.015}},"granite-4.2-8b":{"id":"granite-4.2-8b","name":"Granite 4.2 8B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"granite","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":0.06,"output":0.25,"cache_read":0.015}},"llama-3.2-3b-instruct":{"id":"llama-3.2-3b-instruct","name":"Llama 3.2 3B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2024-09-18","last_updated":"2024-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32000},"cost":{"input":0.03,"output":0.05}},"grok-4-7":{"id":"grok-4-7","name":"Grok 4.7","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5}},"grok-4-20-beta-0309-reasoning":{"id":"grok-4-20-beta-0309-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.248,"output":1.485},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.135,"output":0.4},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3.05,"cache_read":0.13},"canonical_model_id":"moonshotai/kimi-k2.6"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"openai/gpt-6-luna"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"qwen-max":{"id":"qwen-max","name":"Qwen Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":1.6,"output":6.4},"canonical_model_id":"alibaba/qwen-max"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"glm-5.2-fast":{"id":"glm-5.2-fast","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2.156,"output":6.776,"cache_read":0.429},"canonical_model_id":"zhipuai/glm-5.2"},"claude-opus-4-5-20251101":{"id":"claude-opus-4-5-20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":31999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"fugu-ultra-v2.0":{"id":"fugu-ultra-v2.0","name":"Fugu Ultra v2.0","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":5,"output":30,"cache_read":0.5}},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625,"tiers":[{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5}},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen3-235b-a22b-instruct-2507":{"id":"qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B A22B Instruct (2507)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-08","last_updated":"2025-07-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.09,"output":0.58}},"glm-4.5v":{"id":"glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.6,"output":1.8,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5v"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.05,"cache_write":0.3125},"canonical_model_id":"alibaba/qwen3.6-flash"},"glm-4.6v-flashx":{"id":"glm-4.6v-flashx","name":"GLM-4.6V FlashX","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000},"cost":{"input":0.04,"output":0.4,"cache_read":0.004}},"ernie-4.5-vl-424b-a47b":{"id":"ernie-4.5-vl-424b-a47b","name":"ERNIE 4.5 VL 424B A47B","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ernie","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-06-30","last_updated":"2025-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":123000,"output":123000},"cost":{"input":0.42,"output":1.25}},"qwen3-coder-flash":{"id":"qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":997952,"output":65536},"cost":{"input":0.3,"output":1.5,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"alibaba/qwen3-coder-flash"},"o3-mini":{"id":"o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"kimi-k2.7-code-highspeed":{"id":"kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"gpt-3.5-turbo":{"id":"gpt-3.5-turbo","name":"GPT-3.5-turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":0.5,"output":1.5,"cache_read":0},"canonical_model_id":"openai/gpt-3.5-turbo"},"muse-glimmer-30b":{"id":"muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.168,"output":0.336,"cache_read":0.0034,"tiers":[{"input":0.8,"output":4,"cache_read":0.16,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.8,"output":4,"cache_read":0.16}},"canonical_model_id":"xiaomi/mimo-v2.5"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"kimi-k2":{"id":"kimi-k2","name":"Kimi K2","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.57,"output":2.3,"cache_read":0.5}},"gpt-4o-transcribe":{"id":"gpt-4o-transcribe","name":"GPT-4o Transcribe","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-03-20","last_updated":"2025-03-20","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":16000},"cost":{"input":2.5,"output":10}},"ministral-8b-2512":{"id":"ministral-8b-2512","name":"Ministral 8B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"mistral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.15,"output":0.15}},"glm-4.5-airx":{"id":"glm-4.5-airx","name":"GLM-4.5 AirX","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.1,"output":4.5,"cache_read":0.22}},"qwen3-235b-a22b-fp8":{"id":"qwen3-235b-a22b-fp8","name":"Qwen3 235B A22B FP8","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":8192},"cost":{"input":0.2,"output":0.8}},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"claude-sonnet-4-5-20250929":{"id":"claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":63999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"qwen3-vl-235b-a22b-thinking":{"id":"qwen3-vl-235b-a22b-thinking","name":"Qwen3 VL 235B A22B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.98,"output":3.95}},"step-3.7-flash":{"id":"step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"qwen3.8-2.4t-a95b":{"id":"qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1010000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"hy3":{"id":"hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":128000},"cost":{"input":0.132,"output":0.528,"cache_read":0.033},"canonical_model_id":"tencent/hy3"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.1-flash-lite"},"ling-3.0-flash-vl":{"id":"ling-3.0-flash-vl","name":"Ling 3.0 Flash VL","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":0.06,"output":0.18,"cache_read":0.012}},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":64000},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen3.7-plus"},"llama-3.2-11b-instruct":{"id":"llama-3.2-11b-instruct","name":"Llama 3.2 11B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.07,"output":0.33}},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.7-flash"},"ling-3.0-flash":{"id":"ling-3.0-flash","name":"InclusionAI Ling 3.0 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-02","last_updated":"2026-08-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.06,"output":0.18,"cache_read":0.012}},"glm-4.5-air":{"id":"glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":98304},"cost":{"input":0.13,"output":0.85,"cache_read":0.025,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-air"},"qwen3-next-80b-a3b-instruct":{"id":"qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"Qwen3.8-27B":{"id":"Qwen3.8-27B","name":"Qwen3.8 27B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":32768},"cost":{"input":0.08,"output":0.35,"cache_read":0.05}},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.5,"output":2,"cache_read":0.15},"canonical_model_id":"zhipuai/glm-5.3"},"hy-mt2-plus":{"id":"hy-mt2-plus","name":"Hy-MT2 Plus","description":"Tool-capable chat model for instruction following and agentic application workflows","family":"Hy","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0.074,"output":0.295}},"qwen3.5-27b":{"id":"qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":2.4},"canonical_model_id":"alibaba/qwen3.5-27b"},"devstral-2512":{"id":"devstral-2512","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/devstral-2512"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333}},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.38,"output":1.98,"cache_read":0.19,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7"},"seed-2-0-lite-260428":{"id":"seed-2-0-lite-260428","name":"Seed 2.0 Lite (260428)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"qwen3-next-80b-a3b-thinking":{"id":"qwen3-next-80b-a3b-thinking","name":"Qwen3-Next 80B-A3B (Thinking)","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.931,"output":2.93,"cache_read":0.173,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"ministral-3b-2512":{"id":"ministral-3b-2512","name":"Ministral 3B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"mistral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.1,"output":0.1}},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.5-flash"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"grok-4-6":{"id":"grok-4-6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"minimax-m2.1":{"id":"minimax-m2.1","name":"MiniMax-M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.27,"output":1.1},"canonical_model_id":"minimax/MiniMax-M2.1"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32766},"cost":{"input":0.04,"output":0.19,"cache_read":0.01}},"minimax-text-01":{"id":"minimax-text-01","name":"MiniMax Text 01","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-01-15","last_updated":"2025-01-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.2,"output":1.1}},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"o3":{"id":"o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.12,"output":0.25,"cache_read":0.0025},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"glm-4-32b-0414-128k":{"id":"glm-4-32b-0414-128k","name":"GLM-4 32B (0414-128k)","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.1,"output":0.1}},"o1":{"id":"o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"gpt-5.1-codex-mini":{"id":"gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"minimax-m2.5-highspeed":{"id":"minimax-m2.5-highspeed","name":"MiniMax-M2.5-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5-highspeed"},"mistral-medium-3-5":{"id":"mistral-medium-3-5","name":"Mistral Medium 3.5","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5}},"qwen3.5-35b-a3b":{"id":"qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Gemini multimodal model for text, image, audio, video, and document tasks","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-30","last_updated":"2026-09-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":0.375}},"qwen3-coder-480b-a35b-instruct":{"id":"qwen3-coder-480b-a35b-instruct","name":"Qwen3-Coder 480B-A35B Instruct","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.38,"output":1.55},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"seed-1-6-250915":{"id":"seed-1-6-250915","name":"Seed 1.6 (250915)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":63999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"sonar":{"id":"sonar","name":"Sonar","description":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","family":"sonar","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":130000,"output":4096},"cost":{"input":0.25,"output":2.5,"cache_write":0.25},"canonical_model_id":"perplexity/sonar"},"seed-2-0-mini-260428":{"id":"seed-2-0-mini-260428","name":"Seed 2.0 Mini (260428)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.4,"cache_read":0.02}},"gpt-5.2-pro":{"id":"gpt-5.2-pro","name":"GPT-5.2 Pro","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"gemma-4-26b-a4b-it":{"id":"gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.07,"output":0.34},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-03","last_updated":"2026-09-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":1050000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5}},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":384000},"cost":{"input":0.076,"output":0.153,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash"},"seed-2-0-pro-260328":{"id":"seed-2-0-pro-260328","name":"Seed 2.0 Pro (260328)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-28","last_updated":"2026-03-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":3,"cache_read":0.1}},"mistral-large-2512":{"id":"mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/mistral-large-2512"},"muse-spark-1.2-contributor":{"id":"muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"qwen3-32b":{"id":"qwen3-32b","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.36,"output":0.87,"reasoning":8.4},"canonical_model_id":"alibaba/qwen3-32b"},"gpt-4-turbo":{"id":"gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"muse-spark-1.3":{"id":"muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":63999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"mistral-small-2603":{"id":"mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"mistral/mistral-small-2603"},"glm-4.7-flash":{"id":"glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.06,"output":0.4,"cache_read":0.01,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"grok-build-0-1":{"id":"grok-build-0-1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}},"canonical_model_id":"xai/grok-build-0.1"},"llama-4-scout-17b-instruct":{"id":"llama-4-scout-17b-instruct","name":"Llama 4 Scout 17B Instruct","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":2048},"cost":{"input":0.18,"output":0.59}},"minimax-m3.1-flash":{"id":"minimax-m3.1-flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":512000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"grok-4-20-reasoning":{"id":"grok-4-20-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"glm-5":{"id":"glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131072},"cost":{"input":0.72,"output":2.3,"cache_read":0.144,"cache_write":0},"canonical_model_id":"zhipuai/glm-5"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.737,"output":3.08,"cache_read":0.143},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"qwen3-coder-30b-a3b-instruct":{"id":"qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":65536},"cost":{"input":0.07,"output":0.27},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5},"canonical_model_id":"openai/gpt-5.6-sol"},"muse-spark-1.1":{"id":"muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"minimax-m2.7-highspeed":{"id":"minimax-m2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"openai/gpt-6.1-sol"},"qwen-plus-latest":{"id":"qwen-plus-latest","name":"Qwen Plus Latest","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-01-25","last_updated":"2025-01-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":8192},"cost":{"input":0.4,"output":1.2,"cache_read":0.08,"cache_write":0.5}},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":16384},"cost":{"input":0.26,"output":0.38,"cache_read":0.13}},"ministral-14b-2512":{"id":"ministral-14b-2512","name":"Ministral 14B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"mistral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.2,"output":0.2}},"glm-5v-turbo":{"id":"glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"grok-4-20-non-reasoning":{"id":"grok-4-20-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"llama-4-maverick-17b-instruct":{"id":"llama-4-maverick-17b-instruct","name":"Llama 4 Maverick 17B Instruct","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":2048},"cost":{"input":0.27,"output":0.85}},"glm-5-turbo":{"id":"glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"grok-4":{"id":"grok-4","name":"Grok 4","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":3,"output":15,"cache_read":0.75}},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.28,"output":1.1,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M3"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.5-flash-lite"},"mistral-large-4":{"id":"mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":262144},"cost":{"input":1.36,"output":4.18,"cache_read":0.14},"canonical_model_id":"mistral/mistral-large-4"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32766},"cost":{"input":0.032,"output":0.14,"cache_read":0.032}},"mistral-small-2506":{"id":"mistral-small-2506","name":"Mistral Small 3.2","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/mistral-small-2506"},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"codestral-2508":{"id":"codestral-2508","name":"Codestral","description":"Mistral coding model for code completion, generation, and developer workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.3,"output":0.9}},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"qwen3-coder-next":{"id":"qwen3-coder-next","name":"Qwen3 Coder Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.108,"output":0.675,"cache_read":0.06}},"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":1050000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003}},"mistral-large-latest":{"id":"mistral-large-latest","name":"Mistral Large (latest)","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":262144},"cost":{"input":4,"output":12},"canonical_model_id":"mistral/mistral-large-latest"},"gemma-4-31b-it":{"id":"gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.1,"output":0.25,"cache_read":0.01},"canonical_model_id":"google/gemma-4-31b-it"},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.25,"output":3.75,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.7-max"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-6-sol"},"nemotron-3-ultra-550b":{"id":"nemotron-3-ultra-550b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"cost":{"input":0.5,"output":2.2,"cache_read":0.1},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"grok-4-1-fast-non-reasoning":{"id":"grok-4-1-fast-non-reasoning","name":"Grok 4.1 Fast Non-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"anthropic/claude-haiku-5-5"},"reka-edge-2603":{"id":"reka-edge-2603","name":"Reka Edge 2603","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"reka","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-09","last_updated":"2026-10-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":16384},"cost":{"input":0.1,"output":0.1}},"gpt-4o-mini-transcribe":{"id":"gpt-4o-mini-transcribe","name":"GPT-4o Mini Transcribe","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-03-20","last_updated":"2025-03-20","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":16000},"cost":{"input":1.25,"output":5}},"llama-3.1-70b-instruct":{"id":"llama-3.1-70b-instruct","name":"Llama 3.1 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":2048},"status":"beta","cost":{"input":0.72,"output":0.72}},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"grok-4-3":{"id":"grok-4-3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"muse-wonder-1.0":{"id":"muse-wonder-1.0","name":"Muse Wonder","description":"Muse Wonder is a frontier multimodal reasoning model from Meta for personal and professional agents, long-horizon coding, and computer use. It combines Active Memory, which carries relevant context across conversations, with Active Focus, which keeps it working through blockers toward a goal.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"glm-4.6":{"id":"glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.55,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6"},"grok-4-20-beta-0309-non-reasoning":{"id":"grok-4-20-beta-0309-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.605,"output":2.123,"cache_read":0.154,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":384000},"cost":{"input":0.135,"output":0.54,"cache_read":0.0027},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"glm-4.5":{"id":"glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5"},"glm-4.7-flashx":{"id":"glm-4.7-flashx","name":"GLM-4.7-FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.07,"output":0.4,"cache_read":0.01,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.6-flash"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.405,"output":1.98,"cache_read":0.225},"canonical_model_id":"moonshotai/kimi-k2.5"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-5.5-pro":{"id":"gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"fugu-ultra":{"id":"fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-22","last_updated":"2026-06-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":5,"output":30,"cache_read":0.5}},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"qwen3-vl-plus":{"id":"qwen3-vl-plus","name":"Qwen3-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":1.6,"reasoning":4.8,"cache_read":0.04,"cache_write":0.25},"canonical_model_id":"alibaba/qwen3-vl-plus"},"seed-1-8-251228":{"id":"seed-1-8-251228","name":"Seed 1.8 (251228)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"grok-4-1-fast-reasoning":{"id":"grok-4-1-fast-reasoning","name":"Grok 4.1 Fast Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max Preview","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-07-19","last_updated":"2026-07-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.54,"output":4.631,"cache_read":0.198,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"minimax-m2.5":{"id":"minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":228700,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5"},"qwen3-max":{"id":"qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.845,"output":3.38,"cache_read":0.6,"cache_write":3.75},"canonical_model_id":"alibaba/qwen3-max"},"glm-4.5-x":{"id":"glm-4.5-x","name":"GLM-4.5 X","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"beta","cost":{"input":2.2,"output":8.9,"cache_read":0.45}},"qwen35-397b-a17b":{"id":"qwen35-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"gpt-transcribe":{"id":"gpt-transcribe","type":"transcription","name":"GPT-Transcribe","description":"High-accuracy speech-to-text model for audio files and committed Realtime turns","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-28","last_updated":"2026-07-28","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-transcribe"},"ling-3.1-flash":{"id":"ling-3.1-flash","name":"Ling 3.1 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"inclusionai/ling-3.1-flash"},"kimi-k2-thinking":{"id":"kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":2.5,"cache_read":0.06},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"muse-spark-1.2":{"id":"muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"}}},"poolside":{"id":"poolside","env":["POOLSIDE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.poolside.ai/v1","name":"Poolside","doc":"https://platform.poolside.ai","models":{"poolside/laguna-m.1":{"id":"poolside/laguna-m.1","name":"Laguna M.1","description":"Poolside's open-weight model for agentic coding and long-horizon work","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"poolside/laguna-m.1"},"poolside/laguna-xs-2.1":{"id":"poolside/laguna-xs-2.1","name":"Laguna XS 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-07-02","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"poolside/laguna-xs-2.1"},"poolside/laguna-s-2.1":{"id":"poolside/laguna-s-2.1","name":"Laguna S 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"poolside/laguna-s-2.1"}}},"opper":{"id":"opper","env":["OPPER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.opper.ai/v3/compat","name":"Opper","doc":"https://opper.ai/models","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.2,"output":0.5,"cache_read":0.07},"canonical_model_id":"zhipuai/glm-5.3-flash"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"gpt-5.3-chat-latest":{"id":"gpt-5.3-chat-latest","name":"GPT-5.3 Chat (latest)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-chat-latest"},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.69732,"output":2.78928},"canonical_model_id":"minimax/MiniMax-M2.7"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3-flash-preview"},"gpt-5.4-pro":{"id":"gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":1.78812,"output":3.57624},"canonical_model_id":"deepseek/deepseek-v4-pro"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.248,"output":1.485},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15},"canonical_model_id":"moonshotai/kimi-k3"},"sonar-pro":{"id":"sonar-pro","name":"Sonar Pro","description":"Deeper Sonar search model with broader retrieval and stronger synthesis","family":"sonar-pro","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8000},"cost":{"input":3,"output":15},"canonical_model_id":"perplexity/sonar-pro"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"qwen3.8-2.4t-a95b":{"id":"qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":2.5,"output":6,"cache_read":0.63},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.75,"output":4.6488,"cache_read":0.44},"canonical_model_id":"zhipuai/glm-5.3"},"devstral-2512":{"id":"devstral-2512","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/devstral-2512"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.825,"output":4.125,"cache_read":0.0825},"canonical_model_id":"google/gemini-3.8-flash"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"google/gemini-3.5-flash"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.11622,"output":0.488124},"canonical_model_id":"openai/gpt-oss-20b"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125,"tiers":[{"input":6.6,"output":24.75,"cache_read":0.66,"cache_write":8.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6.6,"output":24.75,"cache_read":0.66,"cache_write":8.25}},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"sonar":{"id":"sonar","name":"Sonar","description":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","family":"sonar","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":1,"output":1},"canonical_model_id":"perplexity/sonar"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.25,"output":0.66},"canonical_model_id":"deepseek/deepseek-v4-flash"},"grok-build-0.1":{"id":"grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2},"canonical_model_id":"xai/grok-build-0.1"},"mistral-large-2512":{"id":"mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/mistral-large-2512"},"muse-spark-1.3":{"id":"muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"mistral-small-2603":{"id":"mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.5811,"output":2.44062},"canonical_model_id":"mistral/mistral-small-2603"},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.5"},"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.5811,"output":3},"canonical_model_id":"alibaba/qwen3.8-27b"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":45,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-7"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.12,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.24,"tier":{"type":"context","size":524288}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.24}},"canonical_model_id":"minimax/MiniMax-M3"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":1.1622,"output":4.88124},"canonical_model_id":"openai/gpt-oss-120b"},"qwen3-coder-next":{"id":"qwen3-coder-next","name":"Qwen3 Coder Next","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.5811,"output":2.3244},"canonical_model_id":"alibaba/qwen3-coder-next"},"gemma-4-31b-it":{"id":"gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.46488,"output":2.44062},"canonical_model_id":"google/gemma-4-31b-it"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.62708,"output":5.811},"canonical_model_id":"zhipuai/glm-5.2"},"sonar-reasoning-pro":{"id":"sonar-reasoning-pro","name":"Sonar Reasoning Pro","description":"Web-grounded Sonar for multi-step research questions that need cited reasoning","family":"sonar-reasoning","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":2,"output":8},"canonical_model_id":"perplexity/sonar-reasoning-pro"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-5.5-pro":{"id":"gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.5-pro"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max"},"muse-spark-1.2":{"id":"muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"}}},"infer":{"id":"infer","env":["INFER_API_KEY"],"npm":"@ai-sdk/openai","api":"https://infer.flow7.org/v1","name":"Infer by Flow7","doc":"https://infer.flow7.org/opencode","models":{"infer/gpt-6-astra:official":{"id":"infer/gpt-6-astra:official","name":"GPT-6 Astra (Official API)","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":271999,"output":128000},"provider":{"shape":"responses"},"cost":{"input":12.5,"output":62.5,"cache_read":1.25,"cache_write":15.625},"canonical_model_id":"openai/gpt-6-astra"},"infer/gpt-5.6-sol:official":{"id":"infer/gpt-5.6-sol:official","name":"GPT-5.6 Sol (Official API)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":271999,"output":128000},"provider":{"shape":"responses"},"cost":{"input":2.5,"output":12.5,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"openai/gpt-5.6-sol"}}},"alibaba-coding-plan-cn":{"id":"alibaba-coding-plan-cn","env":["ALIBABA_CODING_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://coding.dashscope.aliyuncs.com/v1","name":"Alibaba Coding Plan (China)","doc":"https://help.aliyun.com/zh/model-studio/coding-plan","models":{"qwen3-coder-plus":{"id":"qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3.5-plus":{"id":"qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":24576},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.7-plus"},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3-max-2026-01-23":{"id":"qwen3-max-2026-01-23","name":"Qwen3 Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-23","last_updated":"2026-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"glm-5":{"id":"glm-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":16384},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3-coder-next":{"id":"qwen3-coder-next","name":"Qwen3 Coder Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}}}},"stepfun-ai":{"id":"stepfun-ai","env":["STEPFUN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.stepfun.ai/v1","name":"StepFun (Global)","doc":"https://platform.stepfun.ai/docs/en/overview/concept","models":{"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"input":1024000,"output":65536},"cost":{"input":1,"output":2.7,"cache_read":0.05},"canonical_model_id":"stepfun/step-5-preview"},"step-3.5-flash":{"id":"step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-06-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.1,"output":0.3,"cache_read":0.02}},"step-1-32k":{"id":"step-1-32k","name":"Step 1 (32K)","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-01-01","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":2.05,"output":9.59,"cache_read":0.41}},"step-3.5-flash-2603":{"id":"step-3.5-flash-2603","name":"Step 3.5 Flash 2603","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.1,"output":0.3,"cache_read":0.02}},"step-tts-2":{"id":"step-tts-2","name":"Step TTS 2","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"step","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-01","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"step-3.7-flash":{"id":"step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-06-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.185,"output":1.11,"cache_read":0.037},"canonical_model_id":"stepfun/step-3.7-flash"},"step-2-16k":{"id":"step-2-16k","name":"Step 2 (16K)","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-01-01","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":5.21,"output":16.44,"cache_read":1.04}},"stepaudio-2.5-asr":{"id":"stepaudio-2.5-asr","name":"StepAudio 2.5 ASR","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"step","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-24","last_updated":"2026-07-02","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0}},"stepaudio-2.5-tts":{"id":"stepaudio-2.5-tts","name":"StepAudio 2.5 TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"step","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-16","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}}}},"modal":{"id":"modal","env":["MODAL_PROXY_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.us-west.modal.direct/v1","name":"Modal","doc":"https://modal.com/docs/guide/endpoints","models":{"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"reasoning":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"thinkingmachines/Inkling-NVFP4":{"id":"thinkingmachines/Inkling-NVFP4","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":1.2,"output":5,"cache_read":0.27},"canonical_model_id":"thinkingmachines/inkling"},"Qwen/Qwen3.8-2.4T-A95B":{"id":"Qwen/Qwen3.8-2.4T-A95B","name":"Qwen3.8-Max","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1010000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.45,"output":1.5,"cache_read":0.09},"canonical_model_id":"zhipuai/glm-5.3-flash"}}},"meganova":{"id":"meganova","env":["MEGANOVA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.meganova.ai/v1","name":"Meganova","doc":"https://docs.meganova.ai","models":{"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-25","last_updated":"2025-08-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.27,"output":1}},"deepseek-ai/DeepSeek-V3.2-Exp":{"id":"deepseek-ai/DeepSeek-V3.2-Exp","name":"DeepSeek V3.2 Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-10","last_updated":"2025-10-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.27,"output":0.4}},"deepseek-ai/DeepSeek-R1-0528":{"id":"deepseek-ai/DeepSeek-R1-0528","name":"DeepSeek R1 0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-07","release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":64000},"cost":{"input":0.5,"output":2.15}},"deepseek-ai/DeepSeek-V3-0324":{"id":"deepseek-ai/DeepSeek-V3-0324","name":"DeepSeek V3 0324","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.25,"output":0.88}},"deepseek-ai/DeepSeek-V3.2":{"id":"deepseek-ai/DeepSeek-V3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-03","last_updated":"2025-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.26,"output":0.38}},"meta-llama/Llama-3.3-70B-Instruct":{"id":"meta-llama/Llama-3.3-70B-Instruct","name":"Llama 3.3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.1,"output":0.3}},"moonshotai/Kimi-K2-Thinking":{"id":"moonshotai/Kimi-K2-Thinking","name":"Kimi K2 Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":2.6}},"moonshotai/Kimi-K2.5":{"id":"moonshotai/Kimi-K2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2026-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.45,"output":2.8}},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2}},"MiniMaxAI/MiniMax-M2.1":{"id":"MiniMaxAI/MiniMax-M2.1","name":"MiniMax M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.28,"output":1.2}},"XiaomiMiMo/MiMo-V2-Flash":{"id":"XiaomiMiMo/MiMo-V2-Flash","name":"MiMo V2 Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32000},"cost":{"input":0.1,"output":0.3}},"mistralai/Mistral-Nemo-Instruct-2407":{"id":"mistralai/Mistral-Nemo-Instruct-2407","name":"Mistral Nemo Instruct 2407","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.02,"output":0.04}},"mistralai/Mistral-Small-3.2-24B-Instruct-2506":{"id":"mistralai/Mistral-Small-3.2-24B-Instruct-2506","name":"Mistral Small 3.2 24B Instruct","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0}},"Qwen/Qwen3.5-Plus":{"id":"Qwen/Qwen3.5-Plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02","last_updated":"2026-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.4,"output":2.4,"reasoning":2.4}},"Qwen/Qwen2.5-VL-32B-Instruct":{"id":"Qwen/Qwen2.5-VL-32B-Instruct","name":"Qwen2.5 VL 32B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":0.2,"output":0.6}},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.09,"output":0.6}},"zai-org/GLM-4.7":{"id":"zai-org/GLM-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.2,"output":0.8}},"zai-org/GLM-4.6":{"id":"zai-org/GLM-4.6","name":"GLM-4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.45,"output":1.9}},"zai-org/GLM-5":{"id":"zai-org/GLM-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.8,"output":2.56}}}},"greenpt":{"id":"greenpt","env":["GREENPT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.greenpt.ai/v1","name":"GreenPT","doc":"https://docs.greenpt.ai","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.127754,"output":0.511016,"cache_read":0.0255508},"canonical_model_id":"zhipuai/glm-5.3-flash"},"green-s-pro":{"id":"green-s-pro","name":"Green S Pro","description":"GreenPT advanced speech-to-text model with multilingual transcription support","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-02","last_updated":"2025-02","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":8192},"cost":{"input":0.00437,"output":0}},"glm-5.2-ponytail-lite":{"id":"glm-5.2-ponytail-lite","name":"GLM-5.2 Ponytail Lite","description":"glm-5.2 carrying a built-in ruleset that compresses generated code, preferring platform features over custom code. The gentlest tier: it cuts filler only and keeps the explanation intact. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.254,"output":5.016,"cache_read":0.3135},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5.2-honey-ultra":{"id":"glm-5.2-honey-ultra","name":"GLM-5.2 Honey Ultra","description":"glm-5.2 carrying a built-in ruleset that compresses both generated code and prose. The most aggressive tier, close to answer-only. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.254,"output":5.016,"cache_read":0.3135},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.1596,"output":0.399,"cache_read":0.0456},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"gemma4":{"id":"gemma4","name":"gemma4","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.57,"output":1.71},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"green-l-raw":{"id":"green-l-raw","name":"Green L Raw","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.285,"output":0.912},"canonical_model_id":"mistral/mistral-small-2506"},"green-l":{"id":"green-l","name":"Green L","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.285,"output":0.912},"canonical_model_id":"mistral/mistral-small-2506"},"green-r":{"id":"green-r","name":"Green R","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.399,"output":1.083},"canonical_model_id":"openai/gpt-oss-120b"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.342,"output":2.052},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":100000,"output":16384},"cost":{"input":1.254,"output":1.254},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.7524,"output":4.275,"cache_read":0.2508},"canonical_model_id":"moonshotai/kimi-k2.6"},"gemma-3-27b-it":{"id":"gemma-3-27b-it","name":"Gemma 3 27B","description":"Google Gemma 3 multimodal model for chat, reasoning, and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":40000,"output":8192},"cost":{"input":0.342,"output":0.684}},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3.762,"output":18.81,"cache_read":0.9405},"canonical_model_id":"moonshotai/kimi-k3"},"qwen3-235b-a22b-instruct-2507":{"id":"qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Qwen3 235B MoE instruct model for long-context multilingual chat and reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":1.026,"output":3.078}},"glm-5.2-honey-lite":{"id":"glm-5.2-honey-lite","name":"GLM-5.2 Honey Lite","description":"glm-5.2 carrying a built-in ruleset that compresses both generated code and prose. The gentlest tier: it cuts filler only and keeps the explanation intact. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.254,"output":5.016,"cache_read":0.3135},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5.2-ponytail":{"id":"glm-5.2-ponytail","name":"GLM-5.2 Ponytail","description":"glm-5.2 carrying a built-in ruleset that compresses generated code, preferring platform features over custom code. The middle tier, and the ruleset as its authors wrote it. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.254,"output":5.016,"cache_read":0.3135},"canonical_model_id":"zhipuai/glm-5.2"},"devstral-2-123b-instruct-2512":{"id":"devstral-2-123b-instruct-2512","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":16384},"cost":{"input":0.57,"output":2.736},"canonical_model_id":"mistral/devstral-2512"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.27754,"output":5.11016,"cache_read":0.319385},"canonical_model_id":"zhipuai/glm-5.3"},"glm-5.2-caveman-lite":{"id":"glm-5.2-caveman-lite","name":"GLM-5.2 Caveman Lite","description":"glm-5.2 carrying a built-in ruleset that compresses prose, keeping code and technical detail verbatim. The gentlest tier: it cuts filler only and keeps the explanation intact. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.254,"output":5.016,"cache_read":0.3135},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5.2-ponytail-ultra":{"id":"glm-5.2-ponytail-ultra","name":"GLM-5.2 Ponytail Ultra","description":"glm-5.2 carrying a built-in ruleset that compresses generated code, preferring platform features over custom code. The most aggressive tier, close to answer-only. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.254,"output":5.016,"cache_read":0.3135},"canonical_model_id":"zhipuai/glm-5.2"},"green-r-raw":{"id":"green-r-raw","name":"Green R Raw","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.399,"output":1.083},"canonical_model_id":"openai/gpt-oss-120b"},"glm-5.2-caveman":{"id":"glm-5.2-caveman","name":"GLM-5.2 Caveman","description":"glm-5.2 carrying a built-in ruleset that compresses prose, keeping code and technical detail verbatim. The middle tier, and the ruleset as its authors wrote it. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.254,"output":5.016,"cache_read":0.3135},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"status":"deprecated","cost":{"input":1.756,"output":5.518},"canonical_model_id":"zhipuai/glm-5.1"},"glm-5.2-honey":{"id":"glm-5.2-honey","name":"GLM-5.2 Honey","description":"glm-5.2 carrying a built-in ruleset that compresses both generated code and prose. The middle tier, and the ruleset as its authors wrote it. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.254,"output":5.016,"cache_read":0.3135},"canonical_model_id":"zhipuai/glm-5.2"},"green-s":{"id":"green-s","name":"Green S","description":"GreenPT speech-to-text model for pre-recorded and live transcription","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-01","last_updated":"2025-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":8192},"cost":{"input":0.00437,"output":0}},"mistral-small-3.2-24b-instruct-2506":{"id":"mistral-small-3.2-24b-instruct-2506","name":"Mistral Small 3.2","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.228,"output":0.456},"canonical_model_id":"mistral/mistral-small-2506"},"glm-5.2-caveman-ultra":{"id":"glm-5.2-caveman-ultra","name":"GLM-5.2 Caveman Ultra","description":"glm-5.2 carrying a built-in ruleset that compresses prose, keeping code and technical detail verbatim. The most aggressive tier, close to answer-only. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.254,"output":5.016,"cache_read":0.3135},"canonical_model_id":"zhipuai/glm-5.2"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.9006,"output":4.389,"cache_read":0.1881},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"qwen3-coder-30b-a3b-instruct":{"id":"qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.285,"output":1.083},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.798,"output":4.959},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"voxtral-small-24b-2507":{"id":"voxtral-small-24b-2507","name":"Voxtral Small 24B","description":"Mistral Voxtral audio-understanding model for speech and transcription tasks","family":"mistral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.228,"output":0.513}},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.228,"output":0.798},"canonical_model_id":"openai/gpt-oss-120b"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.254,"output":5.016,"cache_read":0.3135},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.255552,"output":1.27776,"cache_read":0.0127776},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"holo2-30b-a3b":{"id":"holo2-30b-a3b","name":"Holo2 30B A3B","description":"H Company Holo2 vision model for GUI navigation and computer-use agents","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-11","last_updated":"2025-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":22016,"output":16384},"cost":{"input":0.399,"output":0.969}},"mistral-medium-3.5-128b":{"id":"mistral-medium-3.5-128b","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":2.052,"output":10.26},"canonical_model_id":"mistral/mistral-medium-2604"},"pixtral-12b-2409":{"id":"pixtral-12b-2409","name":"Pixtral 12B","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2024-09-11","last_updated":"2024-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.285,"output":0.285},"canonical_model_id":"mistral/pixtral-12b"},"minimax-m2.5":{"id":"minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.1938,"output":1.129,"cache_read":0.0627},"canonical_model_id":"minimax/MiniMax-M2.5"},"kimi-k2.6-fast":{"id":"kimi-k2.6-fast","name":"Kimi K2.6 Fast","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":1.655,"output":8.778},"canonical_model_id":"moonshotai/kimi-k2.6"}}},"cloudflare-ai-gateway":{"id":"cloudflare-ai-gateway","env":["CLOUDFLARE_API_TOKEN","CLOUDFLARE_ACCOUNT_ID","CLOUDFLARE_GATEWAY_ID"],"npm":"ai-gateway-provider","name":"Cloudflare AI Gateway","doc":"https://developers.cloudflare.com/ai-gateway/","models":{"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"xai/grok-4.3":{"id":"xai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"xai/grok-4.20-0309-reasoning":{"id":"xai/grok-4.20-0309-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"xai/grok-4.7":{"id":"xai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"xai/grok-4.5":{"id":"xai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"xai/grok-4.6":{"id":"xai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"xai/grok-4.20-0309-non-reasoning":{"id":"xai/grok-4.20-0309-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":384000},"cost":{"input":1.74,"output":3.48,"cache_read":0.145},"canonical_model_id":"deepseek/deepseek-v4-pro"},"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.16},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.16},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-open":{"id":"meta/muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1040000,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"meta/muse-wonder-1.0":{"id":"meta/muse-wonder-1.0","name":"Muse Wonder","description":"Muse Wonder is a frontier multimodal reasoning model from Meta for personal and professional agents, long-horizon coding, and computer use. It combines Active Memory, which carries relevant context across conversations, with Active Focus, which keeps it working through blockers toward a goal.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.16},"canonical_model_id":"meta/muse-spark-1.2"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"fireworks/ember-1":{"id":"fireworks/ember-1","name":"Ember-1","description":"Specialized model from Fireworks built on Kimi K3, producing shorter reasoning traces with approximately 40% fewer tokens while maintaining comparable quality","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1040000,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"fireworks/ember-1"},"reflection-ai/beam-1.0-501b-a23b":{"id":"reflection-ai/beam-1.0-501b-a23b","name":"Beam 1.0 501B A23B","description":"Sparse mixture-of-experts reasoning model with 23B active parameters for coding and agentic workflows","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"unbiased/pareto":{"id":"unbiased/pareto","name":"Pareto","description":"Blended multimodal model that runs multiple language models in parallel and returns a synthesized answer","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.25},"canonical_model_id":"unbiased/pareto"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"google/gemini-4-argon"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":5,"output":30},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":1.25,"output":5,"cache_read":0.625},"canonical_model_id":"openai/gpt-4o"},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":10,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"alibaba/qwen3.7-plus":{"id":"alibaba/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.32,"output":1.28,"cache_read":0.064},"canonical_model_id":"alibaba/qwen3.7-plus"},"alibaba/qwen3.5-397b-a17b":{"id":"alibaba/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"alibaba/qwen3.7-max":{"id":"alibaba/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.25,"output":3.75,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.7-max"},"alibaba/qwen3.8-max":{"id":"alibaba/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max"},"alibaba/qwen3-max":{"id":"alibaba/qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":1.2,"output":6},"canonical_model_id":"alibaba/qwen3-max"}}},"clarifai":{"id":"clarifai","env":["CLARIFAI_PAT"],"npm":"@ai-sdk/openai-compatible","api":"https://api.clarifai.com/v2/ext/openai/v1","name":"Clarifai","doc":"https://docs.clarifai.com/compute/inference/","models":{"deepseek-ai/deepseek-ocr/models/DeepSeek-OCR":{"id":"deepseek-ai/deepseek-ocr/models/DeepSeek-OCR","name":"DeepSeek OCR","description":"OCR model for extracting structured text from documents and screenshots","family":"deepseek","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-10-20","last_updated":"2026-02-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.2,"output":0.7}},"arcee_ai/AFM/models/trinity-mini":{"id":"arcee_ai/AFM/models/trinity-mini","name":"Trinity Mini","description":"Reasoning-tuned 26B MoE model with 3B active parameters for agents, tools, and multi-step workloads","family":"trinity","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-12-01","last_updated":"2026-02-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.045,"output":0.15},"canonical_model_id":"arcee-ai/trinity-mini"},"qwen/qwenCoder/models/Qwen3-Coder-30B-A3B-Instruct":{"id":"qwen/qwenCoder/models/Qwen3-Coder-30B-A3B-Instruct","name":"Qwen3 Coder 30B A3B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.11458,"output":0.74812}},"qwen/qwenLM/models/Qwen3-30B-A3B-Thinking-2507":{"id":"qwen/qwenLM/models/Qwen3-30B-A3B-Thinking-2507","name":"Qwen3 30B A3B Thinking 2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-31","last_updated":"2026-02-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.36,"output":1.3}},"qwen/qwenLM/models/Qwen3-30B-A3B-Instruct-2507":{"id":"qwen/qwenLM/models/Qwen3-30B-A3B-Instruct-2507","name":"Qwen3 30B A3B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2026-02-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":0.5}},"moonshotai/chat-completion/models/Kimi-K2_6":{"id":"moonshotai/chat-completion/models/Kimi-K2_6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4},"canonical_model_id":"moonshotai/kimi-k2.6"},"clarifai/main/models/mm-poly-8b":{"id":"clarifai/main/models/mm-poly-8b","name":"MM Poly 8B","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"mm-poly","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-06","last_updated":"2026-02-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":4096},"cost":{"input":0.658,"output":1.11}},"minimaxai/chat-completion/models/MiniMax-M2_5-high-throughput":{"id":"minimaxai/chat-completion/models/MiniMax-M2_5-high-throughput","name":"MiniMax-M2.5 High Throughput","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2}},"mistralai/completion/models/Ministral-3-14B-Reasoning-2512":{"id":"mistralai/completion/models/Ministral-3-14B-Reasoning-2512","name":"Ministral 3 14B Reasoning 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-01","last_updated":"2025-12-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":2.5,"output":1.7}},"mistralai/completion/models/Ministral-3-3B-Reasoning-2512":{"id":"mistralai/completion/models/Ministral-3-3B-Reasoning-2512","name":"Ministral 3 3B Reasoning 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12","last_updated":"2026-02-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.039,"output":0.54825}},"openai/chat-completion/models/gpt-oss-20b":{"id":"openai/chat-completion/models/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-12-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.045,"output":0.18}},"openai/chat-completion/models/gpt-oss-120b-high-throughput":{"id":"openai/chat-completion/models/gpt-oss-120b-high-throughput","name":"GPT OSS 120B High Throughput","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2026-02-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.09,"output":0.36}}}},"minimax-cn":{"id":"minimax-cn","env":["MINIMAX_API_KEY"],"npm":"@ai-sdk/anthropic","api":"https://api.minimax.cn/anthropic/v1","name":"MiniMax (minimax.cn)","doc":"https://platform.minimaxi.com/docs/guides/quickstart","models":{"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":524288},"cost":{"input":0.3137,"output":1.2549,"cache_read":0.0627,"tiers":[{"input":0.6275,"output":2.5098,"cache_read":0.1255,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":0.6275,"output":2.5098,"cache_read":0.1255}}},"MiniMax-M2.7-highspeed":{"id":"MiniMax-M2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.6275,"output":2.5098,"cache_read":0.0627,"cache_write":0.3922}},"MiniMax-M3.1-Flash":{"id":"MiniMax-M3.1-Flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0.1494,"output":0.5976,"cache_read":0.0149},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"MiniMax-M2":{"id":"MiniMax-M2","name":"MiniMax-M2","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3137,"output":1.2549,"cache_read":0.0314,"cache_write":0.3922}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3137,"output":1.2549,"cache_read":0.0314,"cache_write":0.3922}},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3137,"output":1.2549,"cache_read":0.0627,"cache_write":0.3922}},"MiniMax-M2.1":{"id":"MiniMax-M2.1","name":"MiniMax-M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3137,"output":1.2549,"cache_read":0.0314,"cache_write":0.3922}},"MiniMax-M2.5-highspeed":{"id":"MiniMax-M2.5-highspeed","name":"MiniMax-M2.5-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.6275,"output":2.5098,"cache_read":0.0314,"cache_write":0.3922}}}},"nebius":{"id":"nebius","env":["NEBIUS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.tokenfactory.nebius.com/v1","name":"Nebius Token Factory","doc":"https://docs.tokenfactory.nebius.com/","models":{"nvidia/Nemotron-3-Ultra-550b-a55b":{"id":"nvidia/Nemotron-3-Ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1,"output":3,"cache_read":1},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/Nemotron-3_5-Lightning":{"id":"nvidia/Nemotron-3_5-Lightning","name":"Nemotron 3.5 Lightning 30B A3B","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.06,"output":0.24,"cache_read":0.06},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-3-super-120b-a12b":{"id":"nvidia/nemotron-3-super-120b-a12b","name":"Nemotron-3-Super-120B-A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-03-11","last_updated":"2026-03-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.3,"output":0.9},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/Nemotron-3.5-Super-VL":{"id":"nvidia/Nemotron-3.5-Super-VL","name":"Nemotron 3.5 Super VL 120B A12B","description":"NVIDIA vision-language reasoning model for coding, visual understanding, and agentic tool use","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.3,"output":0.9,"cache_read":0.3},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":1024000},"cost":{"input":0.14,"output":0.28,"cache_read":0.14},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-V4.1-Flash":{"id":"deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048000,"output":1048000},"cost":{"input":0.3,"output":1.2,"cache_read":0.3},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.75,"output":3.5,"cache_read":0.15},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":979000,"output":979000},"cost":{"input":1.32,"output":3.96,"cache_read":1.32},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8000},"cost":{"input":0.95,"output":4},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":8000},"cost":{"input":3,"output":15,"cache_read":3},"canonical_model_id":"moonshotai/kimi-k3"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M3"},"reflection/beam":{"id":"reflection/beam","name":"Beam 1.0 501B A23B","description":"Sparse mixture-of-experts reasoning model with 23B active parameters for coding and agentic workflows","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.25,"output":1,"cache_read":0.25},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"meta-models/muse-spark-open":{"id":"meta-models/muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.25,"output":4.25,"cache_read":1.25},"canonical_model_id":"meta/muse-spark-open"},"Qwen/Qwen3-Embedding-8B":{"id":"Qwen/Qwen3-Embedding-8B","name":"Qwen3-Embedding-8B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-10","release_date":"2026-01-10","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"input":40960,"output":0},"cost":{"input":0.01,"output":0}},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8 27B","description":"Dense 27B reasoning model for coding, research, and long-horizon agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.45,"output":3,"cache_read":0.45},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3-30B-A3B-Instruct-2507":{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507","name":"Qwen3-30B-A3B-Instruct-2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2026-01-28","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":8192},"cost":{"input":0.1,"output":0.3,"cache_read":0.01,"cache_write":0.125}},"Qwen/Qwen3.5-397B-A17B":{"id":"Qwen/Qwen3.5-397B-A17B","name":"Qwen3.5-397B-A17B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-15","last_updated":"2026-05-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":250000,"output":8192},"cost":{"input":0.6,"output":3.6,"cache_read":0.06,"cache_write":0.75}},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-25","last_updated":"2025-10-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":8192},"cost":{"input":0.2,"output":0.6}},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":1024000},"cost":{"input":1.4,"output":4.4,"cache_read":1.4},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.4,"output":4.4},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":1024000},"cost":{"input":0.15,"output":0.5,"cache_read":0.15},"canonical_model_id":"zhipuai/glm-5.3-flash"},"NousResearch/Hermes-4-405B":{"id":"NousResearch/Hermes-4-405B","name":"Hermes-4-405B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-11","release_date":"2026-01-30","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":120000,"output":8192},"cost":{"input":1,"output":3,"reasoning":3,"cache_read":0.1,"cache_write":1.25}},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma-3-27b-it","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-10","release_date":"2026-01-20","last_updated":"2026-02-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":110000,"input":100000,"output":8192},"cost":{"input":0.1,"output":0.3,"cache_read":0.01,"cache_write":0.125}},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"gpt-oss-120b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-01-10","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":124000,"output":8192},"cost":{"input":0.15,"output":0.6,"reasoning":0.6,"cache_read":0.015,"cache_write":0.18}}}},"ai21":{"id":"ai21","env":["AI21_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.ai21.com/studio/v1","name":"AI21 Labs","doc":"https://docs.ai21.com/docs/jamba-foundation-models","models":{"jamba-mini":{"id":"jamba-mini","name":"Jamba Mini","description":"AI21's efficient, lightweight hybrid SSM-Transformer model for a wide range of tasks","family":"jamba","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08-22","release_date":"2026-01-08","last_updated":"2026-01-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":4096},"cost":{"input":0.2,"output":0.4},"canonical_model_id":"ai21/jamba-mini"},"jamba-large":{"id":"jamba-large","name":"Jamba Large","description":"AI21's hybrid SSM-Transformer long-context model for enterprise agents and grounded generation","family":"jamba","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08-22","release_date":"2025-07-03","last_updated":"2025-07-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":4096},"cost":{"input":2,"output":8},"canonical_model_id":"ai21/jamba-large"}}},"neon":{"id":"neon","env":["NEON_AI_GATEWAY_BASE_URL","NEON_AI_GATEWAY_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"${NEON_AI_GATEWAY_BASE_URL}/v1","name":"Neon","doc":"https://neon.com/docs","models":{"meta-llama-3-1-8b-instruct":{"id":"meta-llama-3-1-8b-instruct","name":"Llama 3.1 8B Instruct","description":"Meta's compact open-weight Llama 3.1 model for fast, low-cost text generation","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.15,"output":0.45}},"gemini-3-flash":{"id":"gemini-3-flash","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"gpt-5-6-sol":{"id":"gpt-5-6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":45,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"openai/gpt-5.6-sol"},"gemini-3-1-pro":{"id":"gemini-3-1-pro","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"inkling":{"id":"inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"claude-opus-4-1":{"id":"claude-opus-4-1","name":"Claude Opus 4.1 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"gpt-5-4":{"id":"gpt-5-4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"gpt-5-4-nano":{"id":"gpt-5-4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"gpt-5-1":{"id":"gpt-5-1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"gpt-5-3-codex":{"id":"gpt-5-3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"grok-4-7":{"id":"grok-4-7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":2.5,"output":7.5,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"gemma-3-12b":{"id":"gemma-3-12b","name":"Gemma 3 12B","description":"Google's open-weight Gemma 3 vision-language model for text and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.15,"output":0.5}},"gpt-5-6-terra":{"id":"gpt-5-6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"deepseek-v4-1-flash":{"id":"deepseek-v4-1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":127999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"llama-4-maverick":{"id":"llama-4-maverick","name":"Llama 4 Maverick 17B Instruct","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":8192},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"qwen3-next-80b-a3b-instruct":{"id":"qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":10000},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"gpt-5-5-pro":{"id":"gpt-5-5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"gpt-5-4-mini":{"id":"gpt-5-4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"gemini-3-1-flash-lite":{"id":"gemini-3-1-flash-lite","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"gpt-5-2":{"id":"gpt-5-2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"grok-4-6":{"id":"grok-4-6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":524288},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"qwen35-122b-a10b":{"id":"qwen35-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":25000},"cost":{"input":0.22,"output":2.2},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":25000},"cost":{"input":0.07,"output":0.3}},"gpt-5-5":{"id":"gpt-5-5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"gpt-6-1-sol":{"id":"gpt-6-1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"gpt-5-6-luna":{"id":"gpt-5-6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"glm-5-3":{"id":"glm-5-3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"gemini-3-6-flash":{"id":"gemini-3-6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.6-flash"},"glm-5-3-flash":{"id":"glm-5-3-flash","name":"GLM-5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"meta-llama-3-3-70b-instruct":{"id":"meta-llama-3-3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"glm-5-2":{"id":"glm-5-2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"gemini-3-8-flash":{"id":"gemini-3-8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.8-flash"},"gemini-3-5-flash":{"id":"gemini-3-5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":25000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"gemini-3-5-flash-lite":{"id":"gemini-3-5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gemini-3-7-flash":{"id":"gemini-3-7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.7-flash"}}},"cortecs":{"id":"cortecs","env":["CORTECS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.cortecs.ai/v1","name":"Cortecs","doc":"https://api.cortecs.ai/v1/models","models":{"gpt-4o-mini":{"id":"gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000},"cost":{"input":0.159,"output":0.638,"cache_read":0.081},"canonical_model_id":"openai/gpt-4o-mini"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.1,"output":0.35,"cache_read":0.018},"canonical_model_id":"zhipuai/glm-5.3-flash"},"apertus-70b":{"id":"apertus-70b","name":"Apertus 70B","description":"Apertus 70B is an open, multilingual language model designed for research, long-context reasoning, and sovereignty-focused AI systems.","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-09","release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":1.393,"output":2.228},"canonical_model_id":"swiss-ai/apertus-70b"},"qwen3.5-9b":{"id":"qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.111,"output":0.167},"canonical_model_id":"alibaba/qwen3.5-9b"},"qwen3-vl-235b-a22b":{"id":"qwen3-vl-235b-a22b","name":"qwen3-vl-235b-a22b","description":"Qwen3 VL 235B A22B is a 235B-parameter MoE vision-language flagship model (≈22B active) designed for frontier-level multimodal understanding across text, images, documents, and long videos.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-01-13","last_updated":"2026-01-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.617,"output":3.119,"cache_read":0.052}},"mistral-medium-3.5":{"id":"mistral-medium-3.5","name":"mistral-medium-3.5","description":"Mistral Medium 3.5 is a frontier multimodal 128B model combining reasoning, coding, and instruction-following with strong agentic performance and efficient deployment.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1.532,"output":7.843,"cache_read":0.154}},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":0.996,"output":4.982,"cache_read":0.099,"cache_write":1.186},"canonical_model_id":"anthropic/claude-haiku-4-5"},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.111,"output":0.434,"cache_read":0.056},"canonical_model_id":"openai/gpt-4.1-nano"},"claude-4-6-sonnet":{"id":"claude-4-6-sonnet","name":"Claude Sonnet 4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3.196,"output":15.94,"cache_read":0.32,"cache_write":3.999},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.434,"output":1.704,"cache_read":0.134},"canonical_model_id":"openai/gpt-4.1-mini"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.7,"output":3.4,"cache_read":0.15},"canonical_model_id":"deepseek/deepseek-v4-pro"},"minimax-m2":{"id":"minimax-m2","name":"MiniMax-M2","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":400000,"output":196000},"cost":{"input":0.349,"output":1.405},"canonical_model_id":"minimax/MiniMax-M2"},"claude-opus4-8":{"id":"claude-opus4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.437,"output":27.186,"cache_read":0.544,"cache_write":6.797},"canonical_model_id":"anthropic/claude-opus-4-8"},"nova-2-lite":{"id":"nova-2-lite","name":"Nova 2 Lite","description":"Nova 2 Lite is an advanced multimodal reasoning model that combines efficiency and performance, delivering reliable AI for agentic workflows and enterprise applications.","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.373,"output":3.144},"canonical_model_id":"amazon/nova-2-lite"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.082,"output":0.413,"cache_read":0.06},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":32768},"cost":{"input":0.167,"output":0.557},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.724,"output":0.724},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.475,"output":2.97,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.6"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.12,"output":0.6,"cache_read":0.012,"cache_write":0.15},"canonical_model_id":"openai/gpt-6-luna"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.375,"output":10.96,"cache_read":0.156},"canonical_model_id":"openai/gpt-5"},"gemma-3-27b-it":{"id":"gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Gemma 3 is a family of lightweight, multimodal models from Google, supporting text and image inputs, multilingual capabilities, and a 131K context window.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":110000},"cost":{"input":0.099,"output":0.299},"canonical_model_id":"google/gemma-3-27b-it"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":14.999},"canonical_model_id":"moonshotai/kimi-k3"},"mixtral-8x7B-instruct-v0.1":{"id":"mixtral-8x7B-instruct-v0.1","name":"Mixtral 8x7B Instruct v0.1","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2023-12-11","last_updated":"2023-12-11","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.488,"output":0.758}},"qwen3-235b-a22b-instruct-2507":{"id":"qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B-A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.199,"output":0.598,"cache_read":0.018},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.2,"output":13.199,"cache_read":0.219,"cache_write":2.749},"canonical_model_id":"openai/gpt-5.6-terra"},"nvidia-nemotron-3-nano-30b-a3b":{"id":"nvidia-nemotron-3-nano-30b-a3b","name":"nvidia-nemotron-3-nano-30b-a3b","description":"Nemotron-Nano-3-30B-A3B is a compact Mixture-of-Experts model optimized for efficient reasoning, chat, and coding, with strong multilingual support and long-context RAG and agent workflows.","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-01-12","last_updated":"2026-01-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.06,"output":0.24}},"claude-opus-5.5":{"id":"claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4.399,"output":21.998,"cache_read":0.219,"cache_write":5.5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":1.495,"output":9.964,"cache_read":0.242,"cache_write":0.434},"canonical_model_id":"google/gemini-2.5-pro"},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2.192,"output":8.769,"cache_read":0.546},"canonical_model_id":"openai/gpt-4.1"},"ministral-8b-2512":{"id":"ministral-8b-2512","name":"Ministral 3 8B 2512","description":"Ministral 3 8B is a balanced, efficient multimodal model offering strong text and vision capabilities, optimized for edge and local deployment.","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.179,"output":0.179,"cache_read":0.017},"canonical_model_id":"mistral/ministral-8b-2512"},"mistral-nemo-instruct-2407":{"id":"mistral-nemo-instruct-2407","name":"mistral-nemo-instruct-2407","description":"A 12B parameter, instruct-tuned language model by Mistral AI and NVIDIA, designed for advanced instruction following, multi-turn conversations, and generating text and code across multiple languages.","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2024-08-07","last_updated":"2024-08-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.145,"output":0.145,"cache_read":0.014}},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.279,"output":2.192,"cache_read":0.056},"canonical_model_id":"openai/gpt-5-mini"},"qwen3.8-2.4t-a95b":{"id":"qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":2.5,"output":6,"cache_read":0.625},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"nova-lite-v1":{"id":"nova-lite-v1","name":"Nova Lite 1.0","description":"Nova Lite is a fast, low-cost multimodal foundation model capable of reasoning over text, images, and video in 200+ languages.","family":"nova-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.069,"output":0.275},"canonical_model_id":"amazon/nova-lite-v1"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.272,"output":1.631,"cache_read":0.025,"cache_write":0.082},"canonical_model_id":"google/gemini-3.1-flash-lite"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.038},"canonical_model_id":"google/gemini-3.7-flash"},"claude-opus4-5":{"id":"claude-opus4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5.313,"output":26.568,"cache_read":0.531,"cache_write":6.645},"canonical_model_id":"anthropic/claude-opus-4-5"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":2,"output":3.999,"cache_read":0.5},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.219,"output":1.32,"cache_read":0.022,"cache_write":0.275},"canonical_model_id":"openai/gpt-5.6-luna"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.511,"output":2.045,"cache_read":0.154},"canonical_model_id":"zhipuai/glm-5.3"},"claude-opus4-6":{"id":"claude-opus4-6","name":"Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.313,"output":26.561,"cache_read":0.531,"cache_write":6.645},"canonical_model_id":"anthropic/claude-opus-4-6"},"minicpm-v-4.5":{"id":"minicpm-v-4.5","name":"minicpm-v-4.5","description":"MiniCPM-V 4.5 is a compact, high-performance vision-language model excelling in video understanding, OCR, and multimodal reasoning with efficient deployment.","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"cost":{"input":0.651,"output":1.097}},"devstral-2512":{"id":"devstral-2512","name":"Devstral 2","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.478,"output":2.392,"cache_read":0.045},"canonical_model_id":"mistral/devstral-2512"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.825,"output":4.125,"cache_read":0.082,"cache_write":0.084},"canonical_model_id":"google/gemini-3.8-flash"},"claude-sonnet-5.5":{"id":"claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.25,"output":11.249,"cache_read":0.225,"cache_write":2.813},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2.898,"output":15.453,"cache_read":0.242},"canonical_model_id":"openai/gpt-5.4"},"claude-4-5-sonnet":{"id":"claude-4-5-sonnet","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":2.989,"output":14.945,"cache_read":0.326,"cache_write":4.078},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000},"cost":{"input":2.659,"output":10.635,"cache_read":1.33},"canonical_model_id":"openai/gpt-4o"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":202752},"cost":{"input":1.384,"output":4.348,"cache_read":0.346},"canonical_model_id":"zhipuai/glm-5.1"},"ministral-3b-2512":{"id":"ministral-3b-2512","name":"Ministral 3 3B 2512","description":"Ministral 3 3B is a compact, efficient multimodal model with strong language, vision capabilities, and ideal for custom fine-tuning.","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.123,"output":0.123,"cache_read":0.012},"canonical_model_id":"mistral/ministral-3b-2512"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":1.649,"output":9.899,"cache_read":0.165,"cache_write":1},"canonical_model_id":"google/gemini-3.5-flash"},"pixtral-large-2502":{"id":"pixtral-large-2502","name":"Pixtral Large (25.02)","description":"Pixtral Large (25.02) is a 124B open-weight multimodal model built on Mistral Large 2, offering advanced image understanding and strong performance across text and code tasks.","family":"pixtral","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-04-08","last_updated":"2025-04-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":1.993,"output":5.978},"canonical_model_id":"mistral/pixtral-large-2502"},"hermes-4-405b":{"id":"hermes-4-405b","name":"Hermes 4 405B","description":"Hermes 4 405B is a frontier hybrid-mode reasoning model built on Llama 3.1, optimized for advanced logic, math, coding, and structured output generation.","family":"hermes","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-08-31","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.996,"output":2.989},"canonical_model_id":"nousresearch/hermes-4-405b"},"minimax-m2.1":{"id":"minimax-m2.1","name":"MiniMax-M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196000,"output":196000},"cost":{"input":0.359,"output":1.435},"canonical_model_id":"minimax/MiniMax-M2.1"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.045,"output":0.167},"canonical_model_id":"openai/gpt-oss-20b"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":2.25,"output":11.249,"cache_read":0.113},"canonical_model_id":"google/gemini-4-argon"},"qwen3.8-flash-next":{"id":"qwen3.8-flash-next","name":"Qwen3.8 Flash Next","description":"Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":64000},"cost":{"input":0.201,"output":0.5,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.8-flash-next"},"gemma-4-26b-a4b-it":{"id":"gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":81920},"cost":{"input":0.111,"output":0.557},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"qwen3-30b-a3b-instruct-2507":{"id":"qwen3-30b-a3b-instruct-2507","name":"qwen3-30b-a3b-instruct-2507","description":"Qwen3-30B-A3B-Instruct-2507 is an advanced Mixture-of-Experts model optimized for reasoning, coding, and multilingual instruction following.","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.099,"output":0.299}},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":12.272,"output":61.359,"cache_read":1.228,"cache_write":15.34},"canonical_model_id":"openai/gpt-6-astra"},"mistral-small-3.2-24b-instruct-2506":{"id":"mistral-small-3.2-24b-instruct-2506","name":"mistral-small-3.2-24b-instruct-2506","description":"Mistral-Small-3.2-24B-Instruct-2506 is a 24B parameter instruction-tuned model with enhanced long-context support (128k) and state-of-the-art vision understanding.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-05-26","last_updated":"2025-05-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":131000},"cost":{"input":0.1,"output":0.312}},"mistral-large-2512":{"id":"mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.613,"output":1.838,"cache_read":0.061},"canonical_model_id":"mistral/mistral-large-2512"},"qwen3guard-gen-8b":{"id":"qwen3guard-gen-8b","name":"qwen3guard-gen-8b","description":"Qwen3Guard-Gen-8B is a large-scale multilingual safety moderation model designed for high-accuracy prompt and response classification.","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-02-04","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"cost":{"input":0,"output":0}},"nova-pro-v1":{"id":"nova-pro-v1","name":"Nova Pro 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.918,"output":3.671},"canonical_model_id":"amazon/nova-pro-v1"},"qwen3-32b":{"id":"qwen3-32b","name":"Qwen3 32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":0.179,"output":0.697},"canonical_model_id":"alibaba/qwen3-32b"},"mistral-small-2603":{"id":"mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.156,"output":0.625,"cache_read":0.016},"canonical_model_id":"mistral/mistral-small-2603"},"claude-haiku-5.5":{"id":"claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.113,"output":0.563,"cache_read":0.011,"cache_write":0.14},"canonical_model_id":"anthropic/claude-haiku-5-5"},"glm-4.7-flash":{"id":"glm-4.7-flash","name":"GLM-4.7-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":203000},"cost":{"input":0.08,"output":0.478},"canonical_model_id":"zhipuai/glm-4.7-flash"},"nemotron-nano-v2-12b":{"id":"nemotron-nano-v2-12b","name":"nemotron-nano-v2-12b","description":"NVIDIA Nemotron Nano v2 12B is a 12-billion-parameter multimodal reasoning model designed for advanced video understanding, document intelligence, and visual reasoning, built with a hybrid Transformer-Mamba architecture for high efficiency and low latency.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-10-31","last_updated":"2025-10-31","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.24,"output":0.707}},"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":500000},"cost":{"input":0.1,"output":0.4,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.8-27b"},"gpt-oss-safeguard-120b":{"id":"gpt-oss-safeguard-120b","name":"GPT OSS Safeguard 120B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.179,"output":0.697},"canonical_model_id":"openai/gpt-oss-safeguard-120b"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.656,"output":3.3,"cache_read":0.18},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"qwen3-coder-30b-a3b-instruct":{"id":"qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.179,"output":0.697,"cache_read":0.014},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4.399,"output":21.998,"cache_read":0.44,"cache_write":5.5},"canonical_model_id":"openai/gpt-5.6-sol"},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.668,"output":4.01},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"mistral-large-4-0":{"id":"mistral-large-4-0","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":262144},"cost":{"input":0.765,"output":2.351,"cache_read":0.079},"canonical_model_id":"mistral/mistral-large-4"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.498,"cache_read":0.55,"cache_write":6.874},"canonical_model_id":"anthropic/claude-opus-5"},"mistral-7b-instruct-v0.2":{"id":"mistral-7b-instruct-v0.2","name":"mistral-7b-instruct-v0.2","description":"Mistral 7B Instruct is a compact, 7B parameter model optimized for fast and efficient text and code generation with a 32K token context window.","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2025-05-26","last_updated":"2025-05-26","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":8192},"cost":{"input":0.159,"output":0.219}},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.4,"output":11.999,"cache_read":0.12,"cache_write":3},"canonical_model_id":"openai/gpt-6.1-sol"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":65536},"cost":{"input":0.5,"output":1.499,"cache_read":0.15},"canonical_model_id":"deepseek/deepseek-v3.2"},"nova-micro-v1":{"id":"nova-micro-v1","name":"Nova Micro 1.0","description":"Nova Micro is a multilingual text-to-text foundation model with strong reasoning capabilities and broad language coverage across 200+ languages.","family":"nova-micro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.04,"output":0.159},"canonical_model_id":"amazon/nova-micro-v1"},"ministral-14b-2512":{"id":"ministral-14b-2512","name":"Ministral 3 14B 2512","description":"Ministral 3 14B is a frontier-level 14B multimodal model optimized for local deployment, delivering state-of-the-art text and vision reasoning with a 256K context window and strong agentic capabilities.","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.24,"output":0.24,"cache_read":0.022},"canonical_model_id":"mistral/ministral-14b-2512"},"qwen3guard-gen-0.6b":{"id":"qwen3guard-gen-0.6b","name":"qwen3guard-gen-0.6b","description":"Qwen3Guard-Gen-0.6B is a lightweight multilingual safety moderation model that classifies prompts and responses into safe, controversial, or unsafe categories.","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-02-04","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"cost":{"input":0,"output":0}},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.33,"output":2.749,"cache_read":0.033,"cache_write":0.082},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.089,"output":0.446,"cache_read":0.01},"canonical_model_id":"openai/gpt-oss-120b"},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.375,"output":10.96,"cache_read":0.156},"canonical_model_id":"openai/gpt-5.1"},"codestral-2508":{"id":"codestral-2508","name":"Codestral 25.08","description":"Mistral coding model for code completion, generation, and developer workflows","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.368,"output":1.103,"cache_read":0.037},"canonical_model_id":"mistral/codestral-2508"},"qwen3-coder-next":{"id":"qwen3-coder-next","name":"Qwen3 Coder Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.167,"output":0.891},"canonical_model_id":"alibaba/qwen3-coder-next"},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.511,"output":1.534,"cache_read":0.128},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"gemma-4-31b-it":{"id":"gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.223,"output":0.39},"canonical_model_id":"google/gemma-4-31b-it"},"qwen3.6-27b":{"id":"qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.446,"output":3.008},"canonical_model_id":"alibaba/qwen3.6-27b"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.4,"output":11.999,"cache_read":0.24,"cache_write":3},"canonical_model_id":"openai/gpt-6-sol"},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.299,"output":2.491,"cache_read":0.029,"cache_write":0.097},"canonical_model_id":"google/gemini-2.5-flash"},"claude-opus4-7":{"id":"claude-opus4-7","name":"Claude Opus 4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.437,"output":27.186,"cache_read":0.544,"cache_write":6.797},"canonical_model_id":"anthropic/claude-opus-4-7"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1000000},"cost":{"input":0.9,"output":3.24,"cache_read":0.189},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.201,"output":0.6,"cache_read":0.004},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen2.5-vl-72b-instruct":{"id":"qwen2.5-vl-72b-instruct","name":"qwen2.5-vl-72b-instruct","description":"Qwen2.5-VL is a powerful vision-language model with advanced capabilities in visual understanding, long video reasoning, and structured output generation.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":false,"release_date":"2025-01-27","last_updated":"2025-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"cost":{"input":1.014,"output":1.014}},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.038},"canonical_model_id":"google/gemini-3.6-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.219,"cache_write":2.749},"canonical_model_id":"anthropic/claude-sonnet-5"},"voxtral-small-2507":{"id":"voxtral-small-2507","name":"voxtral-small-2507","description":"Voxtral Small is a multimodal model with audio input, combining advanced speech capabilities with strong text performance for transcription, translation, and audio understanding.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-02-02","last_updated":"2026-02-02","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"cost":{"input":0.123,"output":0.368,"cache_read":0.012}},"pixtral-12b-2409":{"id":"pixtral-12b-2409","name":"pixtral-12b-2409","description":"Pixtral 2409 12B is a state-of-the-art multimodal model with 12B parameters and a 400M vision encoder, natively trained on interleaved text and image data. It excels in tasks spanning vision-language reasoning, instruction following, and pure text understanding, making it highly effective for real-world multimodal applications.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2024-11-09","last_updated":"2024-11-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":0.223,"output":0.223}},"mistral-large-2402":{"id":"mistral-large-2402","name":"mistral-large-2402","description":"Mistral Large (24.02) is Mistral AI’s most advanced language model, built for complex multilingual reasoning, code generation, and deep text understanding.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-05-26","last_updated":"2025-05-26","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":8192},"cost":{"input":4.284,"output":12.952}},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.06,"output":0.439,"cache_read":0.019},"canonical_model_id":"openai/gpt-5-nano"},"minimax-m2.5":{"id":"minimax-m2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196000,"output":196000},"cost":{"input":0.348,"output":1.39,"cache_read":0.075},"canonical_model_id":"minimax/MiniMax-M2.5"},"mistral-small-2503":{"id":"mistral-small-2503","name":"mistral-small-2503","description":"Combines advanced text and vision capabilities with 24 billion parameters, supporting multilingual tasks and long contexts up to 131k tokens, making it versatile for various applications without sacrificing performance.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-03-20","last_updated":"2025-03-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.111,"output":0.334}}}},"lmstudio":{"id":"lmstudio","env":["LMSTUDIO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"http://127.0.0.1:1234/v1","name":"LMStudio","doc":"https://lmstudio.ai/models","models":{"qwen/qwen3-coder-30b":{"id":"qwen/qwen3-coder-30b","name":"Qwen3 Coder 30B","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0}},"qwen/qwen3-30b-a3b-2507":{"id":"qwen/qwen3-30b-a3b-2507","name":"Qwen3 30B A3B 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0,"output":0}},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0,"output":0}}}},"claudinio":{"id":"claudinio","env":["CLAUDINIO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.claudin.io/v1","name":"Claudinio","doc":"https://claudin.io","models":{"claudinio":{"id":"claudinio","name":"Claudinio","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"knowledge":"2026-05","release_date":"2026-05-12","last_updated":"2026-06-02","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":0.5,"output":2,"cache_read":0.15}},"claudius":{"id":"claudius","name":"Claudius","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"knowledge":"2026-05","release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":3,"output":8,"cache_read":0.9}}}},"nearai":{"id":"nearai","env":["NEARAI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://cloud-api.near.ai/v1","name":"NEAR AI Cloud","doc":"https://docs.near.ai/","models":{"anthropic/claude-haiku-4-5":{"id":"anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-fable-5-1":{"id":"anthropic/claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-opus-4-8":{"id":"anthropic/claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-sonnet-4-5":{"id":"anthropic/claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-sonnet-5-5":{"id":"anthropic/claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-haiku-5-5":{"id":"anthropic/claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-5-5":{"id":"anthropic/claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2},"canonical_model_id":"anthropic/claude-sonnet-5"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":64000},"cost":{"input":1.1,"output":1.1,"cache_read":0.55},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.08,"output":0.5,"cache_read":0.012},"canonical_model_id":"zhipuai/glm-5.3-flash"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.5,"output":3.3,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.8,"output":7.5},"canonical_model_id":"alibaba/qwen3.7-max"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.81,"output":3.85,"cache_read":0.41},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3.3,"output":16.5,"cache_read":0.33},"canonical_model_id":"moonshotai/kimi-k3"},"Qwen/Qwen3.6-35B-A3B-FP8":{"id":"Qwen/Qwen3.6-35B-A3B-FP8","name":"Qwen 3.6 35B A3B FP8","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.17,"output":1.1,"cache_read":0.056},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen/Qwen3-VL-30B-A3B-Instruct":{"id":"Qwen/Qwen3-VL-30B-A3B-Instruct","name":"Qwen3-VL 30B-A3B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":8192},"cost":{"input":0.15,"output":0.55,"cache_read":0.03}},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.04,"output":1.35,"cache_read":0.018},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3-Embedding-0.6B":{"id":"Qwen/Qwen3-Embedding-0.6B","name":"Qwen3 Embedding 0.6B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-03","last_updated":"2025-06-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":1024},"cost":{"input":0.01,"output":0.01}},"Qwen/Qwen3-Reranker-0.6B":{"id":"Qwen/Qwen3-Reranker-0.6B","name":"Qwen3 Reranker 0.6B","description":"Reranking model for improving retrieval quality in search and recommendation systems","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-03","last_updated":"2025-06-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":1024},"cost":{"input":0.01,"output":0.01}},"black-forest-labs/FLUX.2-klein-4B":{"id":"black-forest-labs/FLUX.2-klein-4B","name":"FLUX.2 Klein 4B","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-14","last_updated":"2026-01-14","modalities":{"input":["text"],"output":["image"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":1,"output":1}},"x-ai/grok-4.7":{"id":"x-ai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/whisper-large-v3":{"id":"openai/whisper-large-v3","name":"Whisper Large v3","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":448},"cost":{"input":0.01,"output":0.01}},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"}}},"zhipuai":{"id":"zhipuai","env":["ZHIPU_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://open.bigmodel.cn/api/paas/v4","name":"Zhipu AI","doc":"https://docs.bigmodel.cn/cn/guide/start/pricing","models":{"glm-4.6v":{"id":"glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.1494,"output":0.44819,"cache_read":0.02988,"cache_write":0,"tiers":[{"input":0.29879,"output":0.89637,"cache_read":0.05976,"cache_write":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6v"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.11952,"output":0.41831,"cache_read":0.03436,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-4.6v-flash":{"id":"glm-4.6v-flash","name":"GLM-4.6V-Flash","description":"Lightweight GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6v-flash"},"glm-4.5v":{"id":"glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":16384},"cost":{"input":0.29879,"output":0.89637,"cache_read":0.05976,"cache_write":0,"tiers":[{"input":0.59758,"output":1.79274,"cache_read":0.11952,"cache_write":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.5v"},"glm-4.5-air":{"id":"glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.11952,"output":0.89637,"cache_read":0.0239,"cache_write":0,"tiers":[{"input":0.17927,"output":1.19516,"cache_read":0.03585,"cache_write":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.5-air"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.19516,"output":4.18306,"cache_read":0.29879,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.44819,"output":2.09153,"cache_read":0.08964,"cache_write":0,"tiers":[{"input":0.59758,"output":2.39032,"cache_read":0.11952,"cache_write":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.7"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.89637,"output":3.58548,"cache_read":0.19421,"cache_write":0,"tiers":[{"input":1.19516,"output":4.18306,"cache_read":0.29879,"cache_write":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5.1"},"glm-4.7-flash":{"id":"glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"glm-5":{"id":"glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.59758,"output":2.68911,"cache_read":0.1494,"cache_write":0,"tiers":[{"input":0.89637,"output":3.28669,"cache_read":0.22409,"cache_write":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5"},"glm-5v-turbo":{"id":"glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":0.74698,"output":3.28669,"cache_read":0.17927,"cache_write":0,"tiers":[{"input":1.04577,"output":3.88427,"cache_read":0.26891,"cache_write":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5v-turbo"},"glm-5.3-flashx":{"id":"glm-5.3-flashx","name":"GLM-5.3-FlashX","description":"High-speed GLM-5.3-Flash serving option for coding and agent workflows","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.29879,"output":1.04577,"cache_read":0.08516,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-4.6":{"id":"glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.19516,"output":4.18306,"cache_read":0.29879,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"glm-4.5":{"id":"glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5"},"glm-4.7-flashx":{"id":"glm-4.7-flashx","name":"GLM-4.7-FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.0747,"output":0.44819,"cache_read":0.01494,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"glm-4.5-flash":{"id":"glm-4.5-flash","name":"GLM-4.5-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":98304},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-flash"}}},"requesty":{"id":"requesty","env":["REQUESTY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://router.requesty.ai/v1","name":"Requesty","doc":"https://requesty.ai/solution/llm-routing/models","models":{"leanstral-1-5@eu":{"id":"leanstral-1-5@eu","name":"leanstral-1-5@eu","description":"Leanstral 1.5 is an updated Lean 4 formal proof engineering model from Mistral AI, optimized for automated theorem proving and autoformalization. It has 119B total parameters with 6.5B active and supports a 256K token context window. It supports native function calling and structured output.","family":"leanstral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-05-27","last_updated":"2026-05-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0}},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":262144},"cost":{"input":0.2,"output":0.6,"cache_read":0.07},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5.2@eu":{"id":"glm-5.2@eu","name":"GLM-5.2 (EU)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.2,"output":4.2,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4-pro-0813@eu":{"id":"deepseek-v4-pro-0813@eu","name":"DeepSeek V4 Pro 0813 (EU)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":1.75,"output":3.5,"cache_read":0.44},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"claude-opus-4-8@eu":{"id":"claude-opus-4-8@eu","name":"Claude Opus 4.8 (EU)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-8"},"nemotron-3.5-content-safety":{"id":"nemotron-3.5-content-safety","name":"Nemotron 3.5 Content Safety","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5.5,"output":33,"cache_read":0.55},"canonical_model_id":"openai/gpt-5.5"},"nemotron-3.5-lightning-30b-a3b":{"id":"nemotron-3.5-lightning-30b-a3b","name":"nemotron-3.5-lightning-30b-a3b","description":"NVIDIA Nemotron 3.5 Lightning 30B-A3B is a hybrid Mamba-2 + MoE + Attention model with 30B total and 3B active parameters, pre-trained on over 20T tokens with an NVFP4 recipe and Multi-Token Prediction for fast generation. Up to 1M token context for long-running autonomous agents, sub-agent workhorse deployments, and agentic workflows. Supports reasoning and tool calling. English and coding languages plus Spanish, French, German, Italian, and Japanese. Open weights under the OpenMDW License Agreement v1.1. Part of the NVIDIA Nemotron family.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0}},"inkling":{"id":"inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":32768},"cost":{"input":1.87,"output":4.68,"cache_read":0.374},"canonical_model_id":"thinkingmachines/inkling"},"claude-sonnet-5-5@eu":{"id":"claude-sonnet-5-5@eu","name":"Claude Sonnet 5.5 (EU)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gemini-3.5-flash-lite@eu":{"id":"gemini-3.5-flash-lite@eu","name":"Gemini 3.5 Flash Lite (EU)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.33,"output":2.75,"cache_read":0.033},"canonical_model_id":"google/gemini-3.5-flash-lite"},"claude-opus-5@eu":{"id":"claude-opus-5@eu","name":"Claude Opus 5 (EU)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-5"},"claude-fable-5.1@eu":{"id":"claude-fable-5.1@eu","name":"Claude Fable 5.1 (EU)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":0.275,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5-1"},"gpt-4o-mini@eu":{"id":"gpt-4o-mini@eu","name":"GPT-4o mini (EU)","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000},"cost":{"input":0.165,"output":0.66,"cache_read":0.0825},"canonical_model_id":"openai/gpt-4o-mini"},"claude-opus-4-1":{"id":"claude-opus-4-1","name":"Claude Opus 4.1 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"gemini-3.1-flash-image":{"id":"gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","video","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":2},"canonical_model_id":"google/gemini-3.1-flash-image"},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":1.2},"canonical_model_id":"minimax/MiniMax-M2.7"},"gpt-5.6-terra@eu":{"id":"gpt-5.6-terra@eu","name":"GPT-5.6 Terra (EU)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.2,"output":13.2,"cache_read":0.22,"cache_write":2.75,"tiers":[{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5}},"canonical_model_id":"openai/gpt-5.6-terra"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"ling-2.6-1t":{"id":"ling-2.6-1t","name":"ling-2.6-1t","description":"Inclusion AI ling-2.6-1t","family":"ling","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":2.5}},"gpt-5.5@eu":{"id":"gpt-5.5@eu","name":"GPT-5.5 (EU)","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5.5,"output":33,"cache_read":0.55,"tiers":[{"input":11,"output":49.5,"cache_read":1.1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":11,"output":49.5,"cache_read":1.1}},"canonical_model_id":"openai/gpt-5.5"},"gpt-6.1-sol@eu":{"id":"gpt-6.1-sol@eu","name":"GPT-6.1 Sol (EU)","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.4,"output":12,"cache_read":0.12,"cache_write":3,"tiers":[{"input":4.8,"output":18,"cache_read":0.24,"cache_write":6,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.8,"output":18,"cache_read":0.24,"cache_write":6}},"canonical_model_id":"openai/gpt-6.1-sol"},"kimi-k2.5@eu":{"id":"kimi-k2.5@eu","name":"Kimi K2.5 (EU)","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16000},"cost":{"input":0.72,"output":3.6,"cache_read":0.72,"cache_write":0.72},"canonical_model_id":"moonshotai/kimi-k2.5"},"claude-haiku-4-5@eu":{"id":"claude-haiku-4-5@eu","name":"Claude Haiku 4.5 (latest) (EU)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5"},"nemotron-lightning-3.5-30b-a3b":{"id":"nemotron-lightning-3.5-30b-a3b","name":"nemotron-lightning-3.5-30b-a3b","description":"Nemotron-Lightning-3.5-30B-A3B is a 30B-parameter Mixture-of-Experts language model (3B active) from NVIDIA's Nemotron-H family, built on a hybrid Mamba-Transformer architecture for efficient long-context inference. Like other models in the family, it responds to queries by first generating a reasoning trace and then concluding with a final response, with reasoning behavior configurable through a flag in the chat template. It includes a multi-token prediction (MTP) speculative decoding head for low-latency serving.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-08-15","last_updated":"2026-08-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.05,"output":0.2,"cache_read":0.01}},"gpt-5.4-pro":{"id":"gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"cache_read":30},"canonical_model_id":"openai/gpt-5.4-pro"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro"},"nvidia-nemotron-3-super-120b-a12b":{"id":"nvidia-nemotron-3-super-120b-a12b","name":"nvidia-nemotron-3-super-120b-a12b","description":"NVIDIA Nemotron 3 Super is a hybrid Mixture-of-Experts (MoE) model engineered for highest compute efficiency and accuracy in multi-agent applications and specialized agentic systems. It is optimized to run many collaborating agents per application on a single GPU, delivering high accuracy for reasoning, tool use, and instruction following.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.5}},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5"},"deepseek-v4.1-flash-1017@eu":{"id":"deepseek-v4.1-flash-1017@eu","name":"DeepSeek V4.1 Flash 1017 (EU)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"seed-2.0-pro":{"id":"seed-2.0-pro","name":"Seed 2.0 Pro","description":"Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.5,"output":3,"cache_read":0.1},"canonical_model_id":"bytedance-seed/seed-2.0-pro"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.28,"output":0.56,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"claude-opus-4-6@eu":{"id":"claude-opus-4-6@eu","name":"Claude Opus 4.6 (EU)","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-6"},"gpt-4.1-mini@eu":{"id":"gpt-4.1-mini@eu","name":"GPT-4.1 mini (EU)","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.44,"output":1.76,"cache_read":0.11},"canonical_model_id":"openai/gpt-4.1-mini"},"deepseek-v4-pro@eu":{"id":"deepseek-v4-pro@eu","name":"DeepSeek V4 Pro (EU)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.75,"output":3.5,"cache_read":0.44},"canonical_model_id":"deepseek/deepseek-v4-pro"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"cache_write":1.25,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"cache_write":2.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4,"cache_write":2.5}},"canonical_model_id":"xai/grok-4.3"},"gpt-5.6-luna@eu":{"id":"gpt-5.6-luna@eu","name":"GPT-5.6 Luna (EU)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.22,"output":1.32,"cache_read":0.022,"cache_write":0.275,"tiers":[{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55}},"canonical_model_id":"openai/gpt-5.6-luna"},"deepseek-v4.1-flash@eu":{"id":"deepseek-v4.1-flash@eu","name":"DeepSeek V4.1 Flash (EU)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":3,"output":15,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"glm-5.2-fast":{"id":"glm-5.2-fast","name":"glm-5.2-fast","description":"GLM-5.2 introduces a robust 1M-token context and advanced, multi-effort coding capabilities to significantly enhance performance on long-horizon tasks. Its new IndexShare architecture and improved MTP layer simultaneously boost efficiency by reducing per-token FLOPs and increasing speculative decoding lengths. A 743B-parameter model in Zhipu AI's GLM series, designed to plan, execute, and iterate autonomously on extended, engineering-grade tasks.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-07-13","last_updated":"2026-07-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.1,"output":6.6,"cache_read":0.21}},"ling-3.0-tiny":{"id":"ling-3.0-tiny","name":"ling-3.0-tiny","description":"Ling-3.0-tiny is an efficient 7.9B parameter MoE model from inclusionAI with only 1.3B active parameters per token. Built for responsive agents, reliable instruction following and multi turn conversation, with a 256K context window, native function calling, prompt caching and switchable Thinking and Instant modes.","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0}},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625},"canonical_model_id":"alibaba/qwen3.6-plus"},"ring-2.6-1t":{"id":"ring-2.6-1t","name":"ring-2.6-1t","description":"Inclusion AI ring-2.6-1t","family":"ring","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-05-08","last_updated":"2026-05-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":2.5}},"inkling-256k":{"id":"inkling-256k","name":"inkling-256k","description":"Inkling 256K is the extended context variant of Inkling, a large MoE hybrid reasoning model from Thinking Machines with audio and vision input support and a 256K context window.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":1.87,"output":4.68,"cache_read":0.374}},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"gemini-2.5-flash-lite@eu":{"id":"gemini-2.5-flash-lite@eu","name":"Gemini 2.5 Flash-Lite (EU)","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":0.18333},"canonical_model_id":"google/gemini-2.5-flash-lite"},"qwen3.8-2.4T-A95B":{"id":"qwen3.8-2.4T-A95B","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"devstral-latest@eu":{"id":"devstral-latest@eu","name":"devstral-latest@eu","description":"An enterprise grade text model, that excels at using tools to explore codebases, editing multiple files and power software engineering agents.","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-05-27","last_updated":"2026-05-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.44,"output":2.2,"cache_read":0.44}},"seed-1.8":{"id":"seed-1.8","name":"seed-1.8","description":"Optimized specifically for multimodal agent scenarios. It features enhanced agent capabilities, upgraded multimodal comprehension, and more flexible context management.","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-05-27","last_updated":"2026-05-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"leanstral-1-5":{"id":"leanstral-1-5","name":"leanstral-1-5","description":"Leanstral 1.5 is an updated Lean 4 formal proof engineering model from Mistral AI, optimized for automated theorem proving and autoformalization. It has 119B total parameters with 6.5B active and supports a 256K token context window. It supports native function calling and structured output.","family":"leanstral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-05-27","last_updated":"2026-05-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0}},"ling-2.6-flash":{"id":"ling-2.6-flash","name":"ling-2.6-flash","description":"Inclusion AI ling-2.6-flash","family":"ling","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.3}},"gemini-2.5-flash@eu":{"id":"gemini-2.5-flash@eu","name":"Gemini 2.5 Flash (EU)","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.3,"output":2.5,"cache_read":0.075,"cache_write":0.55},"canonical_model_id":"google/gemini-2.5-flash"},"gemini-3.5-flash@eu":{"id":"gemini-3.5-flash@eu","name":"Gemini 3.5 Flash (EU)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":1.65,"output":9.9,"cache_read":0.165,"cache_write":1.7413},"canonical_model_id":"google/gemini-3.5-flash"},"muse-glimmer-30b":{"id":"muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":20480},"cost":{"input":0,"output":0},"canonical_model_id":"meta/muse-glimmer-30b"},"kimi-k2.7-code@eu":{"id":"kimi-k2.7-code@eu","name":"Kimi K2.7 Code (EU)","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.25,"output":4.5,"cache_read":0.31},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"gpt-4.1@eu":{"id":"gpt-4.1@eu","name":"GPT-4.1 (EU)","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2.2,"output":8.8,"cache_read":0.55},"canonical_model_id":"openai/gpt-4.1"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"claude-opus-4-5@eu":{"id":"claude-opus-4-5@eu","name":"Claude Opus 4.5 (latest) (EU)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-5"},"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.5"},"claude-opus-4-7@eu":{"id":"claude-opus-4-7@eu","name":"Claude Opus 4.7 (EU)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-7"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.16,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"claude-sonnet-5@eu":{"id":"claude-sonnet-5@eu","name":"Claude Sonnet 5 (EU)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5"},"nemotron-3-ultra-550b-a55b":{"id":"nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"claude-fable-5@eu":{"id":"claude-fable-5@eu","name":"Claude Fable 5 (EU)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5"},"deepseek-v4-flash-0731@eu":{"id":"deepseek-v4-flash-0731@eu","name":"DeepSeek V4 Flash 0731 (EU)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.28,"output":0.56,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"gemini-3.8-flash@eu":{"id":"gemini-3.8-flash@eu","name":"Gemini 3.8 Flash (EU)","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.825,"output":4.125,"cache_read":0.0825},"canonical_model_id":"google/gemini-3.8-flash"},"seed-2.0-mini":{"id":"seed-2.0-mini","name":"Seed 2.0 Mini","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.1,"output":0.4,"cache_read":0.02},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"step-3.7-flash":{"id":"step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"mistral-medium-latest@eu":{"id":"mistral-medium-latest@eu","name":"Mistral Medium (latest) (EU)","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.44,"output":2.2,"cache_read":0.44},"canonical_model_id":"mistral/mistral-medium-latest"},"hy3":{"id":"hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.14,"output":0.58,"cache_read":0.035},"canonical_model_id":"tencent/hy3"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.1-flash-lite"},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.32,"output":1.28,"cache_read":0.032,"cache_write":0.4},"canonical_model_id":"alibaba/qwen3.7-plus"},"gemini-4-argon@eu":{"id":"gemini-4-argon@eu","name":"Gemini 4 Argon (EU)","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2.2,"output":11,"cache_read":0.11},"canonical_model_id":"google/gemini-4-argon"},"mistral-large-4@eu":{"id":"mistral-large-4@eu","name":"Mistral Large 4 (EU)","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":524288},"cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"glm-5.3-flash@eu":{"id":"glm-5.3-flash@eu","name":"GLM-5.3-Flash (EU)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":262144},"cost":{"input":0.2,"output":0.6,"cache_read":0.07},"canonical_model_id":"zhipuai/glm-5.3-flash"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"gpt-5.6-sol@eu":{"id":"gpt-5.6-sol@eu","name":"GPT-5.6 Sol (EU)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.44,"cache_write":5.5,"tiers":[{"input":8.8,"output":33,"cache_read":0.88,"cache_write":11,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8.8,"output":33,"cache_read":0.88,"cache_write":11}},"canonical_model_id":"openai/gpt-5.6-sol"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":4.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":9,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":9}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.2,"output":4.2,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"claude-haiku-5-5@eu":{"id":"claude-haiku-5-5@eu","name":"Claude Haiku 5.5 (EU)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.11,"output":0.55,"cache_read":0.02,"cache_write":0.14,"tiers":[{"input":0.55,"output":2.75,"cache_read":0.06,"cache_write":0.69,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"seed-2.0-code":{"id":"seed-2.0-code","name":"Seed 2.0 Code","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.5,"output":3,"cache_read":0.1},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"gemini-3.7-flash@eu":{"id":"gemini-3.7-flash@eu","name":"Gemini 3.7 Flash (EU)","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.825,"output":4.125,"cache_read":0.0825},"canonical_model_id":"google/gemini-3.7-flash"},"qwen3.5-27b":{"id":"qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.26,"output":2.6},"canonical_model_id":"alibaba/qwen3.5-27b"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.8-flash"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.75,"output":16.5,"cache_read":0.275},"canonical_model_id":"openai/gpt-5.4"},"gpt-4.1-nano@eu":{"id":"gpt-4.1-nano@eu","name":"GPT-4.1 nano (EU)","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.11,"output":0.44,"cache_read":0.0275},"canonical_model_id":"openai/gpt-4.1-nano"},"gemini-3-pro-image":{"id":"gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":1048576,"output":32768},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":4.5},"canonical_model_id":"google/gemini-3-pro-image"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"minimax-m3@eu":{"id":"minimax-m3@eu","name":"MiniMax-M3 (EU)","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.4,"output":2,"cache_read":0.1},"canonical_model_id":"minimax/MiniMax-M3"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":1.583},"canonical_model_id":"google/gemini-3.5-flash"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"nemotron-3-ultra-nvfp4":{"id":"nemotron-3-ultra-nvfp4","name":"nemotron-3-ultra-nvfp4","description":"Nemotron-3-Ultra-550B-A55B-NVFP4 is a frontier-scale large language model (LLM) trained by NVIDIA, designed to deliver strong agentic, reasoning, and conversational capabilities. It is optimized for the most demanding workloads, including complex multi-step agents, long-context analysis, and high-accuracy reasoning over code, math, and science. The model employs a hybrid Latent Mixture-of-Experts (LatentMoE) architecture, utilizing interleaved Mamba-2 and MoE layers, along with select Attention layers. Like the Super model, the Ultra model incorporates Multi-Token Prediction (MTP) layers for faster text generation and improved quality, and it is trained using an NVFP4 pre-training recipe to maximize compute efficiency. The model has 55B active parameters and 550B parameters in total.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":2.4,"cache_read":0.12}},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0027},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"gpt-5-nano@eu":{"id":"gpt-5-nano@eu","name":"GPT-5 Nano (EU)","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":0.055,"output":0.44,"cache_read":0.0055},"canonical_model_id":"openai/gpt-5-nano"},"qwen3.8-2.4T-A95B@eu":{"id":"qwen3.8-2.4T-A95B@eu","name":"Qwen3.8 2.4T A95B (EU)","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":262144},"cost":{"input":2.5,"output":6,"cache_read":0.63},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"claude-sonnet-4@eu":{"id":"claude-sonnet-4@eu","name":"Claude Sonnet 4 (latest) (EU)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-0"},"nemotron-3-nano-omni":{"id":"nemotron-3-nano-omni","name":"nemotron-3-nano-omni","description":"The most open, efficient, and accurate omni modal reasoning model for agentic AI.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-20","last_updated":"2026-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":300000},"cost":{"input":0.06,"output":0.24,"cache_read":0.06}},"mistral-medium-3-5":{"id":"mistral-medium-3-5","name":"Mistral Medium 3.5","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":1.65,"output":8.25,"cache_read":1.65},"canonical_model_id":"mistral/mistral-medium-3-5"},"glm-5.3@eu":{"id":"glm-5.3@eu","name":"GLM-5.3 (EU)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.2,"output":4.2,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"qwen3.5-35b-a3b":{"id":"qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.14,"output":1,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"google/gemini-4-argon"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"qwen3.8-flash-next":{"id":"qwen3.8-flash-next","name":"Qwen3.8 Flash Next","description":"Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.8-flash-next"},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"cache_write":2},"canonical_model_id":"xai/grok-4.7"},"gpt-5@eu":{"id":"gpt-5@eu","name":"GPT-5 (EU)","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.375,"output":11,"cache_read":0.1375},"canonical_model_id":"openai/gpt-5"},"claude-opus-5-5@eu":{"id":"claude-opus-5-5@eu","name":"Claude Opus 5.5 (EU)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.22,"cache_write":5.5},"canonical_model_id":"anthropic/claude-opus-5-5"},"claude-sonnet-4-5@eu":{"id":"claude-sonnet-4-5@eu","name":"Claude Sonnet 4.5 (latest) (EU)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3.3,"output":16.5,"cache_read":0.3,"cache_write":4.125,"tiers":[{"input":6.6,"output":24.75,"cache_read":0.6,"cache_write":8.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6.6,"output":24.75,"cache_read":0.6,"cache_write":8.25}},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"gemma-4-26b-a4b-it":{"id":"gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.07,"output":0.34,"cache_read":0.07},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.28,"output":0.56,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-flash"},"grok-build-0.1":{"id":"grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.1},"canonical_model_id":"xai/grok-build-0.1"},"muse-spark-1.3":{"id":"muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"mistral-small-2603":{"id":"mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.165,"output":0.66,"cache_read":0.165},"canonical_model_id":"mistral/mistral-small-2603"},"qwen3.8-flash-next@eu":{"id":"qwen3.8-flash-next@eu","name":"Qwen3.8 Flash Next (EU)","description":"Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.8-flash-next"},"gpt-5.1@eu":{"id":"gpt-5.1@eu","name":"GPT-5.1 (EU)","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.375,"output":11,"cache_read":0.1375},"canonical_model_id":"openai/gpt-5.1"},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"cache_write":2,"tiers":[{"input":4,"output":12,"cache_read":1,"cache_write":4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1,"cache_write":4}},"canonical_model_id":"xai/grok-4.5"},"gpt-6-sol@eu":{"id":"gpt-6-sol@eu","name":"GPT-6 Sol (EU)","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.4,"output":12,"cache_read":0.24,"cache_write":3,"tiers":[{"input":4.8,"output":18,"cache_read":0.48,"cache_write":6,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.8,"output":18,"cache_read":0.48,"cache_write":6}},"canonical_model_id":"openai/gpt-6-sol"},"devstral-latest":{"id":"devstral-latest","name":"devstral-latest","description":"An enterprise grade text model, that excels at using tools to explore codebases, editing multiple files and power software engineering agents.","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-05-27","last_updated":"2026-05-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.44,"output":2.2,"cache_read":0.44}},"minimax-m3.1-flash":{"id":"minimax-m3.1-flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"nvidia-nemotron-3-ultra":{"id":"nvidia-nemotron-3-ultra","name":"nvidia-nemotron-3-ultra","description":"NVIDIA Nemotron 3 Ultra is NVIDIA's strongest open-weights reasoning model, positioned near GPT-5.4 Mini (xhigh) and ahead of DeepSeek V4-Flash and Qwen3.5-397B-A17B.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":false,"structured_output":false,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.5,"output":2.5}},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":1,"output":2.7,"cache_read":0.05},"canonical_model_id":"stepfun/step-5-preview"},"mistral-small-2603@eu":{"id":"mistral-small-2603@eu","name":"Mistral Small 4 (EU)","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.165,"output":0.66,"cache_read":0.165},"canonical_model_id":"mistral/mistral-small-2603"},"gpt-5.4@eu":{"id":"gpt-5.4@eu","name":"GPT-5.4 (EU)","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.75,"output":16.5,"cache_read":0.275,"tiers":[{"input":5.5,"output":24.75,"cache_read":0.55,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5.5,"output":24.75,"cache_read":0.55}},"canonical_model_id":"openai/gpt-5.4"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"nemotron-3-nano-omni@eu":{"id":"nemotron-3-nano-omni@eu","name":"nemotron-3-nano-omni@eu","description":"The most open, efficient, and accurate omni modal reasoning model for agentic AI.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-20","last_updated":"2026-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":300000},"cost":{"input":0.06,"output":0.24,"cache_read":0.06}},"minimax-m2.7-highspeed":{"id":"minimax-m2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":1.2},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"nemotron-3-nano-omni-30b-a3b-reasoning":{"id":"nemotron-3-nano-omni-30b-a3b-reasoning","name":"Nemotron 3 Nano Omni 30B A3B Reasoning","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":20480},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"o4-mini@eu":{"id":"o4-mini@eu","name":"o4-mini (EU)","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.21,"output":4.84,"cache_read":0.3025},"canonical_model_id":"openai/o4-mini"},"laguna-xs.2":{"id":"laguna-xs.2","name":"Laguna XS.2","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-xs.2"},"nemotron-3-super-120b-a12b":{"id":"nemotron-3-super-120b-a12b","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"cache_write":2,"tiers":[{"input":4,"output":12,"cache_read":1,"cache_write":4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1,"cache_write":4}},"canonical_model_id":"xai/grok-4.6"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"glm-5.1@eu":{"id":"glm-5.1@eu","name":"GLM-5.1 (EU)","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":200000},"cost":{"input":1.4,"output":4.4,"cache_read":1.4},"canonical_model_id":"zhipuai/glm-5.1"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"muse-spark-open":{"id":"muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"mistral-large-4":{"id":"mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":524288},"cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.43,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"laguna-m.1":{"id":"laguna-m.1","name":"Laguna M.1","description":"Poolside's open-weight model for agentic coding and long-horizon work","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-m.1"},"gemma-4-31b-it":{"id":"gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-31b-it"},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.7-max"},"mistral-medium-latest":{"id":"mistral-medium-latest","name":"Mistral Medium (latest)","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.44,"output":2.2,"cache_read":0.44},"canonical_model_id":"mistral/mistral-medium-latest"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"kimi-k3@eu":{"id":"kimi-k3@eu","name":"Kimi K3 (EU)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":3,"output":15,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.13,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.63,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"gpt-6-luna@eu":{"id":"gpt-6-luna@eu","name":"GPT-6 Luna (EU)","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.12,"output":0.6,"cache_read":0.012,"cache_write":0.15,"tiers":[{"input":0.24,"output":0.9,"cache_read":0.024,"cache_write":0.3,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.24,"output":0.9,"cache_read":0.024,"cache_write":0.3}},"canonical_model_id":"openai/gpt-6-luna"},"gpt-5-mini@eu":{"id":"gpt-5-mini@eu","name":"GPT-5 Mini (EU)","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":0.275,"output":2.2,"cache_read":0.0275},"canonical_model_id":"openai/gpt-5-mini"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"muse-wonder-1.0":{"id":"muse-wonder-1.0","name":"Muse Wonder","description":"Muse Wonder is a frontier multimodal reasoning model from Meta for personal and professional agents, long-horizon coding, and computer use. It combines Active Memory, which carries relevant context across conversations, with Active Focus, which keeps it working through blockers toward a goal.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.8,"output":2.55,"cache_read":0.16},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"claude-sonnet-4-6@eu":{"id":"claude-sonnet-4-6@eu","name":"Claude Sonnet 4.6 (EU)","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3.3,"output":16.5,"cache_read":0.3,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"mistral-medium-3-5@eu":{"id":"mistral-medium-3-5@eu","name":"Mistral Medium 3.5 (EU)","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":1.65,"output":8.25,"cache_read":1.65},"canonical_model_id":"mistral/mistral-medium-3-5"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":1.5,"output":7,"cache_read":0.15},"canonical_model_id":"google/gemini-3.6-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-5.5-pro":{"id":"gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.5-pro"},"qwen3.5-2b":{"id":"qwen3.5-2b","name":"qwen3.5-2b","description":"Qwen3.5-2B is a compact yet capable model from Alibaba's Qwen3.5 series. It features a 262K token context window, support for 201 languages, thinking/reasoning mode, and tool calling for agentic workflows. A strong choice for prototyping, fine-tuning, and efficient multilingual deployments.","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-10","last_updated":"2026-03-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.02,"output":0.1}},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"kimi-k2.6@eu":{"id":"kimi-k2.6@eu","name":"Kimi K2.6 (EU)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":128000},"cost":{"input":0.95,"output":4,"cache_read":0.95},"canonical_model_id":"moonshotai/kimi-k2.6"},"fugu-ultra":{"id":"fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra"},"kat-coder-pro":{"id":"kat-coder-pro","name":"kat-coder-pro","description":"KAT-Coder-Pro V2 by KwaiKAT is a non-reasoning model optimized for agentic coding. It delivers strong performance on reasoning-style tasks while requiring significantly fewer output tokens than peer models. With the 1210 release, it achieved a score of 64 on the Artificial Analysis Intelligence Index, placing it in the global Top 10 and ranking first among all non-reasoning models.","family":"kat-coder","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-03-27","last_updated":"2026-03-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.3,"output":1.2}},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"gemini-3.1-flash-lite@eu":{"id":"gemini-3.1-flash-lite@eu","name":"Gemini 3.1 Flash Lite (EU)","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.275,"output":1.65,"cache_read":0.0275,"cache_write":0.091663},"canonical_model_id":"google/gemini-3.1-flash-lite"},"gemini-2.5-pro@eu":{"id":"gemini-2.5-pro@eu","name":"Gemini 2.5 Pro (EU)","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":1.25,"output":10,"cache_read":0.31,"cache_write":2.375,"tiers":[{"input":2.5,"output":15,"cache_read":0.62,"cache_write":4.75,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.62,"cache_write":4.75}},"canonical_model_id":"google/gemini-2.5-pro"},"claude-fable-5.1":{"id":"claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"grok-4.2-beta":{"id":"grok-4.2-beta","name":"grok-4.2-beta","description":"Grok 4.20 Beta is xAI's newest flagship model with industry-leading speed and agentic tool calling capabilities. It combines the lowest hallucination rate on the market with strict prompt adherance, delivering consistently precise and truthful responses.","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-19","last_updated":"2026-03-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":2,"output":6,"cache_read":0.2,"cache_write":2,"tiers":[{"input":4,"output":12,"cache_read":0.4,"cache_write":4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.4,"cache_write":4}}},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"}}},"gitlab":{"id":"gitlab","env":["GITLAB_TOKEN"],"npm":"gitlab-ai-provider","name":"GitLab Duo","doc":"https://docs.gitlab.com/user/duo_agent_platform/","models":{"duo-chat-gpt-5-mini":{"id":"duo-chat-gpt-5-mini","name":"Agentic Chat (GPT-5 Mini)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2026-01-22","last_updated":"2026-01-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0,"output":0}},"duo-chat-gpt-6-1-sol":{"id":"duo-chat-gpt-6-1-sol","name":"Agentic Chat (GPT-6.1 Sol)","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-6.1-sol"},"duo-chat-gpt-6-astra":{"id":"duo-chat-gpt-6-astra","name":"Agentic Chat (GPT-6 Astra)","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-6-astra"},"duo-chat-gpt-5-1":{"id":"duo-chat-gpt-5-1","name":"Agentic Chat (GPT-5.1)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2026-01-22","last_updated":"2026-01-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0,"output":0}},"duo-chat-opus-5-5":{"id":"duo-chat-opus-5-5","name":"Agentic Chat (Claude Opus 5.5)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-opus-5-5"},"duo-chat-fable-5-1":{"id":"duo-chat-fable-5-1","name":"Agentic Chat (Claude Fable 5.1)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-fable-5-1"},"duo-chat-opus-4-8":{"id":"duo-chat-opus-4-8","name":"Agentic Chat (Claude Opus 4.8)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"duo-chat-opus-5":{"id":"duo-chat-opus-5","name":"Agentic Chat (Claude Opus 5)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-opus-5"},"duo-chat-gpt-6-luna":{"id":"duo-chat-gpt-6-luna","name":"Agentic Chat (GPT-6 Luna)","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-6-luna"},"duo-chat-gpt-5-codex":{"id":"duo-chat-gpt-5-codex","name":"Agentic Chat (GPT-5 Codex)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2026-01-22","last_updated":"2026-01-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0,"output":0}},"duo-chat-gpt-6-sol":{"id":"duo-chat-gpt-6-sol","name":"Agentic Chat (GPT-6 Sol)","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-6-sol"},"duo-chat-opus-4-6":{"id":"duo-chat-opus-4-6","name":"Agentic Chat (Claude Opus 4.6)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"duo-chat-gpt-5-6-luna":{"id":"duo-chat-gpt-5-6-luna","name":"Agentic Chat (GPT-5.6 Luna)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-5.6-luna"},"duo-chat-gpt-5-4":{"id":"duo-chat-gpt-5-4","name":"Agentic Chat (GPT-5.4)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0}},"duo-chat-sonnet-4-6":{"id":"duo-chat-sonnet-4-6","name":"Agentic Chat (Claude Sonnet 4.6)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"duo-chat-gpt-5-5":{"id":"duo-chat-gpt-5-5","name":"Agentic Chat (GPT-5.5)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0}},"duo-chat-gpt-5-2-codex":{"id":"duo-chat-gpt-5-2-codex","name":"Agentic Chat (GPT-5.2 Codex)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-01-22","last_updated":"2026-01-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0,"output":0}},"duo-chat-gpt-5-6-sol":{"id":"duo-chat-gpt-5-6-sol","name":"Agentic Chat (GPT-5.6 Sol)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-5.6-sol"},"duo-chat-opus-4-7":{"id":"duo-chat-opus-4-7","name":"Agentic Chat (Claude Opus 4.7)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"duo-chat-gpt-5-4-nano":{"id":"duo-chat-gpt-5-4-nano","name":"Agentic Chat (GPT-5.4 Nano)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0,"output":0}},"duo-chat-gpt-5-2":{"id":"duo-chat-gpt-5-2","name":"Agentic Chat (GPT-5.2)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-01-23","last_updated":"2026-01-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0,"output":0}},"duo-chat-fable-5":{"id":"duo-chat-fable-5","name":"Agentic Chat (Claude Fable 5)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-fable-5"},"duo-chat-sonnet-4-5":{"id":"duo-chat-sonnet-4-5","name":"Agentic Chat (Claude Sonnet 4.5)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2026-01-08","last_updated":"2026-01-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"duo-chat-gpt-5-6-terra":{"id":"duo-chat-gpt-5-6-terra","name":"Agentic Chat (GPT-5.6 Terra)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-5.6-terra"},"duo-chat-gpt-5-3-codex":{"id":"duo-chat-gpt-5-3-codex","name":"Agentic Chat (GPT-5.3 Codex)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0,"output":0}},"duo-chat-sonnet-5-5":{"id":"duo-chat-sonnet-5-5","name":"Agentic Chat (Claude Sonnet 5.5)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"duo-chat-sonnet-5":{"id":"duo-chat-sonnet-5","name":"Agentic Chat (Claude Sonnet 5)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-sonnet-5"},"duo-chat-opus-4-5":{"id":"duo-chat-opus-4-5","name":"Agentic Chat (Claude Opus 4.5)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2026-01-08","last_updated":"2026-01-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"duo-chat-haiku-4-5":{"id":"duo-chat-haiku-4-5","name":"Agentic Chat (Claude Haiku 4.5)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2026-01-08","last_updated":"2026-01-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"duo-chat-gpt-5-4-mini":{"id":"duo-chat-gpt-5-4-mini","name":"Agentic Chat (GPT-5.4 Mini)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0,"output":0}}}},"aki-io":{"id":"aki-io","env":["AKI_IO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://aki.io/v1","name":"AKI.IO","doc":"https://aki.io/docs/","models":{"deepseek-v4.1-flash-552b":{"id":"deepseek-v4.1-flash-552b","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":81920},"cost":{"input":0.25,"output":1,"cache_read":0.1},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"gemma4-26b":{"id":"gemma4-26b","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"cost":{"input":0.1,"output":0.5},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"glm5.3-754b":{"id":"glm5.3-754b","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":81920},"cost":{"input":1,"output":3.5,"cache_read":0.25},"canonical_model_id":"zhipuai/glm-5.3"},"qwen3.6-35b":{"id":"qwen3.6-35b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"cost":{"input":0.15,"output":0.5},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"mistral4-119b":{"id":"mistral4-119b","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":81920},"cost":{"input":0.2,"output":0.6},"canonical_model_id":"mistral/mistral-small-2603"},"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.3,"output":2.2,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3.8-27b"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.15,"output":0.55},"canonical_model_id":"openai/gpt-oss-120b"}}},"stepfun-ai-step-plan":{"id":"stepfun-ai-step-plan","env":["STEPFUN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.stepfun.ai/step_plan/v1","name":"StepFun Step Plan (Global)","doc":"https://platform.stepfun.ai/docs/en/step-plan/integrations/reasoning-api","models":{"step-3.7-flash":{"id":"step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"canonical_model_id":"stepfun/step-3.7-flash"},"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"canonical_model_id":"stepfun/step-5-preview"},"step-3.5-flash-2603":{"id":"step-3.5-flash-2603","name":"Step 3.5 Flash 2603","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"canonical_model_id":"stepfun/step-3.5-flash-2603"},"step-3.5-flash":{"id":"step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"canonical_model_id":"stepfun/step-3.5-flash"}}},"thinkingmachines":{"id":"thinkingmachines","env":["TINKER_API_KEY"],"npm":"@ai-sdk/anthropic","api":"https://tinker.thinkingmachines.dev/services/tinker-prod/anthropic/api/v1","name":"Thinking Machines","doc":"https://tinker-docs.thinkingmachines.ai/tinker/compatible-apis/anthropic/","models":{"thinkingmachines/Inkling":{"id":"thinkingmachines/Inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":1.87,"output":4.68,"cache_read":0.374},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/Inkling:peft:262144":{"id":"thinkingmachines/Inkling:peft:262144","name":"Inkling (256K)","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":3.74,"output":9.36,"cache_read":0.748},"canonical_model_id":"thinkingmachines/inkling"}}},"neuralwatt":{"id":"neuralwatt","env":["NEURALWATT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.neuralwatt.com/v1","name":"Neuralwatt","doc":"https://portal.neuralwatt.com/docs","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"kimi-k3-fast":{"id":"kimi-k3-fast","name":"Kimi K3 Fast","description":"Kimi K3 with thinking disabled for low-latency tool calling, vision, and JSON work","family":"kimi-k3","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"kimi-k3-flex":{"id":"kimi-k3-flex","name":"Kimi K3 Flex","description":"Kimi K3 on the flex tier: discounted, best-effort latency, requests may be held under load","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"cost":{"input":1.95,"output":9.75,"cache_read":0.195},"canonical_model_id":"moonshotai/kimi-k3"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"status":"deprecated","cost":{"input":1,"output":3,"cache_read":0.1},"canonical_model_id":"deepseek/deepseek-v4-pro"},"qwen3.6-35b-flex":{"id":"qwen3.6-35b-flex","name":"Qwen3.6 35B Flex","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131056,"output":131056},"cost":{"input":0.1885,"output":0.7475,"cache_read":0.01885},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"glm-5.2-short-fast":{"id":"glm-5.2-short-fast","name":"GLM 5.2 Short Fast","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-06-17","last_updated":"2026-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":199984,"output":32000},"status":"deprecated","cost":{"input":1.45,"output":4.5,"cache_read":0.145},"canonical_model_id":"zhipuai/glm-5.2"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"glm-5.2-fast":{"id":"glm-5.2-fast","name":"GLM 5.2 Fast","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-06-17","last_updated":"2026-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"status":"deprecated","cost":{"input":1.45,"output":4.5,"cache_read":0.145},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5.2-flex":{"id":"glm-5.2-flex","name":"GLM 5.2 Flex","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-06-17","last_updated":"2026-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"status":"deprecated","cost":{"input":0.9425,"output":2.925,"cache_read":0.09425},"canonical_model_id":"zhipuai/glm-5.2"},"kimi-k2.7-code-flex":{"id":"kimi-k2.7-code-flex","name":"Kimi K2.7 Code Flex","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262128,"output":262128},"cost":{"input":0.6175,"output":2.6,"cache_read":0.06175},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"gemma-4-31b":{"id":"gemma-4-31b","name":"Gemma 4 31B","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262128,"output":16384},"cost":{"input":0.144,"output":0.42,"cache_read":0.0144},"canonical_model_id":"google/gemma-4-31b-it"},"deepseek-v4-flash-flex":{"id":"deepseek-v4-flash-flex","name":"DeepSeek V4 Flash Flex","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"cost":{"input":0.091,"output":0.182,"cache_read":0.0182},"canonical_model_id":"deepseek/deepseek-v4-flash"},"qwen3.6-35b":{"id":"qwen3.6-35b","name":"Qwen3.6 35B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131056,"output":131056},"cost":{"input":0.29,"output":1.15,"cache_read":0.029},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"glm-5.3":{"id":"glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"cost":{"input":1.45,"output":4.5,"cache_read":0.145},"canonical_model_id":"zhipuai/glm-5.3"},"qwen-3.8-27b":{"id":"qwen-3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262128,"output":131072},"cost":{"input":0.45,"output":3.2,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-27b"},"kimi-k2.7-code-fast":{"id":"kimi-k2.7-code-fast","name":"Kimi K2.7 Code Fast","description":"Kimi K2.7 Code with reasoning capped to a short budget for lower latency; reasoning cannot be disabled on this model","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262128,"output":262128},"cost":{"input":0.95,"output":4,"cache_read":0.095},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"qwen-3.8-27b-flex":{"id":"qwen-3.8-27b-flex","name":"Qwen3.8 27B Flex","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262128,"output":131072},"cost":{"input":0.2925,"output":2.08,"cache_read":0.1625},"canonical_model_id":"alibaba/qwen3.8-27b"},"glm-5.3-flex":{"id":"glm-5.3-flex","name":"GLM 5.3 Flex","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"cost":{"input":0.9425,"output":2.925,"cache_read":0.09425},"canonical_model_id":"zhipuai/glm-5.3"},"deepseek-v4.1-flash-flex":{"id":"deepseek-v4.1-flash-flex","name":"DeepSeek V4.1 Flash Flex","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"cost":{"input":0.0975,"output":0.39,"cache_read":0.00975},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"glm-5.2-short":{"id":"glm-5.2-short","name":"GLM 5.2 Short","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-06-17","last_updated":"2026-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":199984,"output":32000},"status":"deprecated","cost":{"input":1.45,"output":4.5,"cache_read":0.145},"canonical_model_id":"zhipuai/glm-5.2"},"mimo-v2.6-pro-flex":{"id":"mimo-v2.6-pro-flex","name":"MiMo-V2.6-Pro Flex","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"cost":{"input":0.5655,"output":1.131,"cache_read":0.0234},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262128,"output":262128},"cost":{"input":0.95,"output":4,"cache_read":0.095},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"deepseek-v4-flash-speed":{"id":"deepseek-v4-flash-speed","name":"DeepSeek V4 Flash (Speed)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"qwen3.6-35b-fast":{"id":"qwen3.6-35b-fast","name":"Qwen3.6 35B Fast","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"qwen3.6","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131056,"output":131056},"cost":{"input":0.29,"output":1.15,"cache_read":0.029},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"cost":{"input":0.87,"output":1.74,"cache_read":0.036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"glm-5.2-short-flex":{"id":"glm-5.2-short-flex","name":"GLM 5.2 Short Flex","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-06-17","last_updated":"2026-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":199984,"output":32000},"status":"deprecated","cost":{"input":0.9425,"output":2.925,"cache_read":0.09425},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5.2":{"id":"glm-5.2","name":"GLM 5.2","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-06-17","last_updated":"2026-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"status":"deprecated","cost":{"input":1.45,"output":4.5,"cache_read":0.145},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"glm-5.2-short-fast-flex":{"id":"glm-5.2-short-fast-flex","name":"GLM 5.2 Short Fast Flex","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-06-17","last_updated":"2026-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":199984,"output":32000},"status":"deprecated","cost":{"input":0.9425,"output":2.925,"cache_read":0.09425},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5.3-flash-flex":{"id":"glm-5.3-flash-flex","name":"GLM-5.3 Flash Flex","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"cost":{"input":0.0975,"output":0.325,"cache_read":0.0195},"canonical_model_id":"zhipuai/glm-5.3-flash"}}},"kuae-cloud-coding-plan":{"id":"kuae-cloud-coding-plan","env":["KUAE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://coding-plan-endpoint.kuaecloud.net/v1","name":"KUAE Cloud Coding Plan","doc":"https://docs.mthreads.com/kuaecloud/kuaecloud-doc-online/coding_plan/","models":{"GLM-4.7":{"id":"GLM-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}}}},"amazon-bedrock":{"id":"amazon-bedrock","env":["AWS_ACCESS_KEY_ID","AWS_SECRET_ACCESS_KEY","AWS_REGION","AWS_BEARER_TOKEN_BEDROCK"],"npm":"@ai-sdk/amazon-bedrock","name":"Amazon Bedrock","doc":"https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html","models":{"openai.gpt-oss-safeguard-120b":{"id":"openai.gpt-oss-safeguard-120b","name":"GPT OSS Safeguard 120B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-safeguard-120b"},"global.anthropic.claude-fable-5-1":{"id":"global.anthropic.claude-fable-5-1","name":"Claude Fable 5.1 (Global)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"us.amazon.nova-premier-v1:0":{"id":"us.amazon.nova-premier-v1:0","name":"Nova Premier (US)","description":"Multimodal model for complex analysis, long-context understanding, tool use, and model distillation","family":"nova","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2025-04-30","last_updated":"2025-04-30","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":10000},"status":"deprecated","cost":{"input":2.5,"output":12.5,"cache_read":0.625,"cache_write":2.5},"canonical_model_id":"amazon/nova-premier"},"global.openai.gpt-6-astra":{"id":"global.openai.gpt-6-astra","name":"GPT-6 Astra (Global)","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"ultrafast":{"cost":{"input":60,"output":300,"cache_read":6,"cache_write":75},"provider":{"body":{"additionalModelRequestFields":{"service_tier":"ultrafast"}}}}}},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"jp.anthropic.claude-opus-4-8":{"id":"jp.anthropic.claude-opus-4-8","name":"Claude Opus 4.8 (JP)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-8"},"zai.glm-4.7":{"id":"zai.glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"status":"deprecated","cost":{"input":0.6,"output":2.2},"canonical_model_id":"zhipuai/glm-4.7"},"us.anthropic.claude-opus-5-5":{"id":"us.anthropic.claude-opus-5-5","name":"Claude Opus 5.5 (US)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.22,"cache_write":5.5},"canonical_model_id":"anthropic/claude-opus-5-5"},"us.openai.gpt-5.6-terra":{"id":"us.openai.gpt-5.6-terra","name":"GPT-5.6 Terra (US)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.2,"output":13.2,"cache_read":0.22,"cache_write":2.75,"tiers":[{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5}},"canonical_model_id":"openai/gpt-5.6-terra"},"mistral.ministral-3-8b-instruct":{"id":"mistral.ministral-3-8b-instruct","name":"Ministral 3 8B","description":"Compact open vision-language model for edge deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.15,"output":0.15},"canonical_model_id":"mistral/ministral-3-8b-instruct-2512"},"us.anthropic.claude-fable-5":{"id":"us.anthropic.claude-fable-5","name":"Claude Fable 5 (US)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5"},"google.gemma-4-e2b":{"id":"google.gemma-4-e2b","name":"Gemma 4 E2B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":0.04,"output":0.08},"canonical_model_id":"google/gemma-4-E2B-it"},"global.anthropic.claude-opus-4-8":{"id":"global.anthropic.claude-opus-4-8","name":"Claude Opus 4.8 (Global)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"us.amazon.nova-2-lite-v1:0":{"id":"us.amazon.nova-2-lite-v1:0","name":"Nova 2 Lite (US)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.33,"output":2.75,"cache_read":0.0825,"cache_write":0.33},"canonical_model_id":"amazon/nova-2-lite"},"global.zai.glm-5.3":{"id":"global.zai.glm-5.3","name":"GLM-5.3 (Global)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.68,"output":5.28,"cache_read":0.312,"cache_write":2.1},"canonical_model_id":"zhipuai/glm-5.3"},"global.moonshotai.kimi-k3":{"id":"global.moonshotai.kimi-k3","name":"Kimi K3 (Global)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"moonshotai/kimi-k3"},"meta.llama4-maverick-17b-instruct-v1:0":{"id":"meta.llama4-maverick-17b-instruct-v1:0","name":"Llama 4 Maverick 17B Instruct","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":8192},"status":"deprecated","cost":{"input":0.24,"output":0.97},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"global.anthropic.claude-haiku-4-5-20251001-v1:0":{"id":"global.anthropic.claude-haiku-4-5-20251001-v1:0","name":"Claude Haiku 4.5 (Global)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"us.anthropic.claude-haiku-4-5-20251001-v1:0":{"id":"us.anthropic.claude-haiku-4-5-20251001-v1:0","name":"Claude Haiku 4.5 (US)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"eu.amazon.nova-pro-v1:0":{"id":"eu.amazon.nova-pro-v1:0","name":"Nova Pro (EU)","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.92,"output":3.68,"cache_read":0.23,"cache_write":0.92},"canonical_model_id":"amazon/nova-pro"},"us.meta.llama4-scout-17b-instruct-v1:0":{"id":"us.meta.llama4-scout-17b-instruct-v1:0","name":"Llama 4 Scout 17B Instruct (US)","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":10000000,"output":8192},"status":"deprecated","cost":{"input":0.17,"output":0.66},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"minimax.minimax-m2.5":{"id":"minimax.minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":98304},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.5"},"global.anthropic.claude-sonnet-4-5-20250929-v1:0":{"id":"global.anthropic.claude-sonnet-4-5-20250929-v1:0","name":"Claude Sonnet 4.5 (Global)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"eu.anthropic.claude-opus-4-7":{"id":"eu.anthropic.claude-opus-4-7","name":"Claude Opus 4.7 (EU)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-7"},"google.gemma-4-26b-a4b":{"id":"google.gemma-4-26b-a4b","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":0.13,"output":0.4},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"openai.gpt-6-astra":{"id":"openai.gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"ultrafast":{"cost":{"input":66,"output":330,"cache_read":6.6,"cache_write":82.5},"provider":{"body":{"service_tier":"ultrafast"}}}}},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75,"tiers":[{"input":22,"output":82.5,"cache_read":2.2,"cache_write":27.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":22,"output":82.5,"cache_read":2.2,"cache_write":27.5}},"canonical_model_id":"openai/gpt-6-astra"},"openai.gpt-6.1-sol":{"id":"openai.gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"ultrafast":{"cost":{"input":13.2,"output":66,"cache_read":0.66,"cache_write":16.5},"provider":{"body":{"service_tier":"ultrafast"}}}}},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":2.2,"output":11,"cache_read":0.11,"cache_write":2.75,"tiers":[{"input":4.4,"output":16.5,"cache_read":0.22,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":16.5,"cache_read":0.22,"cache_write":5.5}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai.gpt-5.5":{"id":"openai.gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-06-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":5.5,"output":33,"cache_read":0.55},"canonical_model_id":"openai/gpt-5.5"},"global.openai.gpt-5.6-sol":{"id":"global.openai.gpt-5.6-sol","name":"GPT-5.6 Sol (Global)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"us.anthropic.claude-sonnet-5":{"id":"us.anthropic.claude-sonnet-5","name":"Claude Sonnet 5 (US)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5"},"us.meta.llama3-1-70b-instruct-v1:0":{"id":"us.meta.llama3-1-70b-instruct-v1:0","name":"Llama 3.1 70B Instruct (US)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-3.1-70b-instruct"},"mistral.magistral-small-2509":{"id":"mistral.magistral-small-2509","name":"Magistral Small 1.2","description":"Open multimodal reasoning model for transparent analysis of text and images","family":"magistral","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-18","last_updated":"2025-09-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":40000},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/magistral-small-2509"},"jp.anthropic.claude-sonnet-5":{"id":"jp.anthropic.claude-sonnet-5","name":"Claude Sonnet 5 (JP)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5"},"openai.gpt-6-luna":{"id":"openai.gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":0.11,"output":0.55,"cache_read":0.011,"cache_write":0.1375,"tiers":[{"input":0.22,"output":0.825,"cache_read":0.022,"cache_write":0.275,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.22,"output":0.825,"cache_read":0.022,"cache_write":0.275}},"canonical_model_id":"openai/gpt-6-luna"},"openai.gpt-oss-safeguard-20b":{"id":"openai.gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.07,"output":0.2},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"us.openai.gpt-6-sol":{"id":"us.openai.gpt-6-sol","name":"GPT-6 Sol (US)","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75,"tiers":[{"input":4.4,"output":16.5,"cache_read":0.44,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":16.5,"cache_read":0.44,"cache_write":5.5}},"canonical_model_id":"openai/gpt-6-sol"},"mistral.voxtral-mini-3b-2507":{"id":"mistral.voxtral-mini-3b-2507","name":"Voxtral Mini 3B 2507","description":"Open audio-language model for speech transcription, audio understanding, and voice-driven tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":4096},"status":"deprecated","cost":{"input":0.04,"output":0.04},"canonical_model_id":"mistral/voxtral-mini-3b-2507"},"mistral.voxtral-small-24b-2507":{"id":"mistral.voxtral-small-24b-2507","name":"Voxtral Small 24B 2507","description":"Open audio-language model for speech transcription, audio understanding, and voice-driven tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"status":"deprecated","cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/voxtral-small-24b-2507"},"eu.amazon.nova-micro-v1:0":{"id":"eu.amazon.nova-micro-v1:0","name":"Nova Micro (EU)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.04,"output":0.16,"cache_read":0.01,"cache_write":0.04},"canonical_model_id":"amazon/nova-micro"},"jp.amazon.nova-2-lite-v1:0":{"id":"jp.amazon.nova-2-lite-v1:0","name":"Nova 2 Lite (JP)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.396,"output":3.311,"cache_read":0.099,"cache_write":0.396},"canonical_model_id":"amazon/nova-2-lite"},"us.amazon.nova-micro-v1:0":{"id":"us.amazon.nova-micro-v1:0","name":"Nova Micro (US)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.035,"output":0.14,"cache_read":0.00875,"cache_write":0.035},"canonical_model_id":"amazon/nova-micro"},"qwen.qwen3-coder-480b-a35b-v1:0":{"id":"qwen.qwen3-coder-480b-a35b-v1:0","name":"Qwen3-Coder 480B-A35B Instruct","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"status":"deprecated","cost":{"input":0.45,"output":1.8},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"global.anthropic.claude-opus-4-6-v1":{"id":"global.anthropic.claude-opus-4-6-v1","name":"Claude Opus 4.6 (Global)","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"au.anthropic.claude-opus-5-5":{"id":"au.anthropic.claude-opus-5-5","name":"Claude Opus 5.5 (AU)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.22,"cache_write":5.5},"canonical_model_id":"anthropic/claude-opus-5-5"},"meta.llama3-1-70b-instruct-v1:0":{"id":"meta.llama3-1-70b-instruct-v1:0","name":"Llama 3.1 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-3.1-70b-instruct"},"au.anthropic.claude-sonnet-4-6":{"id":"au.anthropic.claude-sonnet-4-6","name":"AU Anthropic Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic.claude-sonnet-5":{"id":"anthropic.claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"qwen.qwen3-coder-30b-a3b-v1:0":{"id":"qwen.qwen3-coder-30b-a3b-v1:0","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"status":"deprecated","cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"us.anthropic.claude-sonnet-4-20250514-v1:0":{"id":"us.anthropic.claude-sonnet-4-20250514-v1:0","name":"Claude Sonnet 4 (US)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"us.openai.gpt-6-astra":{"id":"us.openai.gpt-6-astra","name":"GPT-6 Astra (US)","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"ultrafast":{"cost":{"input":66,"output":330,"cache_read":6.6,"cache_write":82.5},"provider":{"body":{"additionalModelRequestFields":{"service_tier":"ultrafast"}}}}}},"cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75,"tiers":[{"input":22,"output":82.5,"cache_read":2.2,"cache_write":27.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":22,"output":82.5,"cache_read":2.2,"cache_write":27.5}},"canonical_model_id":"openai/gpt-6-astra"},"meta.llama3-1-8b-instruct-v1:0":{"id":"meta.llama3-1-8b-instruct-v1:0","name":"Llama 3.1 8B Instruct","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.22,"output":0.22},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"us.anthropic.claude-opus-4-6-v1":{"id":"us.anthropic.claude-opus-4-6-v1","name":"Claude Opus 4.6 (US)","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-6"},"minimax.minimax-m2":{"id":"minimax.minimax-m2","name":"MiniMax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204608,"output":128000},"status":"deprecated","cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2"},"deepseek.v3-v1:0":{"id":"deepseek.v3-v1:0","name":"DeepSeek-V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":81920},"status":"deprecated","cost":{"input":0.58,"output":1.68},"canonical_model_id":"deepseek/deepseek-v3.1"},"anthropic.claude-opus-5-5":{"id":"anthropic.claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"apac.anthropic.claude-sonnet-4-20250514-v1:0":{"id":"apac.anthropic.claude-sonnet-4-20250514-v1:0","name":"Claude Sonnet 4 (APAC)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"eu.anthropic.claude-opus-5-5":{"id":"eu.anthropic.claude-opus-5-5","name":"Claude Opus 5.5 (EU)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.22,"cache_write":5.5},"canonical_model_id":"anthropic/claude-opus-5-5"},"global.openai.gpt-6-luna":{"id":"global.openai.gpt-6-luna","name":"GPT-6 Luna (Global)","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"eu.anthropic.claude-fable-5":{"id":"eu.anthropic.claude-fable-5","name":"Claude Fable 5 (EU)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5"},"google.gemma-4-31b":{"id":"google.gemma-4-31b","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"global.openai.gpt-6.1-sol":{"id":"global.openai.gpt-6.1-sol","name":"GPT-6.1 Sol (Global)","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"ultrafast":{"cost":{"input":12,"output":60,"cache_read":0.6,"cache_write":15},"provider":{"body":{"additionalModelRequestFields":{"service_tier":"ultrafast"}}}}}},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"us.writer.palmyra-x4-v1:0":{"id":"us.writer.palmyra-x4-v1:0","name":"Palmyra X4 (US)","description":"Enterprise language model for workflow automation, coding, data analysis, and tool use","family":"palmyra","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2024-10-09","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":122880,"output":8192},"status":"deprecated","cost":{"input":2.5,"output":10},"canonical_model_id":"writer/palmyra-x4"},"openai.gpt-oss-20b-1:0":{"id":"openai.gpt-oss-20b-1:0","name":"gpt-oss-20b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":128000},"cost":{"input":0.07,"output":0.3},"canonical_model_id":"openai/gpt-oss-20b"},"in.openai.gpt-5.6-terra":{"id":"in.openai.gpt-5.6-terra","name":"GPT-5.6 Terra (India)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.2,"output":13.2,"cache_read":0.22,"cache_write":2.75,"tiers":[{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5}},"canonical_model_id":"openai/gpt-5.6-terra"},"au.anthropic.claude-opus-4-6-v1":{"id":"au.anthropic.claude-opus-4-6-v1","name":"AU Anthropic Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-6"},"global.xai.grok-4.6":{"id":"global.xai.grok-4.6","name":"Grok 4.6 (Global)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"global.anthropic.claude-haiku-5-5":{"id":"global.anthropic.claude-haiku-5-5","name":"Claude Haiku 5.5 (Global)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"us.anthropic.claude-opus-5":{"id":"us.anthropic.claude-opus-5","name":"Claude Opus 5 (US)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-5"},"mistral.pixtral-large-2502-v1:0":{"id":"mistral.pixtral-large-2502-v1:0","name":"Pixtral Large (25.02)","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-08","last_updated":"2025-04-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/pixtral-large-2502"},"us.zai.glm-5.3":{"id":"us.zai.glm-5.3","name":"GLM-5.3 (US)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.848,"output":5.808,"cache_read":0.3432,"cache_write":2.31},"canonical_model_id":"zhipuai/glm-5.3"},"zai.glm-4.7-flash":{"id":"zai.glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"status":"deprecated","cost":{"input":0.07,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"us.xai.grok-4.7":{"id":"us.xai.grok-4.7","name":"Grok 4.7 (US)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2.2,"output":6.6,"cache_read":0.55},"canonical_model_id":"xai/grok-4.7"},"eu.anthropic.claude-haiku-4-5-20251001-v1:0":{"id":"eu.anthropic.claude-haiku-4-5-20251001-v1:0","name":"Claude Haiku 4.5 (EU)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"openai.gpt-oss-120b":{"id":"openai.gpt-oss-120b","name":"gpt-oss-120b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/v1","shape":"responses"},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"us.anthropic.claude-sonnet-4-5-20250929-v1:0":{"id":"us.anthropic.claude-sonnet-4-5-20250929-v1:0","name":"Claude Sonnet 4.5 (US)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"us.mistral.pixtral-large-2502-v1:0":{"id":"us.mistral.pixtral-large-2502-v1:0","name":"Pixtral Large (25.02) (US)","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-08","last_updated":"2025-04-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/pixtral-large-2502"},"eu.anthropic.claude-sonnet-4-20250514-v1:0":{"id":"eu.anthropic.claude-sonnet-4-20250514-v1:0","name":"Claude Sonnet 4 (EU)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"amazon.nova-micro-v1:0":{"id":"amazon.nova-micro-v1:0","name":"Nova Micro","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.035,"output":0.14,"cache_read":0.00875,"cache_write":0.035},"canonical_model_id":"amazon/nova-micro"},"eu.mistral.pixtral-large-2502-v1:0":{"id":"eu.mistral.pixtral-large-2502-v1:0","name":"Pixtral Large (25.02) (EU)","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-08","last_updated":"2025-04-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/pixtral-large-2502"},"in.anthropic.claude-sonnet-5":{"id":"in.anthropic.claude-sonnet-5","name":"Claude Sonnet 5 (India)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic.claude-opus-5":{"id":"anthropic.claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"us.anthropic.claude-opus-4-1-20250805-v1:0":{"id":"us.anthropic.claude-opus-4-1-20250805-v1:0","name":"Claude Opus 4.1 (US)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"us.openai.gpt-5.6-sol":{"id":"us.openai.gpt-5.6-sol","name":"GPT-5.6 Sol (US)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.44,"cache_write":5.5,"tiers":[{"input":8.8,"output":33,"cache_read":0.88,"cache_write":11,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8.8,"output":33,"cache_read":0.88,"cache_write":11}},"canonical_model_id":"openai/gpt-5.6-sol"},"global.anthropic.claude-opus-4-5-20251101-v1:0":{"id":"global.anthropic.claude-opus-4-5-20251101-v1:0","name":"Claude Opus 4.5 (Global)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"amazon.nova-2-lite-v1:0":{"id":"amazon.nova-2-lite-v1:0","name":"Nova 2 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.33,"output":2.75,"cache_read":0.0825,"cache_write":0.33},"canonical_model_id":"amazon/nova-2-lite"},"us.anthropic.claude-fable-5-1":{"id":"us.anthropic.claude-fable-5-1","name":"Claude Fable 5.1 (US)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":0.275,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5-1"},"qwen.qwen3-next-80b-a3b":{"id":"qwen.qwen3-next-80b-a3b","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-11","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262000},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"global.anthropic.claude-sonnet-5":{"id":"global.anthropic.claude-sonnet-5","name":"Claude Sonnet 5 (Global)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"us.amazon.nova-pro-v1:0":{"id":"us.amazon.nova-pro-v1:0","name":"Nova Pro (US)","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.8,"output":3.2,"cache_read":0.2,"cache_write":0.8},"canonical_model_id":"amazon/nova-pro"},"anthropic.claude-haiku-4-5-20251001-v1:0":{"id":"anthropic.claude-haiku-4-5-20251001-v1:0","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"anthropic.claude-haiku-5-5":{"id":"anthropic.claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"jp.anthropic.claude-haiku-4-5-20251001-v1:0":{"id":"jp.anthropic.claude-haiku-4-5-20251001-v1:0","name":"Claude Haiku 4.5 (JP)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"us.meta.llama4-maverick-17b-instruct-v1:0":{"id":"us.meta.llama4-maverick-17b-instruct-v1:0","name":"Llama 4 Maverick 17B Instruct (US)","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":8192},"status":"deprecated","cost":{"input":0.24,"output":0.97},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"global.anthropic.claude-sonnet-4-6":{"id":"global.anthropic.claude-sonnet-4-6","name":"Claude Sonnet 4.6 (Global)","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic.claude-opus-4-8":{"id":"anthropic.claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic.claude-sonnet-4-5-20250929-v1:0":{"id":"anthropic.claude-sonnet-4-5-20250929-v1:0","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"in.anthropic.claude-haiku-4-5-20251001-v1:0":{"id":"in.anthropic.claude-haiku-4-5-20251001-v1:0","name":"Claude Haiku 4.5 (India)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"meta.llama4-scout-17b-instruct-v1:0":{"id":"meta.llama4-scout-17b-instruct-v1:0","name":"Llama 4 Scout 17B Instruct","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":10000000,"output":8192},"status":"deprecated","cost":{"input":0.17,"output":0.66},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"global.anthropic.claude-opus-5":{"id":"global.anthropic.claude-opus-5","name":"Claude Opus 5 (Global)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"global.amazon.nova-2-lite-v1:0":{"id":"global.amazon.nova-2-lite-v1:0","name":"Nova 2 Lite (Global)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.3,"output":2.5,"cache_read":0.075,"cache_write":0.3},"canonical_model_id":"amazon/nova-2-lite"},"anthropic.claude-fable-5":{"id":"anthropic.claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"au.anthropic.claude-opus-5":{"id":"au.anthropic.claude-opus-5","name":"Claude Opus 5 (AU)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-5"},"openai.gpt-5.6-luna":{"id":"openai.gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":0.22,"output":1.32,"cache_read":0.022,"cache_write":0.275,"tiers":[{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55}},"canonical_model_id":"openai/gpt-5.6-luna"},"eu.anthropic.claude-opus-4-5-20251101-v1:0":{"id":"eu.anthropic.claude-opus-4-5-20251101-v1:0","name":"Claude Opus 4.5 (EU)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"au.anthropic.claude-opus-4-7":{"id":"au.anthropic.claude-opus-4-7","name":"Claude Opus 4.7 (AU)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-7"},"ca.amazon.nova-lite-v1:0":{"id":"ca.amazon.nova-lite-v1:0","name":"Nova Lite (CA)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.064,"output":0.256,"cache_read":0.016,"cache_write":0.064},"canonical_model_id":"amazon/nova-lite"},"google.gemma-3-12b-it":{"id":"google.gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.09,"output":0.29},"canonical_model_id":"google/gemma-3-12b-it"},"deepseek.v3.2":{"id":"deepseek.v3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2026-02-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":81920},"cost":{"input":0.62,"output":1.85},"canonical_model_id":"deepseek/deepseek-v3.2"},"eu.anthropic.claude-sonnet-4-6":{"id":"eu.anthropic.claude-sonnet-4-6","name":"Claude Sonnet 4.6 (EU)","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"us.meta.llama3-3-70b-instruct-v1:0":{"id":"us.meta.llama3-3-70b-instruct-v1:0","name":"Llama 3.3 70B Instruct (US)","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"zai.glm-5":{"id":"zai.glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1,"output":3.2},"canonical_model_id":"zhipuai/glm-5"},"eu.amazon.nova-lite-v1:0":{"id":"eu.amazon.nova-lite-v1:0","name":"Nova Lite (EU)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.069,"output":0.276,"cache_read":0.01725,"cache_write":0.069},"canonical_model_id":"amazon/nova-lite"},"jp.anthropic.claude-opus-5":{"id":"jp.anthropic.claude-opus-5","name":"Claude Opus 5 (JP)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-5"},"openai.gpt-oss-120b-1:0":{"id":"openai.gpt-oss-120b-1:0","name":"gpt-oss-120b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":128000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"au.anthropic.claude-sonnet-5":{"id":"au.anthropic.claude-sonnet-5","name":"Claude Sonnet 5 (AU)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5"},"openai.gpt-5.6-terra":{"id":"openai.gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":2.2,"output":13.2,"cache_read":0.22,"cache_write":2.75,"tiers":[{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5}},"canonical_model_id":"openai/gpt-5.6-terra"},"au.anthropic.claude-haiku-5-5":{"id":"au.anthropic.claude-haiku-5-5","name":"Claude Haiku 5.5 (AU)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.11,"output":0.55,"cache_read":0.011,"cache_write":0.1375,"tiers":[{"input":0.55,"output":2.75,"cache_read":0.055,"cache_write":0.6875,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"us.anthropic.claude-opus-4-7":{"id":"us.anthropic.claude-opus-4-7","name":"Claude Opus 4.7 (US)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-7"},"au.anthropic.claude-haiku-4-5-20251001-v1:0":{"id":"au.anthropic.claude-haiku-4-5-20251001-v1:0","name":"Claude Haiku 4.5 (AU)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"au.anthropic.claude-opus-4-8":{"id":"au.anthropic.claude-opus-4-8","name":"Claude Opus 4.8 (AU)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-8"},"eu.amazon.nova-2-lite-v1:0":{"id":"eu.amazon.nova-2-lite-v1:0","name":"Nova 2 Lite (EU)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.374,"output":3.157,"cache_read":0.0935,"cache_write":0.374},"canonical_model_id":"amazon/nova-2-lite"},"eu.anthropic.claude-haiku-5-5":{"id":"eu.anthropic.claude-haiku-5-5","name":"Claude Haiku 5.5 (EU)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.11,"output":0.55,"cache_read":0.011,"cache_write":0.1375,"tiers":[{"input":0.55,"output":2.75,"cache_read":0.055,"cache_write":0.6875,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"us.amazon.nova-lite-v1:0":{"id":"us.amazon.nova-lite-v1:0","name":"Nova Lite (US)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.06,"output":0.24,"cache_read":0.015,"cache_write":0.06},"canonical_model_id":"amazon/nova-lite"},"openai.gpt-6-sol":{"id":"openai.gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75,"tiers":[{"input":4.4,"output":16.5,"cache_read":0.44,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":16.5,"cache_read":0.44,"cache_write":5.5}},"canonical_model_id":"openai/gpt-6-sol"},"eu.anthropic.claude-opus-5":{"id":"eu.anthropic.claude-opus-5","name":"Claude Opus 5 (EU)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-5"},"eu.anthropic.claude-sonnet-4-5-20250929-v1:0":{"id":"eu.anthropic.claude-sonnet-4-5-20250929-v1:0","name":"Claude Sonnet 4.5 (EU)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"us.deepseek.r1-v1:0":{"id":"us.deepseek.r1-v1:0","name":"DeepSeek-R1 (US)","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"status":"deprecated","cost":{"input":1.35,"output":5.4},"canonical_model_id":"deepseek/deepseek-r1"},"amazon.nova-lite-v1:0":{"id":"amazon.nova-lite-v1:0","name":"Nova Lite","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.06,"output":0.24,"cache_read":0.015,"cache_write":0.06},"canonical_model_id":"amazon/nova-lite"},"apac.amazon.nova-pro-v1:0":{"id":"apac.amazon.nova-pro-v1:0","name":"Nova Pro (APAC)","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.84,"output":3.36,"cache_read":0.21,"cache_write":0.84},"canonical_model_id":"amazon/nova-pro"},"us.openai.gpt-6-luna":{"id":"us.openai.gpt-6-luna","name":"GPT-6 Luna (US)","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.11,"output":0.55,"cache_read":0.011,"cache_write":0.1375,"tiers":[{"input":0.22,"output":0.825,"cache_read":0.022,"cache_write":0.275,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.22,"output":0.825,"cache_read":0.022,"cache_write":0.275}},"canonical_model_id":"openai/gpt-6-luna"},"global.xai.grok-4.7":{"id":"global.xai.grok-4.7","name":"Grok 4.7 (Global)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"writer.palmyra-x4-v1:0":{"id":"writer.palmyra-x4-v1:0","name":"Palmyra X4","description":"Enterprise language model for workflow automation, coding, data analysis, and tool use","family":"palmyra","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2024-10-09","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":122880,"output":8192},"status":"deprecated","cost":{"input":2.5,"output":10},"canonical_model_id":"writer/palmyra-x4"},"google.gemma-3-4b-it":{"id":"google.gemma-3-4b-it","name":"Gemma 3 4B IT","description":"Open multimodal Gemma instruction model for efficient text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"status":"deprecated","cost":{"input":0.04,"output":0.08},"canonical_model_id":"google/gemma-3-4b-it"},"eu.anthropic.claude-sonnet-5-5":{"id":"eu.anthropic.claude-sonnet-5-5","name":"Claude Sonnet 5.5 (EU)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.11,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"jp.anthropic.claude-sonnet-4-6":{"id":"jp.anthropic.claude-sonnet-4-6","name":"Claude Sonnet 4.6 (JP)","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"openai.gpt-5.6-sol":{"id":"openai.gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":4.4,"output":22,"cache_read":0.44,"cache_write":5.5,"tiers":[{"input":8.8,"output":33,"cache_read":0.88,"cache_write":11,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8.8,"output":33,"cache_read":0.88,"cache_write":11}},"canonical_model_id":"openai/gpt-5.6-sol"},"in.openai.gpt-5.6-luna":{"id":"in.openai.gpt-5.6-luna","name":"GPT-5.6 Luna (India)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.22,"output":1.32,"cache_read":0.022,"cache_write":0.275,"tiers":[{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55}},"canonical_model_id":"openai/gpt-5.6-luna"},"us.moonshotai.kimi-k3":{"id":"us.moonshotai.kimi-k3","name":"Kimi K3 (US)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"moonshotai/kimi-k3"},"qwen.qwen3-coder-next":{"id":"qwen.qwen3-coder-next","name":"Qwen3 Coder Next","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.5,"output":1.2},"canonical_model_id":"alibaba/qwen3-coder-next"},"minimax.minimax-m2.1":{"id":"minimax.minimax-m2.1","name":"MiniMax-M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"status":"deprecated","cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.1"},"eu.anthropic.claude-opus-4-6-v1":{"id":"eu.anthropic.claude-opus-4-6-v1","name":"Claude Opus 4.6 (EU)","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-6"},"deepseek.r1-v1:0":{"id":"deepseek.r1-v1:0","name":"DeepSeek-R1","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"status":"deprecated","cost":{"input":1.35,"output":5.4},"canonical_model_id":"deepseek/deepseek-r1"},"anthropic.claude-sonnet-4-6":{"id":"anthropic.claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"apac.amazon.nova-micro-v1:0":{"id":"apac.amazon.nova-micro-v1:0","name":"Nova Micro (APAC)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.037,"output":0.148,"cache_read":0.00925,"cache_write":0.037},"canonical_model_id":"amazon/nova-micro"},"global.openai.gpt-5.6-terra":{"id":"global.openai.gpt-5.6-terra","name":"GPT-5.6 Terra (Global)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"qwen.qwen3-32b-v1:0":{"id":"qwen.qwen3-32b-v1:0","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-32b"},"mistral.devstral-2-123b":{"id":"mistral.devstral-2-123b","name":"Devstral 2 123B","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"status":"deprecated","cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/devstral-2512"},"qwen.qwen3-vl-235b-a22b":{"id":"qwen.qwen3-vl-235b-a22b","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262000},"cost":{"input":0.53,"output":2.66},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"anthropic.claude-sonnet-5-5":{"id":"anthropic.claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"global.anthropic.claude-fable-5":{"id":"global.anthropic.claude-fable-5","name":"Claude Fable 5 (Global)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"us.meta.llama3-1-8b-instruct-v1:0":{"id":"us.meta.llama3-1-8b-instruct-v1:0","name":"Llama 3.1 8B Instruct (US)","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.22,"output":0.22},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"anthropic.claude-opus-4-7":{"id":"anthropic.claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"nvidia.nemotron-nano-12b-v2":{"id":"nvidia.nemotron-nano-12b-v2","name":"NVIDIA Nemotron Nano 12B v2 VL BF16","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-28","last_updated":"2025-10-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.2,"output":0.6},"canonical_model_id":"nvidia/nemotron-nano-12b-v2-vl"},"eu.anthropic.claude-opus-4-8":{"id":"eu.anthropic.claude-opus-4-8","name":"Claude Opus 4.8 (EU)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-8"},"apac.amazon.nova-lite-v1:0":{"id":"apac.amazon.nova-lite-v1:0","name":"Nova Lite (APAC)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.063,"output":0.252,"cache_read":0.01575,"cache_write":0.063},"canonical_model_id":"amazon/nova-lite"},"nvidia.nemotron-nano-9b-v2":{"id":"nvidia.nemotron-nano-9b-v2","name":"NVIDIA Nemotron Nano 9B v2","description":"Compact Nemotron model for efficient reasoning and deployable AI agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09","release_date":"2025-08-18","last_updated":"2025-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.06,"output":0.23},"canonical_model_id":"nvidia/nemotron-nano-9b-v2"},"jp.anthropic.claude-opus-5-5":{"id":"jp.anthropic.claude-opus-5-5","name":"Claude Opus 5.5 (JP)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.22,"cache_write":5.5},"canonical_model_id":"anthropic/claude-opus-5-5"},"nvidia.nemotron-nano-3-30b":{"id":"nvidia.nemotron-nano-3-30b","name":"NVIDIA Nemotron Nano 3 30B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.06,"output":0.24},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"jp.anthropic.claude-sonnet-4-5-20250929-v1:0":{"id":"jp.anthropic.claude-sonnet-4-5-20250929-v1:0","name":"Claude Sonnet 4.5 (JP)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"global.anthropic.claude-opus-5-5":{"id":"global.anthropic.claude-opus-5-5","name":"Claude Opus 5.5 (Global)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"in.anthropic.claude-opus-5":{"id":"in.anthropic.claude-opus-5","name":"Claude Opus 5 (India)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-5"},"xai.grok-4.3":{"id":"xai.grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-06-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"mistral.ministral-3-3b-instruct":{"id":"mistral.ministral-3-3b-instruct","name":"Ministral 3 3B","description":"Compact open vision-language model for edge deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.1,"output":0.1},"canonical_model_id":"mistral/ministral-3-3b-instruct-2512"},"global.openai.gpt-6-sol":{"id":"global.openai.gpt-6-sol","name":"GPT-6 Sol (Global)","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"us-gov.openai.gpt-oss-20b-1:0":{"id":"us-gov.openai.gpt-oss-20b-1:0","name":"gpt-oss-20b (GovCloud)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.084,"output":0.36},"canonical_model_id":"openai/gpt-oss-20b"},"moonshotai.kimi-k2.5":{"id":"moonshotai.kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-02-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.6,"output":3},"canonical_model_id":"moonshotai/kimi-k2.5"},"us.xai.grok-4.6":{"id":"us.xai.grok-4.6","name":"Grok 4.6 (US)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2.2,"output":6.6,"cache_read":0.55},"canonical_model_id":"xai/grok-4.6"},"us.anthropic.claude-opus-4-5-20251101-v1:0":{"id":"us.anthropic.claude-opus-4-5-20251101-v1:0","name":"Claude Opus 4.5 (US)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"xai.grok-4.6":{"id":"xai.grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":2.2,"output":6.6,"cache_read":0.55},"canonical_model_id":"xai/grok-4.6"},"anthropic.claude-opus-4-5-20251101-v1:0":{"id":"anthropic.claude-opus-4-5-20251101-v1:0","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"writer.palmyra-x5-v1:0":{"id":"writer.palmyra-x5-v1:0","name":"Palmyra X5","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"palmyra","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1040000,"input":1040000,"output":8192},"status":"deprecated","cost":{"input":0.6,"output":6},"canonical_model_id":"writer/palmyra-x5"},"qwen.qwen3-235b-a22b-2507-v1:0":{"id":"qwen.qwen3-235b-a22b-2507-v1:0","name":"Qwen3 235B-A22B Instruct 2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.22,"output":0.88},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"global.anthropic.claude-opus-4-7":{"id":"global.anthropic.claude-opus-4-7","name":"Claude Opus 4.7 (Global)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"us.anthropic.claude-haiku-5-5":{"id":"us.anthropic.claude-haiku-5-5","name":"Claude Haiku 5.5 (US)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.11,"output":0.55,"cache_read":0.011,"cache_write":0.1375,"tiers":[{"input":0.55,"output":2.75,"cache_read":0.055,"cache_write":0.6875,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"eu.anthropic.claude-sonnet-5":{"id":"eu.anthropic.claude-sonnet-5","name":"Claude Sonnet 5 (EU)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5"},"au.anthropic.claude-sonnet-4-5-20250929-v1:0":{"id":"au.anthropic.claude-sonnet-4-5-20250929-v1:0","name":"Claude Sonnet 4.5 (AU)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"us.anthropic.claude-sonnet-5-5":{"id":"us.anthropic.claude-sonnet-5-5","name":"Claude Sonnet 5.5 (US)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.11,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"us.anthropic.claude-opus-4-8":{"id":"us.anthropic.claude-opus-4-8","name":"Claude Opus 4.8 (US)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-8"},"jp.anthropic.claude-haiku-5-5":{"id":"jp.anthropic.claude-haiku-5-5","name":"Claude Haiku 5.5 (JP)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.11,"output":0.55,"cache_read":0.011,"cache_write":0.1375,"tiers":[{"input":0.55,"output":2.75,"cache_read":0.055,"cache_write":0.6875,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"jp.anthropic.claude-opus-4-7":{"id":"jp.anthropic.claude-opus-4-7","name":"Claude Opus 4.7 (JP)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-7"},"us.writer.palmyra-x5-v1:0":{"id":"us.writer.palmyra-x5-v1:0","name":"Palmyra X5 (US)","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"palmyra","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1040000,"input":1040000,"output":8192},"status":"deprecated","cost":{"input":0.6,"output":6},"canonical_model_id":"writer/palmyra-x5"},"global.anthropic.claude-sonnet-5-5":{"id":"global.anthropic.claude-sonnet-5-5","name":"Claude Sonnet 5.5 (Global)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"moonshot.kimi-k2-thinking":{"id":"moonshot.kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16000},"status":"deprecated","cost":{"input":0.6,"output":2.5},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"us.openai.gpt-5.6-luna":{"id":"us.openai.gpt-5.6-luna","name":"GPT-5.6 Luna (US)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.22,"output":1.32,"cache_read":0.022,"cache_write":0.275,"tiers":[{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55}},"canonical_model_id":"openai/gpt-5.6-luna"},"mistral.mistral-large-3-675b-instruct":{"id":"mistral.mistral-large-3-675b-instruct","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/mistral-large-2512"},"google.gemma-3-27b-it":{"id":"google.gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.23,"output":0.38},"canonical_model_id":"google/gemma-3-27b-it"},"nvidia.nemotron-super-3-120b":{"id":"nvidia.nemotron-super-3-120b","name":"NVIDIA Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.15,"output":0.65},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"openai.gpt-oss-20b":{"id":"openai.gpt-oss-20b","name":"gpt-oss-20b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/v1","shape":"responses"},"cost":{"input":0.07,"output":0.3},"canonical_model_id":"openai/gpt-oss-20b"},"meta.llama3-3-70b-instruct-v1:0":{"id":"meta.llama3-3-70b-instruct-v1:0","name":"Llama 3.3 70B Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"amazon.nova-pro-v1:0":{"id":"amazon.nova-pro-v1:0","name":"Nova Pro","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.8,"output":3.2,"cache_read":0.2,"cache_write":0.8},"canonical_model_id":"amazon/nova-pro"},"us-gov.openai.gpt-oss-120b-1:0":{"id":"us-gov.openai.gpt-oss-120b-1:0","name":"gpt-oss-120b (GovCloud)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.18,"output":0.72},"canonical_model_id":"openai/gpt-oss-120b"},"global.anthropic.claude-sonnet-4-20250514-v1:0":{"id":"global.anthropic.claude-sonnet-4-20250514-v1:0","name":"Claude Sonnet 4 (Global)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"us.anthropic.claude-sonnet-4-6":{"id":"us.anthropic.claude-sonnet-4-6","name":"Claude Sonnet 4.6 (US)","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic.claude-fable-5-1":{"id":"anthropic.claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"us.openai.gpt-6.1-sol":{"id":"us.openai.gpt-6.1-sol","name":"GPT-6.1 Sol (US)","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"ultrafast":{"cost":{"input":13.2,"output":66,"cache_read":0.66,"cache_write":16.5},"provider":{"body":{"additionalModelRequestFields":{"service_tier":"ultrafast"}}}}}},"cost":{"input":2.2,"output":11,"cache_read":0.11,"cache_write":2.75,"tiers":[{"input":4.4,"output":16.5,"cache_read":0.22,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":16.5,"cache_read":0.22,"cache_write":5.5}},"canonical_model_id":"openai/gpt-6.1-sol"},"anthropic.claude-opus-4-6-v1":{"id":"anthropic.claude-opus-4-6-v1","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-6"},"openai.gpt-5.4":{"id":"openai.gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-06-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":2.75,"output":16.5,"cache_read":0.275},"canonical_model_id":"openai/gpt-5.4"},"anthropic.claude-opus-4-1-20250805-v1:0":{"id":"anthropic.claude-opus-4-1-20250805-v1:0","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"global.openai.gpt-5.6-luna":{"id":"global.openai.gpt-5.6-luna","name":"GPT-5.6 Luna (Global)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"mistral.ministral-3-14b-instruct":{"id":"mistral.ministral-3-14b-instruct","name":"Ministral 14B 3.0","description":"Open vision-language model for efficient local deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.2,"output":0.2},"canonical_model_id":"mistral/ministral-3-14b-instruct-2512"}}},"tempr":{"id":"tempr","env":["TEMPR_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.temprhq.io/v1","name":"Tempr Gateway","doc":"https://temprhq.io/docs/gateway-reference.html","models":{"anthropic/claude-haiku-4-5":{"id":"anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4-5-20251101":{"id":"anthropic/claude-opus-4-5-20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-fable-5-1":{"id":"anthropic/claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-sonnet-4-5-20250929":{"id":"anthropic/claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"anthropic/claude-opus-4-8":{"id":"anthropic/claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-sonnet-4-5":{"id":"anthropic/claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-haiku-4-5-20251001":{"id":"anthropic/claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-opus-4-5":{"id":"anthropic/claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"xai/grok-4.3":{"id":"xai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"xai/grok-4.20-0309-reasoning":{"id":"xai/grok-4.20-0309-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"xai/grok-4.7":{"id":"xai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"xai/grok-build-0.1":{"id":"xai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}},"canonical_model_id":"xai/grok-build-0.1"},"xai/grok-4.5":{"id":"xai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"xai/grok-4.6":{"id":"xai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"xai/grok-4.20-0309-non-reasoning":{"id":"xai/grok-4.20-0309-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"moonshot/kimi-k2.6":{"id":"moonshot/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshot/kimi-k3":{"id":"moonshot/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshot/kimi-k2.7-code-highspeed":{"id":"moonshot/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshot/kimi-k2.7-code":{"id":"moonshot/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"reasoning":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-flash":{"id":"deepseek/deepseek-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"reasoning":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"xiaomi-mimo/mimo-v2.6-pro-ultraspeed":{"id":"xiaomi-mimo/mimo-v2.6-pro-ultraspeed","name":"MiMo-V2.6-Pro-UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.35,"output":8.7,"cache_read":0.036},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"xiaomi-mimo/mimo-v2.6-flash":{"id":"xiaomi-mimo/mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"xiaomi-mimo/mimo-v2.6-pro":{"id":"xiaomi-mimo/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"minimax/MiniMax-M3":{"id":"minimax/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"tiers":[{"input":0.6,"output":2.4,"cache_read":0.12,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":0.6,"output":2.4,"cache_read":0.12}},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/MiniMax-M2.7-highspeed":{"id":"minimax/MiniMax-M2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"minimax/MiniMax-M2.7":{"id":"minimax/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"groq/allam-2-7b":{"id":"groq/allam-2-7b","name":"ALLaM-2-7b","description":"ALLaM-2-7b instruction tuned model by SDAIA","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-01-23","last_updated":"2025-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"sdaia/allam-2-7b"},"groq/qwen/qwen3.8-27b":{"id":"groq/qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131042,"output":16384},"cost":{"input":0.8,"output":4},"canonical_model_id":"alibaba/qwen3.8-27b"},"groq/openai/gpt-oss-20b":{"id":"groq/openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-oss-20b"},"groq/openai/gpt-oss-120b":{"id":"groq/openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-10-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-oss-120b"},"zai/glm-4.6v":{"id":"zai/glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.3,"output":0.9},"canonical_model_id":"zhipuai/glm-4.6v"},"zai/glm-5.3-flash":{"id":"zai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai/glm-4.6v-flash":{"id":"zai/glm-4.6v-flash","name":"GLM-4.6V-Flash","description":"Lightweight GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6v-flash"},"zai/glm-4.5v":{"id":"zai/glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":16384},"cost":{"input":0.6,"output":1.8},"canonical_model_id":"zhipuai/glm-4.5v"},"zai/glm-4.5-air":{"id":"zai/glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.2,"output":1.1,"cache_read":0.03,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-air"},"zai/glm-5.3":{"id":"zai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"zai/glm-4.7":{"id":"zai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7"},"zai/glm-5.1":{"id":"zai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"zai/glm-4.7-flash":{"id":"zai/glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"zai/glm-5":{"id":"zai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2,"cache_write":0},"canonical_model_id":"zhipuai/glm-5"},"zai/glm-5v-turbo":{"id":"zai/glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24,"cache_write":0},"canonical_model_id":"zhipuai/glm-5v-turbo"},"zai/glm-5.3-flashx":{"id":"zai/glm-5.3-flashx","name":"GLM-5.3-FlashX","description":"High-speed GLM-5.3-Flash serving option for coding and agent workflows","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.37,"output":1.25,"cache_read":0.075,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai/glm-5-turbo":{"id":"zai/glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24,"cache_write":0},"canonical_model_id":"zhipuai/glm-5-turbo"},"zai/glm-4.6":{"id":"zai/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6"},"zai/glm-5.2":{"id":"zai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"zai/glm-4.5":{"id":"zai/glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5"},"zai/glm-4.7-flashx":{"id":"zai/glm-4.7-flashx","name":"GLM-4.7-FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.07,"output":0.4,"cache_read":0.01,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"zai/glm-4.5-flash":{"id":"zai/glm-4.5-flash","name":"GLM-4.5-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":98304},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-flash"},"google/gemini-flash-lite-latest":{"id":"google/gemini-flash-lite-latest","name":"Gemini Flash-Lite Latest","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-flash-lite-latest"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemini-embedding-001":{"id":"google/gemini-embedding-001","type":"embedding","name":"Gemini Embedding 001","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-05","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2048,"output":1},"cost":{"input":0.15,"output":0},"canonical_model_id":"google/gemini-embedding-001"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-embedding-2":{"id":"google/gemini-embedding-2","type":"embedding","name":"Gemini Embedding 2","description":"Multimodal embedding model mapping text, images, video, audio, and PDFs into a unified embedding space","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-11","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1},"cost":{"input":0.2,"output":0},"canonical_model_id":"google/gemini-embedding-2"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-flash-latest":{"id":"google/gemini-flash-latest","name":"Gemini Flash Latest","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-flash-latest"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"cerebras/qwen-3.8-27b":{"id":"cerebras/qwen-3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-09-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":40960},"cost":{"input":0.99,"output":1.49},"canonical_model_id":"alibaba/qwen3.8-27b"},"cerebras/gpt-oss-120b":{"id":"cerebras/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2026-06-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":40960},"cost":{"input":0.35,"output":0.75},"canonical_model_id":"openai/gpt-oss-120b"},"mistral/mistral-medium-2604":{"id":"mistral/mistral-medium-2604","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"mistral/mistral-medium-2604"},"mistral/zai-glm-5-3":{"id":"mistral/zai-glm-5-3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.3"},"mistral/zai-glm-5-2":{"id":"mistral/zai-glm-5-2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"beta","cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"mistral/mistral-small-latest":{"id":"mistral/mistral-small-latest","name":"Mistral Small (latest)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"mistral/mistral-small-latest"},"mistral/mistral-embed":{"id":"mistral/mistral-embed","type":"embedding","name":"Mistral Embed","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"mistral-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-12-11","last_updated":"2023-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":3072},"cost":{"input":0.1,"output":0},"canonical_model_id":"mistral/mistral-embed"},"mistral/mistral-large-2512":{"id":"mistral/mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/mistral-large-2512"},"mistral/voxtral-small-latest":{"id":"mistral/voxtral-small-latest","name":"Voxtral Small (latest)","description":"Instruct model with native audio input for speech understanding and tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/voxtral-small-latest"},"mistral/mistral-small-2603":{"id":"mistral/mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"mistral/mistral-small-2603"},"mistral/mistral-large-latest":{"id":"mistral/mistral-large-latest","name":"Mistral Large (latest)","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/mistral-large-latest"},"mistral/mistral-medium-latest":{"id":"mistral/mistral-medium-latest","name":"Mistral Medium (latest)","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"mistral/mistral-medium-latest"}}},"fireworks-ai":{"id":"fireworks-ai","env":["FIREWORKS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.fireworks.ai/inference/v1/","name":"Fireworks AI","doc":"https://fireworks.ai/docs/","models":{"accounts/fireworks/routers/kimi-fast-latest":{"id":"accounts/fireworks/routers/kimi-fast-latest","name":"Kimi Fast Latest","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-27","last_updated":"2026-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.5,"output":22.5,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"accounts/fireworks/routers/kimi-k3-fast":{"id":"accounts/fireworks/routers/kimi-k3-fast","name":"Kimi K3 Fast","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-27","last_updated":"2026-07-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.5,"output":22.5,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"accounts/fireworks/routers/glm-fast-latest":{"id":"accounts/fireworks/routers/glm-fast-latest","name":"GLM 5.3 Fast (Latest)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-09-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048572,"output":262144},"cost":{"input":2.1,"output":6.6,"cache_read":0.39},"canonical_model_id":"zhipuai/glm-5.3"},"accounts/fireworks/routers/minimax-latest":{"id":"accounts/fireworks/routers/minimax-latest","name":"MiniMax Latest","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-12","last_updated":"2026-09-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":512000},"experimental":{"modes":{"priority":{"cost":{"input":0.45,"output":1.8,"cache_read":0.09},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"accounts/fireworks/routers/glm-flash-latest":{"id":"accounts/fireworks/routers/glm-flash-latest","name":"GLM Flash Latest (GLM 5.3 Flash)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048573,"output":131072},"experimental":{"modes":{"priority":{"cost":{"input":0.1875,"output":0.625,"cache_read":0.0375},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"accounts/fireworks/routers/qwen-max-latest":{"id":"accounts/fireworks/routers/qwen-max-latest","name":"Qwen Max Latest (Qwen3.8 Max)","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-09-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"experimental":{"modes":{"priority":{"cost":{"input":3,"output":9,"cache_read":0.375},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max"},"accounts/fireworks/routers/deepseek-flash-latest":{"id":"accounts/fireworks/routers/deepseek-flash-latest","name":"DeepSeek Flash Latest","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"experimental":{"modes":{"priority":{"cost":{"input":0.375,"output":1.5,"cache_read":0.0075},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"accounts/fireworks/routers/kimi-latest":{"id":"accounts/fireworks/routers/kimi-latest","name":"Kimi Latest","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-27","last_updated":"2026-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"experimental":{"modes":{"priority":{"cost":{"input":3.75,"output":18.75,"cache_read":0.375},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"accounts/fireworks/routers/glm-latest":{"id":"accounts/fireworks/routers/glm-latest","name":"GLM Latest","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-09-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048573,"output":262144},"experimental":{"modes":{"priority":{"cost":{"input":1.75,"output":5.5,"cache_read":0.325},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"accounts/fireworks/routers/glm-5p3-fast":{"id":"accounts/fireworks/routers/glm-5p3-fast","name":"GLM 5.3 Fast","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-09-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048572,"output":262144},"cost":{"input":2.1,"output":6.6,"cache_read":0.39},"canonical_model_id":"zhipuai/glm-5.3"},"accounts/fireworks/models/inkling":{"id":"accounts/fireworks/models/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"accounts/fireworks/models/qwen3p8-2p4t-a95b":{"id":"accounts/fireworks/models/qwen3p8-2p4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"accounts/fireworks/models/glm-5p3-flash":{"id":"accounts/fireworks/models/glm-5p3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-09-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048573,"output":131072},"experimental":{"modes":{"priority":{"cost":{"input":0.1875,"output":0.625,"cache_read":0.0375},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"accounts/fireworks/models/deepseek-v4p1-flash-1017":{"id":"accounts/fireworks/models/deepseek-v4p1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"experimental":{"modes":{"priority":{"cost":{"input":0.375,"output":1.5,"cache_read":0.0075},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"accounts/fireworks/models/kimi-k3":{"id":"accounts/fireworks/models/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-27","last_updated":"2026-07-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"experimental":{"modes":{"priority":{"cost":{"input":3.75,"output":18.75,"cache_read":0.375},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"accounts/fireworks/models/glm-5p3":{"id":"accounts/fireworks/models/glm-5p3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-09-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048573,"output":262144},"experimental":{"modes":{"priority":{"cost":{"input":1.75,"output":5.5,"cache_read":0.325},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"accounts/fireworks/models/nemotron-3-ultra-nvfp4":{"id":"accounts/fireworks/models/nemotron-3-ultra-nvfp4","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"accounts/fireworks/models/ember-1":{"id":"accounts/fireworks/models/ember-1","name":"Ember-1","description":"Specialized model from Fireworks built on Kimi K3, producing shorter reasoning traces with approximately 40% fewer tokens while maintaining comparable quality","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3}},"accounts/fireworks/models/qwen3p8-max":{"id":"accounts/fireworks/models/qwen3p8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"experimental":{"modes":{"priority":{"cost":{"input":3,"output":9,"cache_read":0.375},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max"},"accounts/fireworks/models/minimax-m3":{"id":"accounts/fireworks/models/minimax-m3","name":"MiniMax-M3","description":"Fireworks text-only MiniMax coding model for long-context reasoning and agent tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":512000},"experimental":{"modes":{"priority":{"cost":{"input":0.45,"output":1.8,"cache_read":0.09},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.3,"output":1.2,"cache_read":0.06}},"accounts/fireworks/models/muse-spark-open":{"id":"accounts/fireworks/models/muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"accounts/fireworks/models/gpt-oss-120b":{"id":"accounts/fireworks/models/gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2026-06-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"experimental":{"modes":{"priority":{"cost":{"input":0.18,"output":0.72,"cache_read":0.018},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.15,"output":0.6,"cache_read":0.015}},"accounts/fireworks/models/deepseek-v4p1-flash":{"id":"accounts/fireworks/models/deepseek-v4p1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"experimental":{"modes":{"priority":{"cost":{"input":0.375,"output":1.5,"cache_read":0.0075},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"accounts/fireworks/models/nemotron-lightning-3p5-30b-a3b":{"id":"accounts/fireworks/models/nemotron-lightning-3p5-30b-a3b","name":"Nemotron 3.5 Lightning 30B A3B","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"experimental":{"modes":{"priority":{"cost":{"input":0.0625,"output":0.25,"cache_read":0.0125},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.05,"output":0.2,"cache_read":0.01},"canonical_model_id":"nvidia/nemotron-3.5-lightning"}}},"reflection":{"id":"reflection","env":["REFLECTION_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.reflection.ai/openai/v1","name":"Reflection","doc":"https://developers.reflection.ai/models","models":{"Beam-1.0-501B-A23B":{"id":"Beam-1.0-501B-A23B","name":"Beam 1.0 501B A23B","description":"Sparse mixture-of-experts reasoning model with 23B active parameters for coding and agentic workflows","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"reflection/beam-1.0-501b-a23b"}}},"fastrouter":{"id":"fastrouter","env":["FASTROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://go.fastrouter.ai/api/v1","name":"FastRouter","doc":"https://fastrouter.ai/models","models":{"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":32000}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-sonnet-4":{"id":"anthropic/claude-sonnet-4","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":32000}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"anthropic/claude-opus-4.1":{"id":"anthropic/claude-opus-4.1","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":32000}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":32000}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.74,"output":3.48},"canonical_model_id":"deepseek/deepseek-v4-pro"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.05,"output":3.5},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.95,"output":3.15}},"deepseek-ai/deepseek-r1-distill-llama-70b":{"id":"deepseek-ai/deepseek-r1-distill-llama-70b","name":"DeepSeek R1 Distill Llama 70B","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-10","release_date":"2025-01-23","last_updated":"2025-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.03,"output":0.14}},"qwen/qwen3-coder":{"id":"qwen/qwen3-coder","name":"Qwen3 Coder","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":66536},"cost":{"input":0.3,"output":1.2}},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.75,"output":3.5},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k2":{"id":"moonshotai/kimi-k2","name":"Kimi K2","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.55,"output":2.2}},"bytedance/seedance-2":{"id":"bytedance/seedance-2","name":"Seedance 2","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"seed","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":4096,"output":0}},"wanx/wan-v2-6":{"id":"wanx/wan-v2-6","name":"Wan 2.6","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text","image"],"output":["video"]},"open_weights":true,"limit":{"context":400000,"output":0}},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5},"canonical_model_id":"xai/grok-4.3"},"x-ai/grok-build-0.1":{"id":"x-ai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2},"canonical_model_id":"xai/grok-build-0.1"},"x-ai/grok-4":{"id":"x-ai/grok-4","name":"Grok 4","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.75,"cache_write":15}},"google/gemini-3-pro-image-preview":{"id":"google/gemini-3-pro-image-preview","type":"image","name":"Nano Banana Pro Preview","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12},"canonical_model_id":"google/gemini-3-pro-image-preview"},"google/veo3.1-fast":{"id":"google/veo3.1-fast","name":"Veo 3.1 Fast","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":400000,"output":0}},"google/gemini-3.1-flash-image-preview":{"id":"google/gemini-3.1-flash-image-preview","type":"image","name":"Nano Banana 2 Preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":65536},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image-preview"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.31}},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9},"canonical_model_id":"google/gemini-3.5-flash"},"google/imagen-4.0-fast":{"id":"google/imagen-4.0-fast","name":"Imagen 4 Fast","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"imagen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":480,"output":0}},"google/imagen-4.0-ultra":{"id":"google/imagen-4.0-ultra","name":"Imagen 4 Ultra","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"imagen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":480,"output":0}},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.13,"output":0.38},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.0375}},"google/veo3.1-lite":{"id":"google/veo3.1-lite","name":"Veo 3.1 Lite","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":400000,"output":0}},"google/veo3.1":{"id":"google/veo3.1","name":"Veo 3.1","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":400000,"output":0}},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-10-01","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-10-01","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"openai/gpt-image-2":{"id":"openai/gpt-image-2","name":"GPT Image 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":128000,"output":0}},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.05,"output":0.2}},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6}},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-realtime-1.5":{"id":"openai/gpt-realtime-1.5","name":"GPT Realtime 1.5","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-06-01","last_updated":"2025-06-01","modalities":{"input":["text","audio","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":32000,"output":4096},"cost":{"input":4,"output":16}},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-10-01","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005}},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14},"canonical_model_id":"openai/gpt-5.3-codex"},"sarvam/sarvam-105b":{"id":"sarvam/sarvam-105b","name":"Sarvam 105B","description":"Flagship Indian-language reasoning model for enterprise multilingual applications","family":"sarvam","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-18","last_updated":"2026-02-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.04,"output":0.16},"canonical_model_id":"sarvam/sarvam-105b"},"sarvam/sarvam-30b":{"id":"sarvam/sarvam-30b","name":"Sarvam 30B","description":"Efficient Indian-language reasoning model for chat, coding, and multilingual work","family":"sarvam","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-18","last_updated":"2026-02-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.02,"output":0.1},"canonical_model_id":"sarvam/sarvam-30b"},"leonardo-ai/lucid-realism":{"id":"leonardo-ai/lucid-realism","name":"Lucid Realism","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"lucid","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-01","last_updated":"2025-06-01","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":4096,"output":0}},"leonardo-ai/lucid-origin":{"id":"leonardo-ai/lucid-origin","name":"Lucid Origin","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"lucid","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-01","last_updated":"2025-06-01","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":4096,"output":0}}}},"scnet-token-plan":{"id":"scnet-token-plan","env":["SCNET_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.scnet.cn/api/llm/v1","name":"SCNet Token Plan","doc":"https://www.scnet.cn/ac/openapi/doc/2.0/moduleapi/plans/token-plan.html","models":{"Kimi-K2.6":{"id":"Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"minimax/MiniMax-M3"},"Kimi-K2.7-Code":{"id":"Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"Qwen3.8-Flash":{"id":"Qwen3.8-Flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"alibaba/qwen3.8-flash"},"GLM-5.1":{"id":"GLM-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.1"},"Kimi-K3":{"id":"Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k3"},"GLM-5.3":{"id":"GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.3"},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"minimax/MiniMax-M2.5"},"DeepSeek-V4-Flash-0731":{"id":"DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"DeepSeek-V4.1-Flash":{"id":"DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"GLM-5":{"id":"GLM-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5"},"GLM-5.2":{"id":"GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.2"},"Qwen3.8-Max":{"id":"Qwen3.8-Max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"alibaba/qwen3.8-max"},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"minimax/MiniMax-M2.7"},"DeepSeek-V4-Flash":{"id":"DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"DeepSeek-V4-Pro":{"id":"DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-pro"},"Kimi-K2.5":{"id":"Kimi-K2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k2.5"},"GLM-5.3-Flash":{"id":"GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"DeepSeek-V4-Pro-0813":{"id":"DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"}}},"stepfun":{"id":"stepfun","env":["STEPFUN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.stepfun.com/v1","name":"StepFun (China)","doc":"https://platform.stepfun.com/docs/zh/overview/concept","models":{"stepaudio-2.5-tts":{"id":"stepaudio-2.5-tts","name":"StepAudio 2.5 TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"step","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-16","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"stepaudio-2.5-asr":{"id":"stepaudio-2.5-asr","name":"StepAudio 2.5 ASR","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"step","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-24","last_updated":"2026-07-02","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0}},"step-2-16k":{"id":"step-2-16k","name":"Step 2 (16K)","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-01-01","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":5.21,"output":16.44,"cache_read":1.04}},"step-3.7-flash":{"id":"step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-06-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.185,"output":1.11,"cache_read":0.037},"canonical_model_id":"stepfun/step-3.7-flash"},"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":0.959,"output":2.741,"cache_read":0.048},"canonical_model_id":"stepfun/step-5-preview"},"step-tts-2":{"id":"step-tts-2","name":"Step TTS 2","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"step","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-01","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"step-3.5-flash-2603":{"id":"step-3.5-flash-2603","name":"Step 3.5 Flash 2603","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.1,"output":0.3,"cache_read":0.02}},"step-1-32k":{"id":"step-1-32k","name":"Step 1 (32K)","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-01-01","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":2.05,"output":9.59,"cache_read":0.41}},"step-3.5-flash":{"id":"step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-06-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.1,"output":0.3,"cache_read":0.02}}}},"xiaomi-token-plan-ams":{"id":"xiaomi-token-plan-ams","env":["XIAOMI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://token-plan-ams.xiaomimimo.com/v1","name":"Xiaomi Token Plan (Europe)","doc":"https://platform.xiaomimimo.com/#/docs","models":{"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"mimo-v2-pro":{"id":"mimo-v2-pro","name":"MiMo-V2-Pro","description":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2-pro"},"mimo-v2.5-tts":{"id":"mimo-v2.5-tts","name":"MiMo-V2.5-TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0}},"mimo-v2-tts":{"id":"mimo-v2-tts","name":"MiMo-V2-TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0}},"mimo-v2.5-tts-voiceclone":{"id":"mimo-v2.5-tts-voiceclone","name":"MiMo-V2.5-TTS-VoiceClone","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0}},"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"mimo-v2.5-tts-voicedesign":{"id":"mimo-v2.5-tts-voicedesign","name":"MiMo-V2.5-TTS-VoiceDesign","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0}},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5-pro"}}},"melious":{"id":"melious","env":["MELIOUS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.melious.ai/v1","name":"Melious","doc":"https://melious.ai/docs/reference/models","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.11592,"output":0.46368,"cache_read":0.023184},"canonical_model_id":"zhipuai/glm-5.3-flash"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.11592,"output":0.2898,"cache_read":0.023184},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.81144,"output":4.0572,"cache_read":0.266616},"canonical_model_id":"moonshotai/kimi-k2.6"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3.1878,"output":15.939,"cache_read":0.788256},"canonical_model_id":"moonshotai/kimi-k3"},"deepseek-r1-0528":{"id":"deepseek-r1-0528","name":"DeepSeek-R1","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":32768},"cost":{"input":0.69552,"output":2.78208,"cache_read":0.185472},"canonical_model_id":"deepseek/deepseek-r1"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.1592,"output":3.4776,"cache_read":0.23184},"canonical_model_id":"zhipuai/glm-5.3"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131072},"cost":{"input":1.50696,"output":4.6368,"cache_read":0.370944},"canonical_model_id":"zhipuai/glm-5.1"},"glm-5":{"id":"glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131072},"cost":{"input":1.10124,"output":3.36168,"cache_read":0.266616},"canonical_model_id":"zhipuai/glm-5"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.81144,"output":3.4776,"cache_read":0.220248},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":64000},"cost":{"input":0.34776,"output":0.5796,"cache_read":0.092736},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.33675,"output":1.347,"cache_read":0.033675},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.23184,"output":1.1592,"cache_read":0.011592},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.5796,"output":2.95596,"cache_read":0.139104},"canonical_model_id":"moonshotai/kimi-k2.5"}}},"huggingface":{"id":"huggingface","env":["HF_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"https://router.huggingface.co/v1","name":"Hugging Face","doc":"https://huggingface.co/docs/inference-providers","models":{"deepseek-ai/DeepSeek-V4-Flash-Vision-Exp":{"id":"deepseek-ai/DeepSeek-V4-Flash-Vision-Exp","name":"DeepSeek V4 Flash Vision Exp","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.44,"output":1.32},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek-V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.27,"output":1},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek-ai/DeepSeek-V3":{"id":"deepseek-ai/DeepSeek-V3","name":"DeepSeek-V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":8192},"cost":{"input":0.4,"output":1.3},"canonical_model_id":"deepseek/deepseek-v3"},"deepseek-ai/DeepSeek-R1-0528":{"id":"deepseek-ai/DeepSeek-R1-0528","name":"DeepSeek-R1-0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":3,"output":5}},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.14,"output":0.28},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-V4.1-Flash":{"id":"deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-ai/DeepSeek-V4.1-Flash-1017":{"id":"deepseek-ai/DeepSeek-V4.1-Flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-ai/DeepSeek-V3-0324":{"id":"deepseek-ai/DeepSeek-V3-0324","name":"DeepSeek V3 0324","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.27,"output":1.12},"canonical_model_id":"deepseek/deepseek-v3-0324"},"deepseek-ai/DeepSeek-V3.2":{"id":"deepseek-ai/DeepSeek-V3.2","name":"DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":65536},"cost":{"input":0.28,"output":0.4}},"deepseek-ai/DeepSeek-V4-Flash":{"id":"deepseek-ai/DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.14,"output":0.28},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.435,"output":0.87,"cache_read":0.003625},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/DeepSeek-R1":{"id":"deepseek-ai/DeepSeek-R1","name":"DeepSeek-R1","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":32768},"cost":{"input":0.7,"output":2.5},"canonical_model_id":"deepseek/deepseek-r1"},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.96},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"meta-llama/Llama-3.3-70B-Instruct":{"id":"meta-llama/Llama-3.3-70B-Instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.59,"output":0.79},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/Llama-3.1-8B-Instruct":{"id":"meta-llama/Llama-3.1-8B-Instruct","name":"Llama-3.1-8B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.06,"output":0.06},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi-K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16}},"moonshotai/Kimi-K2-Instruct-0905":{"id":"moonshotai/Kimi-K2-Instruct-0905","name":"Kimi-K2-Instruct-0905","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-09-04","last_updated":"2025-09-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":1,"output":3}},"moonshotai/Kimi-K2-Instruct":{"id":"moonshotai/Kimi-K2-Instruct","name":"Kimi-K2-Instruct","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-07-14","last_updated":"2025-07-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":1,"output":3}},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K2-Thinking":{"id":"moonshotai/Kimi-K2-Thinking","name":"Kimi-K2-Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":2.5,"cache_read":0.15}},"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3,"output":15},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.5":{"id":"moonshotai/Kimi-K2.5","name":"Kimi-K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-01-01","last_updated":"2026-01-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3,"cache_read":0.1}},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":512000},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMaxAI/MiniMax-M2":{"id":"MiniMaxAI/MiniMax-M2","name":"MiniMax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2"},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03}},"MiniMaxAI/MiniMax-M1-80k":{"id":"MiniMaxAI/MiniMax-M1-80k","name":"MiniMax M1 80K","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":80000},"cost":{"input":0.55,"output":2.2},"canonical_model_id":"minimax/minimax-m1-80k"},"MiniMaxAI/MiniMax-M2.7":{"id":"MiniMaxAI/MiniMax-M2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06}},"MiniMaxAI/MiniMax-M2.1":{"id":"MiniMaxAI/MiniMax-M2.1","name":"MiniMax-M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-10","release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2}},"XiaomiMiMo/MiMo-V2.5-Pro":{"id":"XiaomiMiMo/MiMo-V2.5-Pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1,"output":3},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"XiaomiMiMo/MiMo-V2.5":{"id":"XiaomiMiMo/MiMo-V2.5","name":"MiMo-V2.5","description":"MiMo model for long-context reasoning, perception, and agentic tasks","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.4,"output":2},"canonical_model_id":"xiaomi/mimo-v2.5"},"XiaomiMiMo/MiMo-V2-Flash":{"id":"XiaomiMiMo/MiMo-V2-Flash","name":"MiMo-V2-Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":4096},"cost":{"input":0.1,"output":0.3}},"stepfun-ai/Step-3.5-Flash":{"id":"stepfun-ai/Step-3.5-Flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"stepfun/step-3.5-flash"},"stepfun-ai/Step-3.7-Flash":{"id":"stepfun-ai/Step-3.7-Flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.2,"output":1.15},"canonical_model_id":"stepfun/step-3.7-flash"},"thinkingmachines/Inkling":{"id":"thinkingmachines/Inkling","name":"Inkling","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1,"output":4.05},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/Inkling-Small":{"id":"thinkingmachines/Inkling-Small","name":"Inkling Small","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":1048576},"cost":{"input":0.5,"output":1.2},"canonical_model_id":"thinkingmachines/inkling-small"},"Qwen/Qwen3-Embedding-8B":{"id":"Qwen/Qwen3-Embedding-8B","name":"Qwen 3 Embedding 8B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.01,"output":0}},"Qwen/Qwen3-VL-235B-A22B-Instruct":{"id":"Qwen/Qwen3-VL-235B-A22B-Instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":1.5},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"Qwen/Qwen3-Coder-Next":{"id":"Qwen/Qwen3-Coder-Next","name":"Qwen3-Coder-Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.2,"output":1.5}},"Qwen/Qwen3.5-35B-A3B":{"id":"Qwen/Qwen3.5-35B-A3B","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"Qwen/Qwen3.6-35B-A3B":{"id":"Qwen/Qwen3.6-35B-A3B","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.15,"output":0.95},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen/Qwen3-Coder-480B-A35B-Instruct":{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct","name":"Qwen3-Coder-480B-A35B-Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":66536},"cost":{"input":2,"output":2}},"Qwen/Qwen3-Next-80B-A3B-Instruct":{"id":"Qwen/Qwen3-Next-80B-A3B-Instruct","name":"Qwen3-Next-80B-A3B-Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-11","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":66536},"cost":{"input":0.25,"output":1}},"Qwen/Qwen3-VL-30B-A3B-Instruct":{"id":"Qwen/Qwen3-VL-30B-A3B-Instruct","name":"Qwen3 VL 30B A3B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.2,"output":0.7},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.4,"output":3},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3-Coder-30B-A3B-Instruct":{"id":"Qwen/Qwen3-Coder-30B-A3B-Instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.07,"output":0.26},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"Qwen/Qwen3-235B-A22B-Thinking-2507":{"id":"Qwen/Qwen3-235B-A22B-Thinking-2507","name":"Qwen3-235B-A22B-Thinking-2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.3,"output":3}},"Qwen/Qwen3-VL-235B-A22B-Thinking":{"id":"Qwen/Qwen3-VL-235B-A22B-Thinking","name":"Qwen3 VL 235B A22B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.98,"output":3.95},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-thinking"},"Qwen/Qwen3.5-122B-A10B":{"id":"Qwen/Qwen3.5-122B-A10B","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":3.2},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"Qwen/Qwen3-Next-80B-A3B-Thinking":{"id":"Qwen/Qwen3-Next-80B-A3B-Thinking","name":"Qwen3-Next-80B-A3B-Thinking","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-11","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.3,"output":2}},"Qwen/Qwen3.6-27B":{"id":"Qwen/Qwen3.6-27B","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.47,"output":3.19},"canonical_model_id":"alibaba/qwen3.6-27b"},"Qwen/Qwen2.5-Coder-32B-Instruct":{"id":"Qwen/Qwen2.5-Coder-32B-Instruct","name":"Qwen2.5-Coder-32B-Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.06,"output":0.2},"canonical_model_id":"alibaba/qwen2.5-coder-32b-instruct"},"Qwen/Qwen3.5-9B":{"id":"Qwen/Qwen3.5-9B","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.17,"output":0.25},"canonical_model_id":"alibaba/qwen3.5-9b"},"Qwen/Qwen3-32B":{"id":"Qwen/Qwen3-32B","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.29,"output":0.59},"canonical_model_id":"alibaba/qwen3-32b"},"Qwen/Qwen3-30B-A3B":{"id":"Qwen/Qwen3-30B-A3B","name":"Qwen3 30B A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.5},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"Qwen/Qwen3.5-27B":{"id":"Qwen/Qwen3.5-27B","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":2.4},"canonical_model_id":"alibaba/qwen3.5-27b"},"Qwen/Qwen3.5-397B-A17B":{"id":"Qwen/Qwen3.5-397B-A17B","name":"Qwen3.5-397B-A17B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-02-01","last_updated":"2026-02-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.6,"output":3.6}},"Qwen/Qwen3-235B-A22B":{"id":"Qwen/Qwen3-235B-A22B","name":"Qwen3 235B-A22B","description":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.2,"output":0.8},"canonical_model_id":"alibaba/qwen3-235b-a22b"},"Qwen/Qwen3-Embedding-4B":{"id":"Qwen/Qwen3-Embedding-4B","name":"Qwen 3 Embedding 4B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":2048},"cost":{"input":0.01,"output":0}},"Qwen/Qwen3.8-2.4T-A95B":{"id":"Qwen/Qwen3.8-2.4T-A95B","name":"Qwen3.8 2.4T A95B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":2.5,"output":6.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B-A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.855,"output":2.565},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"zai-org/GLM-4.7":{"id":"zai-org/GLM-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11}},"zai-org/GLM-4.6V-Flash":{"id":"zai-org/GLM-4.6V-Flash","name":"GLM-4.6V-Flash","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":0.9},"canonical_model_id":"zhipuai/glm-4.6v-flash"},"zai-org/GLM-4.6":{"id":"zai-org/GLM-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.55,"output":2.2},"canonical_model_id":"zhipuai/glm-4.6"},"zai-org/AutoGLM-Phone-9B-Multilingual":{"id":"zai-org/AutoGLM-Phone-9B-Multilingual","name":"AutoGLM Phone 9B Multilingual","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-10","last_updated":"2025-12-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.035,"output":0.138},"canonical_model_id":"zhipuai/autoglm-phone-9b-multilingual"},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-03","last_updated":"2026-04-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2}},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-4.5V":{"id":"zai-org/GLM-4.5V","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":0.6,"output":1.8},"canonical_model_id":"zhipuai/glm-4.5v"},"zai-org/GLM-5":{"id":"zai-org/GLM-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2}},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":1.4,"output":4.4},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-4.7-Flash":{"id":"zai-org/GLM-4.7-Flash","name":"GLM-4.7-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-08-08","last_updated":"2025-08-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0,"output":0}},"zai-org/GLM-4.5-Air":{"id":"zai-org/GLM-4.5-Air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.13,"output":0.85},"canonical_model_id":"zhipuai/glm-4.5-air"},"zai-org/GLM-4.5":{"id":"zai-org/GLM-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.6,"output":2.2},"canonical_model_id":"zhipuai/glm-4.5"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5},"canonical_model_id":"zhipuai/glm-5.3-flash"},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.15},"canonical_model_id":"google/gemma-3-12b-it"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.08,"output":0.16},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemma-4-26B-A4B-it":{"id":"google/gemma-4-26B-A4B-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.13,"output":0.4},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemma-3-4b-it":{"id":"google/gemma-3-4b-it","name":"Gemma 3 4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.1},"canonical_model_id":"google/gemma-3-4b-it"},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"tencent/Hy4-preview":{"id":"tencent/Hy4-preview","name":"Hy4 preview","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":64000},"cost":{"input":0.834,"output":2.501},"canonical_model_id":"tencent/hy4-preview"},"tencent/Hy3":{"id":"tencent/Hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":128000},"cost":{"input":0.14,"output":0.58},"canonical_model_id":"tencent/hy3"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.1,"output":0.5},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.25,"output":0.69},"canonical_model_id":"openai/gpt-oss-120b"}}},"klokintegration":{"id":"klokintegration","env":["KLOKINTEGRATION_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api-gw.klok.ipaas.se/proxy/kloker-key/v1","name":"klokintegration.se","doc":"https://klokintegration.se/docs/ai-api","models":{"Kloker-Integration-Developer":{"id":"Kloker-Integration-Developer","name":"Kloker Integration Developer","description":"Knows the customer integration environment and Klok best practices. Opinionated about implementation. The gateway runs ecosystem lookup tools server-side and appends a system-prompt injection. Client system prompts and OpenAI tool calls are preserved.","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-29","last_updated":"2026-08-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":50000},"status":"beta","cost":{"input":0.23,"output":1.16}},"Kloker":{"id":"Kloker","name":"Kloker","description":"Cheap general model with a clean context. Nothing from the customer environment is packed in. It tracks the current best open source model. The Klok team verifies it and upgrades it periodically.","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-29","last_updated":"2026-08-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":50000},"status":"beta","cost":{"input":0.23,"output":1.16}},"Kloker-Integration-Architect":{"id":"Kloker-Integration-Architect","name":"Kloker Integration Architect","description":"Knows the customer integration environment and Klok best practices. Opinionated about structure. The gateway runs ecosystem lookup tools server-side and appends a system-prompt injection (data contracts, CloudEvents, event-driven flows). Client system prompts and OpenAI tool calls are preserved.","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-29","last_updated":"2026-08-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":50000},"status":"beta","cost":{"input":0.23,"output":1.16}}}},"qvac":{"id":"qvac","env":["QVAC_API_KEY"],"npm":"@qvac/ai-sdk-provider","name":"QVAC","doc":"https://www.npmjs.com/package/@qvac/ai-sdk-provider","models":{"qwen3.5-9b":{"id":"qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.5-9b"},"qwen3.5-0.8b":{"id":"qwen3.5-0.8b","name":"Qwen3.5 0.8B","description":"Qwen instruction model for multilingual chat and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0}},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen3.5-4b":{"id":"qwen3.5-4b","name":"Qwen3.5 4B","description":"Qwen instruction model for multilingual chat and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0}},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-oss-20b"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-oss-120b"},"qwen3.6-27b":{"id":"qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.6-27b"},"qwen3.5-2b":{"id":"qwen3.5-2b","name":"Qwen3.5 2B","description":"Qwen instruction model for multilingual chat and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0}},"gemma4-31b":{"id":"gemma4-31b","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-31b-it"}}},"llmtr":{"id":"llmtr","env":["LLMTR_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://llmtr.com/v1","name":"LLMTR","doc":"https://llmtr.com/docs","models":{"muse-glimmer-30b-tr":{"id":"muse-glimmer-30b-tr","name":"Muse Glimmer 30B (TR)","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":2,"output":5,"cache_read":0.5},"canonical_model_id":"meta/muse-glimmer-30b"},"qwen3-6-35b":{"id":"qwen3-6-35b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":5,"output":10},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"magibu-11b-v8":{"id":"magibu-11b-v8","name":"Magibu 11B v8","description":"Turkish-language chat model for instruction following and assistant flows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-06-05","last_updated":"2026-08-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0.1,"output":0.5}},"trendyol-asure-12b":{"id":"trendyol-asure-12b","name":"Trendyol Asure 12B","description":"Turkish-language multimodal instruct model built on Gemma 3 12B for e-commerce text, chat, and image-text tasks","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-02-19","last_updated":"2026-02-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":40960},"cost":{"input":0.1,"output":0.5,"cache_read":0.025},"canonical_model_id":"trendyol/asure-12b"},"gemma-4":{"id":"gemma-4","name":"Gemma 4","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":2,"output":5,"cache_read":0.5},"canonical_model_id":"google/gemma-4-31b-it"},"medgemma-4b":{"id":"medgemma-4b","name":"MedGemma 4B","description":"Multimodal medical-domain Gemma variant for text and image analysis","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-04-26","last_updated":"2026-08-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"status":"deprecated","cost":{"input":3,"output":5}},"meta/muse-spark-1.2-contributor":{"id":"meta/muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.2},"canonical_model_id":"meta/muse-spark-1.2"},"qwen/qwen-plus":{"id":"qwen/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.4,"output":1.2},"canonical_model_id":"alibaba/qwen-plus"},"qwen/qwen-flash":{"id":"qwen/qwen-flash","name":"Qwen Flash","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.05,"output":0.4},"canonical_model_id":"alibaba/qwen-flash"},"qwen/qwen3-coder-plus":{"id":"qwen/qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Hosted Qwen coder for software agents, repo edits, and long-context code","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1,"output":5},"canonical_model_id":"alibaba/qwen3-coder-plus"},"qwen/qwen3.5-plus":{"id":"qwen/qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.4,"output":2.4},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3-coder-flash":{"id":"qwen/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":1.5},"canonical_model_id":"alibaba/qwen3-coder-flash"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":1.6},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3-vl-plus":{"id":"qwen/qwen3-vl-plus","name":"Qwen3-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32768},"cost":{"input":0.2,"output":1.6},"canonical_model_id":"alibaba/qwen3-vl-plus"},"qwen/qwen3-max":{"id":"qwen/qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"cost":{"input":1.2,"output":6},"canonical_model_id":"alibaba/qwen3-max"},"poolside/laguna-xs-2.1":{"id":"poolside/laguna-xs-2.1","name":"Laguna XS 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-02","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-xs-2.1"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.87,"output":4.68},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.58,"output":1.44},"canonical_model_id":"thinkingmachines/inkling-small"},"mimo/mimo-v2.5-pro":{"id":"mimo/mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.435,"output":0.87},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"mimo/mimo-v2.5":{"id":"mimo/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.14,"output":0.28},"canonical_model_id":"xiaomi/mimo-v2.5"},"upstage/solar-pro4":{"id":"upstage/solar-pro4","name":"Solar Pro 4","description":"Upstage's flagship model, specialized for agentic use","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.03,"output":0.12},"canonical_model_id":"upstage/solar-pro4"},"upstage/solar-pro3":{"id":"upstage/solar-pro3","name":"Solar Pro 3","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2026-01-26","last_updated":"2026-03-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"upstage/solar-pro3"},"upstage/solar-pro2":{"id":"upstage/solar-pro2","name":"Solar Pro 2","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":8192},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"upstage/solar-pro2"},"perplexity/sonar-deep-research":{"id":"perplexity/sonar-deep-research","name":"Sonar Deep Research","description":"Sonar search model for autonomous research and citation-backed long-form reports","family":"sonar","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-02-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":2,"output":8,"reasoning":3},"canonical_model_id":"perplexity/sonar-deep-research"},"publicai/apertus-8b-instruct":{"id":"publicai/apertus-8b-instruct","name":"Apertus 8B Instruct","description":"Fully open 8B multilingual LLM supporting 1800+ languages with 65K context. Trained on compliant open data. Apache 2.0, EU AI Act compliant.","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":8192},"cost":{"input":0.1,"output":0.2},"canonical_model_id":"swiss-ai/apertus-8b"},"publicai/apertus-70b-instruct":{"id":"publicai/apertus-70b-instruct","name":"Apertus 70B Instruct","description":"Fully open 70B multilingual LLM supporting 1800+ languages with 65K context. Trained on 15T tokens of compliant open data. Apache 2.0, EU AI Act compliant.","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":8192},"cost":{"input":0.82,"output":2.92},"canonical_model_id":"swiss-ai/apertus-70b"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.1},"canonical_model_id":"google/gemini-2.5-flash-lite"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":5,"output":30},"canonical_model_id":"sakana/fugu-ultra"},"mistral/voxtral-small-latest":{"id":"mistral/voxtral-small-latest","name":"Voxtral Small (latest)","description":"Instruct model with native audio input for speech understanding and tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/voxtral-small-latest"}}},"orcarouter":{"id":"orcarouter","env":["ORCAROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.orcarouter.ai/v1","name":"OrcaRouter","doc":"https://docs.orcarouter.ai","models":{"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":10},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.442,"output":0.884,"cache_read":0.06},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.147,"output":0.295,"cache_read":0.02},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.442,"output":0.884,"cache_read":0.06},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.147,"output":0.295,"cache_read":0.02},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.147,"output":0.295,"cache_read":0.02},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-chat":{"id":"deepseek/deepseek-chat","name":"DeepSeek Chat","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.147,"output":0.295,"cache_read":0.02},"canonical_model_id":"deepseek/deepseek-chat"},"deepseek/deepseek-v4-flash-free":{"id":"deepseek/deepseek-v4-flash-free","name":"DeepSeek V4 Flash (free)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-reasoner":{"id":"deepseek/deepseek-reasoner","name":"DeepSeek Reasoner","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.147,"output":0.295,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-reasoner"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.075,"output":0.25},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-4.5-air":{"id":"z-ai/glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.2,"output":1.1,"cache_read":0.03,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-air"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.26,"output":3.96,"cache_read":0.234},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-5.3-flash-free":{"id":"z-ai/glm-5.3-flash-free","name":"GLM-5.3-Flash (free)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-4.6":{"id":"z-ai/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-4.5":{"id":"z-ai/glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"qwen/qwen3.7-flash":{"id":"qwen/qwen3.7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":131072},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.038},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen/qwen3-vl-235b-a22b-instruct":{"id":"qwen/qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.4,"output":1.6},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.115,"output":0.917},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3.5-plus":{"id":"qwen/qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.115,"output":0.688,"reasoning":2.4},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen/qwen3.6-35b-a3b":{"id":"qwen/qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.248,"output":1.485},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625,"tiers":[{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5}},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3.5-flash":{"id":"qwen/qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1,"output":0.4},"canonical_model_id":"alibaba/qwen3.5-flash"},"qwen/qwen3-vl-235b-a22b-thinking":{"id":"qwen/qwen3-vl-235b-a22b-thinking","name":"Qwen3 VL 235B A22B Thinking","description":"Qwen vision-language thinking model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":40960},"cost":{"input":0.4,"output":4},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-thinking"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.35,"output":1.42,"cache_read":0.071},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.086,"output":0.688},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.057,"output":0.459},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.33,"output":2.4},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.172,"output":1.032},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.25,"output":3.75,"cache_read":0.25,"cache_write":1.563},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.8-max":{"id":"qwen/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen/qwen3-max":{"id":"qwen/qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.359,"output":1.434},"canonical_model_id":"alibaba/qwen3-max"},"orcarouter/auto":{"id":"orcarouter/auto","name":"OrcaRouter Auto","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2026-05-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0,"output":0}},"orcarouter/fusion":{"id":"orcarouter/fusion","name":"OrcaRouter Fusion","description":"Curated fan-out router that runs Claude Opus 4.8, GPT-5.5 and Gemini 3.1 Pro in parallel on hard requests, then has a Claude Opus 4.8 judge return the strongest single answer verbatim. Easy requests fall through to a cheaper default and bill as one call.","family":"model-router","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000}},"orcarouter/fusion-mini":{"id":"orcarouter/fusion-mini","name":"OrcaRouter Fusion Mini","description":"Leaner two-model Fusion panel that runs Claude Opus 4.8 and GPT-5.5 in parallel on hard requests, then has a Claude Opus 4.8 judge return the strongest single answer verbatim. Easy requests fall through to a cheaper default and bill as one call.","family":"model-router","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000}},"orcarouter/free":{"id":"orcarouter/free","name":"OrcaRouter Free","description":"Built-in router over the free tier that scores each request's difficulty and sends light work to the smaller free model and harder work to the stronger one. Priced at zero and never falls back to a paid model.","family":"auto","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":0,"output":0}},"orcarouter/fusion-flash":{"id":"orcarouter/fusion-flash","name":"OrcaRouter Fusion Flash","description":"Budget Fusion panel that runs Gemini 3.5 Flash, MiniMax M2.7 and GLM 5.1 in parallel on hard requests, then has a Claude Opus 4.8 judge return the strongest single answer verbatim. Cost-sensitive fan-out over a 200K window.","family":"model-router","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000}},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2.5-highspeed":{"id":"minimax/minimax-m2.5-highspeed","name":"MiniMax-M2.5-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5-highspeed"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5"},"google/gemini-flash-lite-latest":{"id":"google/gemini-flash-lite-latest","name":"Gemini Flash-Lite Latest","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-flash-lite-latest"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"input_audio":2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":0.08333,"input_audio":3},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.06,"output":0.33,"cache_read":0.0075},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-robotics-er-1.6-preview":{"id":"google/gemini-robotics-er-1.6-preview","name":"Gemini Robotics-ER 1.6 Preview","description":"Vision-language model for embodied reasoning: spatial understanding, task planning, and physical-world agentic robotics","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-14","last_updated":"2026-04-14","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":1,"output":5},"canonical_model_id":"google/gemini-robotics-er-1.6-preview"},"google/gemini-flash-latest":{"id":"google/gemini-flash-latest","name":"Gemini Flash Latest","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.1},"canonical_model_id":"google/gemini-flash-latest"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.13,"output":0.38,"cache_read":0.02},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.1-flash-lite-preview":{"id":"google/gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0.18,"output":0.59,"cache_read":0.059},"canonical_model_id":"tencent/hy3"},"tencent/hy3-free":{"id":"tencent/hy3-free","name":"Hy3 (free)","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"tencent/hy3"},"kimi/kimi-k2.6":{"id":"kimi/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"kimi/kimi-k3":{"id":"kimi/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3.3,"output":16.5,"cache_read":0.33},"canonical_model_id":"moonshotai/kimi-k3"},"kimi/kimi-k2.7-code":{"id":"kimi/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"kimi/kimi-k2.5":{"id":"kimi/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.6,"output":3,"cache_read":0.1,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.5"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4":{"id":"openai/gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":30,"output":60},"canonical_model_id":"openai/gpt-4"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5-turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":0.5,"output":1.5,"cache_read":0},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-4o-2024-08-06":{"id":"openai/gpt-4o-2024-08-06","name":"GPT-4o (2024-08-06)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-08-06"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5.1-chat-latest":{"id":"openai/gpt-5.1-chat-latest","name":"GPT-5.1 Chat","description":"Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-chat-latest"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2 Pro","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/gpt-4o-2024-05-13":{"id":"openai/gpt-4o-2024-05-13","name":"GPT-4o (2024-05-13)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-05-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":5,"output":15},"canonical_model_id":"openai/gpt-4o-2024-05-13"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.2-chat-latest":{"id":"openai/gpt-5.2-chat-latest","name":"GPT-5.2 Chat","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-chat-latest"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.03,"output":0.17},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":1.5,"output":9,"cache_read":0.15},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":100000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-4o-2024-11-20":{"id":"openai/gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"openai/gpt-5-chat-latest":{"id":"openai/gpt-5-chat-latest","name":"GPT-5 Chat (latest)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":100000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5-chat-latest"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"grok/grok-4.3":{"id":"grok/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"grok/grok-4.5":{"id":"grok/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.5"},"grok/grok-4.6":{"id":"grok/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"}}},"privatemode-ai":{"id":"privatemode-ai","env":["PRIVATEMODE_API_KEY","PRIVATEMODE_ENDPOINT"],"npm":"@ai-sdk/openai-compatible","api":"http://localhost:8080/v1","name":"Privatemode AI","doc":"https://docs.privatemode.ai/api/overview","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"beta","cost":{"input":0.2311,"output":0.7511,"cache_read":0.0578},"canonical_model_id":"zhipuai/glm-5.3-flash"},"whisper-large-v3":{"id":"whisper-large-v3","type":"transcription","name":"Whisper large-v3","description":"Open Whisper checkpoint for robust multilingual transcription and captioning","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-10-01","last_updated":"2024-10-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":4096},"cost":{"input":0.01618,"output":0},"canonical_model_id":"openai/whisper-large-v3"},"voxtral-mini-3b":{"id":"voxtral-mini-3b","name":"Voxtral Mini 3B","description":"Speech-to-text model for audio transcription, translation, and audio understanding","family":"voxtral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07","last_updated":"2025-07","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000},"cost":{"input":0.00462,"output":0}},"qwen3-embedding-4b":{"id":"qwen3-embedding-4b","name":"Qwen3-Embedding 4B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-06","release_date":"2025-06-06","last_updated":"2025-06-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":2560},"cost":{"input":0.1502,"output":0}},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":262144},"status":"deprecated","cost":{"input":1.791,"output":8.9436,"cache_read":0.1733},"canonical_model_id":"moonshotai/kimi-k2.6"},"glm-flash-latest":{"id":"glm-flash-latest","name":"GLM Flash (latest)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"beta","cost":{"input":0.2311,"output":0.7511,"cache_read":0.0578},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"beta","cost":{"input":1.791,"output":6.6326,"cache_read":0.1733},"canonical_model_id":"zhipuai/glm-5.3"},"deepseek-ocr-2":{"id":"deepseek-ocr-2","name":"DeepSeek OCR 2","description":"High-accuracy OCR model for extracting text from documents, screenshots, receipts, and natural scenes","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"status":"beta","cost":{"input":0.8897,"output":1.4675,"cache_read":0.0924},"canonical_model_id":"deepseek/deepseek-ocr-2"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"gpt-oss-120b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.2311,"output":0.7511,"cache_read":0.0462},"canonical_model_id":"openai/gpt-oss-120b"},"kimi-latest":{"id":"kimi-latest","name":"Kimi (latest)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":262144},"status":"deprecated","cost":{"input":1.791,"output":8.9436,"cache_read":0.1733},"canonical_model_id":"moonshotai/kimi-k2.6"},"glm-latest":{"id":"glm-latest","name":"GLM (latest)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"beta","cost":{"input":1.791,"output":6.6326,"cache_read":0.1733},"canonical_model_id":"zhipuai/glm-5.3"}}},"nova":{"id":"nova","env":["NOVA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.nova.amazon.com/v1","name":"Nova","doc":"https://nova.amazon.com/dev/documentation","models":{"nova-2-pro-v1":{"id":"nova-2-pro-v1","name":"Nova 2 Pro","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-03","last_updated":"2026-01-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"reasoning":0}},"nova-2-lite-v1":{"id":"nova-2-lite-v1","name":"Nova 2 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"reasoning":0}}}},"unorouter":{"id":"unorouter","env":["UNOROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.unorouter.com/v1","name":"UnoRouter","doc":"https://unorouter.com/models","models":{"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1875,"output":1.125},"canonical_model_id":"openai/gpt-5.5"},"minimax-m2.7:free":{"id":"minimax-m2.7:free","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.819,"output":3.276},"canonical_model_id":"minimax/MiniMax-M2.7"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.8999,"output":1.7999},"canonical_model_id":"deepseek/deepseek-v4-pro"},"gemma-4-31b-it:free":{"id":"gemma-4-31b-it:free","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-31b-it"},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.05,"output":8.4},"canonical_model_id":"openai/gpt-5.2"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.2675,"output":5.3368},"canonical_model_id":"moonshotai/kimi-k2.6"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":1.8,"output":10.8},"canonical_model_id":"openai/gpt-5.4"},"gpt-5.5:free":{"id":"gpt-5.5:free","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.5"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1857,"output":1.1142},"canonical_model_id":"google/gemini-3.5-flash"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.425,"output":2.125},"canonical_model_id":"anthropic/claude-opus-4-8"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.0625,"output":0.125},"canonical_model_id":"deepseek/deepseek-v4-flash"},"glm-4.5-flash:free":{"id":"glm-4.5-flash:free","name":"GLM-4.5-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":98304},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.5-flash"},"deepseek-v4-pro:free":{"id":"deepseek-v4-pro:free","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-pro"},"nemotron-3-ultra-550b-a55b:free":{"id":"nemotron-3-ultra-550b-a55b:free","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.2,"output":6},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.6001,"output":5.0288},"canonical_model_id":"zhipuai/glm-5.2"},"gpt-5.4:free":{"id":"gpt-5.4:free","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.4"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":1.44,"output":7.2},"canonical_model_id":"anthropic/claude-sonnet-5"},"deepseek-v4-flash:free":{"id":"deepseek-v4-flash:free","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"glm-5.2:free":{"id":"glm-5.2:free","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.2"},"qwen3.5-397b-a17b:free":{"id":"qwen3.5-397b-a17b:free","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"step-3.7-flash:free":{"id":"step-3.7-flash:free","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0,"output":0},"canonical_model_id":"stepfun/step-3.7-flash"}}},"openrouter":{"id":"openrouter","env":["OPENROUTER_API_KEY"],"npm":"@openrouter/ai-sdk-provider","api":"https://openrouter.ai/api/v1","name":"OpenRouter","doc":"https://openrouter.ai/models","models":{"meituan/longcat-2.0":{"id":"meituan/longcat-2.0","name":"LongCat-2.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048756,"output":262144},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"meituan/longcat-2.0"},"~deepseek/deepseek-pro-latest":{"id":"~deepseek/deepseek-pro-latest","name":"DeepSeek Pro Latest","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-14","last_updated":"2026-09-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":393216},"cost":{"input":0.225,"output":8,"cache_read":0.23}},"~deepseek/deepseek-flash-latest":{"id":"~deepseek/deepseek-flash-latest","name":"DeepSeek Flash Latest","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-14","last_updated":"2026-09-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.07,"output":0.5,"cache_read":0.008}},"~deepseek/deepseek-v4-flash-latest":{"id":"~deepseek/deepseek-v4-flash-latest","name":"DeepSeek V4 Flash Latest","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-01","last_updated":"2026-08-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.01208,"output":0.12,"cache_read":0.00336}},"undi95/remm-slerp-l2-13b":{"id":"undi95/remm-slerp-l2-13b","name":"ReMM SLERP 13B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-06-30","release_date":"2023-07-22","last_updated":"2023-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":6144,"output":5529},"cost":{"input":0.35,"output":0.65},"canonical_model_id":"undi95/remm-slerp-l2-13b"},"aion-labs/aion-3.0-mini":{"id":"aion-labs/aion-3.0-mini","name":"Aion-3.0-Mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-14","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.7,"output":1.4,"cache_read":0.18},"canonical_model_id":"aion-labs/aion-3.0-mini"},"aion-labs/aion-rp-llama-3.1-8b":{"id":"aion-labs/aion-rp-llama-3.1-8b","name":"Aion-RP 1.0 (8B)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12-31","release_date":"2025-02-04","last_updated":"2025-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":29491},"cost":{"input":0.8,"output":1.6},"canonical_model_id":"aion-labs/aion-rp-llama-3.1-8b"},"aion-labs/aion-3.5":{"id":"aion-labs/aion-3.5","name":"Aion 3.5","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":3,"output":6,"cache_read":0.75}},"aion-labs/aion-3.5-mini":{"id":"aion-labs/aion-3.5-mini","name":"Aion 3.5 Mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.7,"output":1.4,"cache_read":0.18}},"aion-labs/aion-2.0":{"id":"aion-labs/aion-2.0","name":"Aion-2.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-21","last_updated":"2025-12-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.8,"output":1.6,"cache_read":0.2},"canonical_model_id":"aion-labs/aion-2.0"},"aion-labs/aion-3.0":{"id":"aion-labs/aion-3.0","name":"Aion-3.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-05","last_updated":"2026-05-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":3,"output":6,"cache_read":0.75},"canonical_model_id":"aion-labs/aion-3.0"},"inception/mercury-2":{"id":"inception/mercury-2","name":"Mercury 2","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":50000},"cost":{"input":0.25,"output":0.75,"cache_read":0.025},"canonical_model_id":"inception/mercury-2"},"inception/mercury-2.5":{"id":"inception/mercury-2.5","name":"Mercury 2.5","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":260000,"output":65536},"cost":{"input":0.04,"output":0.15,"cache_read":0.004},"canonical_model_id":"inception/mercury-2.5"},"inception/mercury-3":{"id":"inception/mercury-3","name":"Mercury 3","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1,"output":0.35,"cache_read":0.01}},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-opus-4.1":{"id":"anthropic/claude-opus-4.1","name":"Claude Opus 4.1 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"nex-agi/nex-n2.5-pro":{"id":"nex-agi/nex-n2.5-pro","name":"Nex-N2.5-Pro","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"agi","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.075,"output":0.25,"cache_read":0.015}},"nex-agi/nex-n2.5-mini":{"id":"nex-agi/nex-n2.5-mini","name":"Nex-N2.5-Mini","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"agi","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.025,"output":0.1,"cache_read":0.0025}},"amazon/nova-premier-v1":{"id":"amazon/nova-premier-v1","name":"Nova Premier 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2025-04-30","last_updated":"2025-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"status":"deprecated","cost":{"input":2.5,"output":12.5,"cache_read":0.625},"canonical_model_id":"amazon/nova-premier"},"amazon/nova-2-lite-v1":{"id":"amazon/nova-2-lite-v1","name":"Nova 2 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.3,"output":2.5},"canonical_model_id":"amazon/nova-2-lite-v1"},"amazon/nova-lite-v1":{"id":"amazon/nova-lite-v1","name":"Nova Lite 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":5120},"cost":{"input":0.06,"output":0.24},"canonical_model_id":"amazon/nova-lite-v1"},"amazon/nova-pro-v1":{"id":"amazon/nova-pro-v1","name":"Nova Pro 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":5120},"cost":{"input":0.8,"output":3.2},"canonical_model_id":"amazon/nova-pro-v1"},"amazon/nova-micro-v1":{"id":"amazon/nova-micro-v1","name":"Nova Micro 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":5120},"cost":{"input":0.035,"output":0.14},"canonical_model_id":"amazon/nova-micro-v1"},"~z-ai/glm-flash-latest":{"id":"~z-ai/glm-flash-latest","name":"GLM Flash Latest","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.037476,"output":0.16,"cache_read":0.006}},"~z-ai/glm-latest":{"id":"~z-ai/glm-latest","name":"GLM Latest","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-19","last_updated":"2026-08-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.079,"output":1.4,"cache_read":0.078}},"scaledown/summarize":{"id":"scaledown/summarize","name":"Summarize","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":4000},"cost":{"input":0.04,"output":0}},"scaledown/extract":{"id":"scaledown/extract","name":"Extract","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":900000},"cost":{"input":0.04,"output":0}},"anthracite-org/magnum-v4-72b":{"id":"anthracite-org/magnum-v4-72b","name":"Magnum v4 72B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":4096},"cost":{"input":2.5,"output":5},"canonical_model_id":"anthracite-org/magnum-v4-72b"},"mancer/weaver":{"id":"mancer/weaver","name":"Weaver (alpha)","description":"General-purpose chat model for instruction following, writing, and analysis","family":"alpha","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-06-30","release_date":"2023-08-02","last_updated":"2023-08-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":6000},"cost":{"input":0.4,"output":0.75},"canonical_model_id":"mancer/weaver"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.95526,"output":1.91052,"cache_read":0.079605},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-r1-0528":{"id":"deepseek/deepseek-r1-0528","name":"DeepSeek R1 0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.5,"output":2.15,"cache_read":0.35},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek/deepseek-v3.2-exp":{"id":"deepseek/deepseek-v3.2-exp","name":"DeepSeek V3.2 Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":147456},"cost":{"input":0.27,"output":0.41}},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-chat-v3.1":{"id":"deepseek/deepseek-chat-v3.1","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.25,"output":0.95,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-chat-v3.1"},"deepseek/deepseek-v3.1-terminus":{"id":"deepseek/deepseek-v3.1-terminus","name":"DeepSeek V3.1 Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-22","last_updated":"2025-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":147456},"cost":{"input":0.27,"output":1}},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0.2156,"output":0.6468,"cache_read":0.00686},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943717},"cost":{"input":0.091,"output":0.182,"cache_read":0.0182},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-chat":{"id":"deepseek/deepseek-chat","name":"DeepSeek Chat","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":16000},"cost":{"input":0.2574,"output":1.0287},"canonical_model_id":"deepseek/deepseek-chat"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":65536},"cost":{"input":0.28,"output":0.42,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-chat-v3-0324":{"id":"deepseek/deepseek-chat-v3-0324","name":"DeepSeek V3 0324","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07-31","release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":115200},"cost":{"input":0.29,"output":1.14,"cache_read":0.11},"canonical_model_id":"deepseek/deepseek-chat-v3-0324"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":0.056,"output":0.12,"cache_read":0.0016},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"microsoft/phi-4":{"id":"microsoft/phi-4","name":"Phi 4","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"phi","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2025-01-10","last_updated":"2025-01-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":14745},"cost":{"input":0.07,"output":0.14}},"gryphe/mythomax-l2-13b":{"id":"gryphe/mythomax-l2-13b","name":"MythoMax L2 13B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2023-08-10","last_updated":"2023-08-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":3686},"cost":{"input":0.08,"output":0.11},"canonical_model_id":"gryphe/mythomax-l2-13b"},"nvidia/nemotron-3.5-content-safety":{"id":"nvidia/nemotron-3.5-content-safety","name":"Nemotron 3.5 Content Safety","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.2,"output":0.2},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"nvidia/nemotron-3.5-lightning":{"id":"nvidia/nemotron-3.5-lightning","name":"Nemotron 3.5 Lightning 30B A3B","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.07,"output":0.2,"cache_read":0.04},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-3.5-super-vl":{"id":"nvidia/nemotron-3.5-super-vl","name":"Nemotron 3.5 Super VL","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":235929},"cost":{"input":0.15,"output":0.5,"cache_read":0.02},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/nemotron-3-ultra-550b-a55b":{"id":"nvidia/nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":182520},"cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/nemotron-3.5-super-vl:free":{"id":"nvidia/nemotron-3.5-super-vl:free","name":"Nemotron 3.5 Super VL (free)","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/nemotron-3.5-content-safety:free":{"id":"nvidia/nemotron-3.5-content-safety:free","name":"Nemotron 3.5 Content Safety (free)","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"nvidia/nemotron-3.5-lightning:free":{"id":"nvidia/nemotron-3.5-lightning:free","name":"Nemotron 3.5 Lightning (free)","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-3-nano-30b-a3b":{"id":"nvidia/nemotron-3-nano-30b-a3b","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.06,"output":0.24},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/nemotron-3-super-120b-a12b":{"id":"nvidia/nemotron-3-super-120b-a12b","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.08,"output":0.45},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/nemotron-3-ultra-550b-a55b:free":{"id":"nvidia/nemotron-3-ultra-550b-a55b:free","name":"Nemotron 3 Ultra (free)","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/nemotron-3-super-120b-a12b:free":{"id":"nvidia/nemotron-3-super-120b-a12b:free","name":"Nemotron 3 Super (free)","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free":{"id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free","name":"Nemotron 3 Nano Omni (free)","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"z-ai/glm-4.6v":{"id":"z-ai/glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.055},"canonical_model_id":"zhipuai/glm-4.6v"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":0.037476,"output":0.16,"cache_read":0.006},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.3-prime":{"id":"z-ai/glm-5.3-prime","name":"GLM 5.3 Prime","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.8,"output":8.8,"cache_read":0.56}},"z-ai/glm-4.5v":{"id":"z-ai/glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":0.6,"output":1.8,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5v"},"z-ai/glm-4.5-air":{"id":"z-ai/glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.13,"output":0.85,"cache_read":0.025},"canonical_model_id":"zhipuai/glm-4.5-air"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":0.079,"output":1.4,"cache_read":0.078},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.54,"output":1.98,"cache_read":0.099},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128000},"cost":{"input":0.966,"output":3.036,"cache_read":0.1794},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-4.7-flash":{"id":"z-ai/glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":117964},"cost":{"input":0.0605,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128000},"cost":{"input":0.6,"output":1.92,"cache_read":0.12},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-5v-turbo":{"id":"z-ai/glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"z-ai/glm-5.3-flashx":{"id":"z-ai/glm-5.3-flashx","name":"GLM 5.3 FlashX","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.37,"output":1.25,"cache_read":0.09}},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"z-ai/glm-4.6":{"id":"z-ai/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.55,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":0.62,"output":2.42,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-4.5":{"id":"z-ai/glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"status":"deprecated","cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5"},"inclusionai/ling-3.0-flash-vl":{"id":"inclusionai/ling-3.0-flash-vl","name":"Ling 3.0 Flash VL","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012},"canonical_model_id":"inclusionai/ling-3.0-flash-vl"},"inclusionai/ling-3.0-flash":{"id":"inclusionai/ling-3.0-flash","name":"Ling 3.0 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.021,"output":0.063,"cache_read":0.0042},"canonical_model_id":"inclusionai/ling-3.0-flash"},"inclusionai/ling-3.1-flash":{"id":"inclusionai/ling-3.1-flash","name":"Ling 3.1 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.075,"output":0.225,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.1-flash"},"meta/muse-spark-1.3-contributor":{"id":"meta/muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/muse-wonder":{"id":"meta/muse-wonder","name":"Muse Wonder","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":4,"output":20,"cache_read":0.2}},"meta/muse-glimmer-30b":{"id":"meta/muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.3,"output":1.2,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"meta/muse-spark-1.2-contributor":{"id":"meta/muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-08-21","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-open":{"id":"meta/muse-spark-open","name":"Muse Spark Open","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"bytedance-seed/seed-1.6-flash":{"id":"bytedance-seed/seed-1.6-flash","name":"Seed 1.6 Flash","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.075,"output":0.3,"tiers":[{"input":0.1,"output":0.8,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-6-flash"},"bytedance-seed/seed-2-1-turbo":{"id":"bytedance-seed/seed-2-1-turbo","name":"Seed 2.1 Turbo","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0.5,"output":2.5},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"bytedance-seed/seed-2.0-mini":{"id":"bytedance-seed/seed-2.0-mini","name":"Seed 2.0 Mini","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.1,"output":0.4,"tiers":[{"input":0.2,"output":0.8,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"bytedance-seed/seed-1.6":{"id":"bytedance-seed/seed-1.6","name":"Seed 1.6","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.25,"output":2,"tiers":[{"input":0.5,"output":4,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-6"},"bytedance-seed/seed-2.0-code":{"id":"bytedance-seed/seed-2.0-code","name":"Seed 2.0 Code","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"status":"deprecated","cost":{"input":0.5,"output":3,"tiers":[{"input":1,"output":6,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"bytedance-seed/seed-2.0-lite":{"id":"bytedance-seed/seed-2.0-lite","name":"Seed 2.0 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.25,"output":2,"tiers":[{"input":0.5,"output":4,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"meta-llama/llama-3.1-8b-instruct":{"id":"meta-llama/llama-3.1-8b-instruct","name":"Llama-3.1-8B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.05,"output":0.08,"cache_read":0.025},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta-llama/llama-3.2-3b-instruct":{"id":"meta-llama/llama-3.2-3b-instruct","name":"Llama 3.2 3B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":72000},"cost":{"input":0.0509,"output":0.335}},"meta-llama/llama-3.3-70b-instruct":{"id":"meta-llama/llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.1,"output":0.32},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/llama-4-maverick":{"id":"meta-llama/llama-4-maverick","name":"Llama 4 Maverick","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":16384},"cost":{"input":0.1875,"output":0.6525,"cache_read":0.05},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"meta-llama/llama-4-scout":{"id":"meta-llama/llama-4-scout","name":"Llama 4 Scout","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1310720,"output":16384},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"meta-llama/llama-3.2-1b-instruct":{"id":"meta-llama/llama-3.2-1b-instruct","name":"Llama 3.2 1B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":60000,"output":54000},"cost":{"input":0.027,"output":0.201},"canonical_model_id":"meta/llama-3.2-1b-instruct"},"meta-llama/llama-guard-4-12b":{"id":"meta-llama/llama-guard-4-12b","name":"Llama Guard 4 12B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-04-30","last_updated":"2025-04-30","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":16384},"cost":{"input":0.18,"output":0.18}},"meta-llama/llama-3.1-70b-instruct":{"id":"meta-llama/llama-3.1-70b-instruct","name":"Llama-3.1-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.4,"output":0.4},"canonical_model_id":"meta/llama-3.1-70b-instruct"},"qwen/qwen3.5-plus-02-15":{"id":"qwen/qwen3.5-plus-02-15","name":"Qwen3.5 Plus 2026-02-15","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.26,"output":1.56,"tiers":[{"input":0.325,"output":1.95,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.325,"output":1.95}}},"qwen/qwen3-235b-a22b-thinking-2507":{"id":"qwen/qwen3-235b-a22b-thinking-2507","name":"Qwen3 235B A22B Thinking 2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-06-30","release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.45,"output":3.5}},"qwen/qwen3.7-flash":{"id":"qwen/qwen3.7-flash","name":"Qwen3.7 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":65536},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.038,"tiers":[{"input":0.1,"output":0.4,"cache_read":0.02,"cache_write":0.125,"tier":{"type":"context","size":32000}},{"input":0.2,"output":0.8,"cache_read":0.04,"cache_write":0.25,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen/qwen3-vl-30b-a3b-instruct":{"id":"qwen/qwen3-vl-30b-a3b-instruct","name":"Qwen3 VL 30B A3B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"qwen/qwen3.5-9b":{"id":"qwen/qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.1,"output":0.15},"canonical_model_id":"alibaba/qwen3.5-9b"},"qwen/qwen-2.5-7b-instruct":{"id":"qwen/qwen-2.5-7b-instruct","name":"Qwen2.5 7B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-10-16","last_updated":"2024-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":29491},"cost":{"input":0.1,"output":0.2}},"qwen/qwen3-vl-235b-a22b-instruct":{"id":"qwen/qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.21,"output":1.9,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"qwen/qwen-plus":{"id":"qwen/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.26,"output":0.78,"cache_read":0.052,"cache_write":0.325,"tiers":[{"input":0.78,"output":2.34,"cache_read":0.156,"cache_write":0.975,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.78,"output":2.34,"cache_read":0.156,"cache_write":0.975}},"canonical_model_id":"alibaba/qwen-plus"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.26,"output":2.08},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3.6-35b-a3b":{"id":"qwen/qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.1,"output":0.95,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3.8-omni-flash":{"id":"qwen/qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"qwen/qwen3-235b-a22b-2507":{"id":"qwen/qwen3-235b-a22b-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.09,"output":0.55},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.325,"output":1.95,"cache_write":0.40625,"tiers":[{"input":1.3,"output":3.9,"cache_write":1.625,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.3,"output":3.9,"cache_write":1.625}},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1875,"output":1.125,"cache_write":0.234375,"tiers":[{"input":0.75,"output":3,"cache_write":0.9375,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.75,"output":3,"cache_write":0.9375}},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3-coder-flash":{"id":"qwen/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.195,"output":0.975,"cache_read":0.039,"cache_write":0.24375,"tiers":[{"input":0.325,"output":1.625,"cache_read":0.065,"cache_write":0.40625,"tier":{"type":"context","size":32000}},{"input":0.52,"output":2.6,"cache_read":0.104,"cache_write":0.65,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3-coder-flash"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3.8-max-prime":{"id":"qwen/qwen3.8-max-prime","name":"Qwen 3.8 Max Prime","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.8-max","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":4,"output":12,"cache_read":0.5},"canonical_model_id":"alibaba/qwen3.8-max-prime"},"qwen/qwen-2.5-coder-32b-instruct":{"id":"qwen/qwen-2.5-coder-32b-instruct","name":"Qwen2.5-Coder-32B-Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":29491},"cost":{"input":0.66,"output":1},"canonical_model_id":"alibaba/qwen2.5-coder-32b-instruct"},"qwen/qwen3-30b-a3b":{"id":"qwen/qwen3-30b-a3b","name":"Qwen3 30B A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.5},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"qwen/qwen3.8-2.4t-a95b":{"id":"qwen/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.32,"output":1.28,"cache_read":0.064,"cache_write":0.4,"tiers":[{"input":0.96,"output":3.84,"cache_read":0.192,"cache_write":1.2,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.96,"output":3.84,"cache_read":0.192,"cache_write":1.2}},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3-vl-30b-a3b-thinking":{"id":"qwen/qwen3-vl-30b-a3b-thinking","name":"Qwen3 VL 30B A3B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.29,"output":1}},"qwen/qwen3-next-80b-a3b-instruct":{"id":"qwen/qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.09,"output":1.1},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.27,"output":2.16,"cache_read":0.27},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3-next-80b-a3b-thinking":{"id":"qwen/qwen3-next-80b-a3b-thinking","name":"Qwen3-Next 80B-A3B (Thinking)","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"qwen/qwen3.5-plus-20260420":{"id":"qwen/qwen3.5-plus-20260420","name":"Qwen3.5 Plus 2026-04-20","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":1.8,"cache_write":0.375,"tiers":[{"input":0.375,"output":2.25,"cache_write":0.46875,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.375,"output":2.25,"cache_write":0.46875}}},"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.15,"output":1,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3-30b-a3b-instruct-2507":{"id":"qwen/qwen3-30b-a3b-instruct-2507","name":"Qwen3 30B A3B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-30","release_date":"2025-07-29","last_updated":"2025-07-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.09,"output":0.3}},"qwen/qwen3-32b":{"id":"qwen/qwen3-32b","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.08,"output":0.28},"canonical_model_id":"alibaba/qwen3-32b"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0.04,"output":1.35,"cache_read":0.018},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3-coder-30b-a3b-instruct":{"id":"qwen/qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.07,"output":0.28},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.39,"output":2.34},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3-vl-8b-instruct":{"id":"qwen/qwen3-vl-8b-instruct","name":"Qwen3 VL 8B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-14","last_updated":"2025-10-14","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.25,"output":0.75,"cache_read":0.12}},"qwen/qwen3-coder-next":{"id":"qwen/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.12,"output":0.8,"cache_read":0.07},"canonical_model_id":"alibaba/qwen3-coder-next"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"alibaba/qwen3.6-27b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.475,"output":4.425,"cache_read":0.295,"cache_write":1.84375},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3-coder":{"id":"qwen/qwen3-coder","name":"Qwen3 Coder 480B A35B","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":1,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"qwen/qwen2.5-vl-72b-instruct":{"id":"qwen/qwen2.5-vl-72b-instruct","name":"Qwen2.5-VL 72B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-01-26","last_updated":"2025-01-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":115200},"cost":{"input":0.8,"output":1,"cache_read":0.4},"canonical_model_id":"alibaba/qwen2-5-vl-72b-instruct"},"qwen/qwen-2.5-72b-instruct":{"id":"qwen/qwen-2.5-72b-instruct","name":"Qwen2.5 72B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-09-19","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.36,"output":0.4}},"qwen/qwen3-14b":{"id":"qwen/qwen3-14b","name":"Qwen3 14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.24}},"qwen/qwen3.8-max-0902":{"id":"qwen/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"qwen/qwen3.5-flash-02-23":{"id":"qwen/qwen3.5-flash-02-23","name":"Qwen3.5-Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-25","last_updated":"2026-02-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.065,"output":0.26}},"inference-net/schematron-v2-small":{"id":"inference-net/schematron-v2-small","name":"Schematron V2 Small","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":0.05,"output":0.23,"cache_read":0.05},"canonical_model_id":"inference-net/schematron-v2-small"},"inference-net/schematron-v2-turbo":{"id":"inference-net/schematron-v2-turbo","name":"Schematron V2 Turbo","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.03,"output":0.15,"cache_read":0.03},"canonical_model_id":"inference-net/schematron-v2-turbo"},"~x-ai/grok-latest":{"id":"~x-ai/grok-latest","name":"Grok Latest","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}}},"~unbiased/pareto-latest":{"id":"~unbiased/pareto-latest","name":"Pareto Latest","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.8,"output":3.2,"cache_read":0.03}},"thedrummer/cydonia-24b-v4.1":{"id":"thedrummer/cydonia-24b-v4.1","name":"Cydonia 24B V4.1","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04-30","release_date":"2025-09-27","last_updated":"2025-09-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.3,"output":0.5,"cache_read":0.15},"canonical_model_id":"thedrummer/cydonia-24b-v4.1"},"thedrummer/unslopnemo-12b":{"id":"thedrummer/unslopnemo-12b","name":"UnslopNemo 12B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04-30","release_date":"2024-11-08","last_updated":"2024-11-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":819200},"cost":{"input":0.4,"output":0.4},"canonical_model_id":"thedrummer/unslopnemo-12b"},"thedrummer/skyfall-36b-v2":{"id":"thedrummer/skyfall-36b-v2","name":"Skyfall 36B V2","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2025-03-10","last_updated":"2025-03-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":29491},"cost":{"input":0.55,"output":0.8,"cache_read":0.25},"canonical_model_id":"thedrummer/skyfall-36b-v2"},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 Instruct 0905","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.6,"output":2.5},"canonical_model_id":"moonshotai/kimi-k2-0905"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.65,"output":3.41,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":0.239,"output":11.936,"cache_read":0.21},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2":{"id":"moonshotai/kimi-k2","name":"Kimi K2 0711","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-12-31","release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.57,"output":2.3},"canonical_model_id":"moonshotai/kimi-k2"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.6712,"output":3.35,"cache_read":0.18},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.5,"output":2.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.6,"output":2.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"writer/palmyra-x5":{"id":"writer/palmyra-x5","name":"Palmyra X5","description":"General-purpose chat model for instruction following, writing, and analysis","family":"palmyra","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1040000,"output":8192},"cost":{"input":0.6,"output":6},"canonical_model_id":"writer/palmyra-x5"},"xingchen-agi/xing4.0-29b-a4b:free":{"id":"xingchen-agi/xing4.0-29b-a4b:free","name":"Xing4.0 29B A4B (free)","description":"Free provider route for experiments, demos, and cost-sensitive chat workloads","family":"agi","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0,"output":0}},"poolside/laguna-s-2.1:free":{"id":"poolside/laguna-s-2.1:free","name":"Laguna S 2.1 (free)","description":"Free provider route for experiments, demos, and cost-sensitive chat workloads","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-s-2.1"},"poolside/laguna-xs-2.1:free":{"id":"poolside/laguna-xs-2.1:free","name":"Laguna XS 2.1 (free)","description":"Free provider route for experiments, demos, and cost-sensitive chat workloads","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-02","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-xs-2.1"},"poolside/laguna-xs-2.1":{"id":"poolside/laguna-xs-2.1","name":"Laguna XS 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-02","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.06,"output":0.12,"cache_read":0.03},"canonical_model_id":"poolside/laguna-xs-2.1"},"poolside/laguna-s-2.1":{"id":"poolside/laguna-s-2.1","name":"Laguna S 2.1","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.09,"output":0.18,"cache_read":0.009},"canonical_model_id":"poolside/laguna-s-2.1"},"fireworks/ember-1":{"id":"fireworks/ember-1","name":"Ember-1","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"fireworks/ember-1"},"bytedance/ui-tars-1.5-7b":{"id":"bytedance/ui-tars-1.5-7b","name":"UI-TARS 7B ","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01-31","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":2048},"cost":{"input":0.1,"output":0.2,"cache_read":0.1}},"sao10k/l3-lunaris-8b":{"id":"sao10k/l3-lunaris-8b","name":"Llama 3 8B Lunaris","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-13","last_updated":"2024-08-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":7372},"cost":{"input":0.04,"output":0.05},"canonical_model_id":"sao10k/l3-lunaris-8b"},"sao10k/l3.1-euryale-70b":{"id":"sao10k/l3.1-euryale-70b","name":"Llama 3.1 Euryale 70B v2.2","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-28","last_updated":"2024-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.85,"output":0.85},"canonical_model_id":"sao10k/l3.1-euryale-70b"},"dots-studio/dots-3-note-preview:free":{"id":"dots-studio/dots-3-note-preview:free","name":"Dots3-Note Preview (free)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":512000,"output":460800},"status":"deprecated","cost":{"input":0,"output":0}},"~moonshotai/kimi-latest":{"id":"~moonshotai/kimi-latest","name":"Kimi Latest","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.238,"output":16,"cache_read":0.228}},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":471859},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/inkling-small:free":{"id":"thinkingmachines/inkling-small:free","name":"Inkling Small (free)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":262144},"cost":{"input":0.45,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/inkling:free":{"id":"thinkingmachines/inkling:free","name":"Inkling (free)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"thinkingmachines/inkling"},"reflection/beam-1.0-501b-a23b":{"id":"reflection/beam-1.0-501b-a23b","name":"Beam 1.0 501B A23B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":0.3,"output":1,"cache_read":0.04},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"stepfun/step-3.7-flash":{"id":"stepfun/step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":256000,"output":230400},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"stepfun/step-5-preview":{"id":"stepfun/step-5-preview","name":"Step 5 Preview","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1000000,"output":64000},"cost":{"input":1,"output":2.7,"cache_read":0.1},"canonical_model_id":"stepfun/step-5-preview"},"stepfun/step-3.5-flash":{"id":"stepfun/step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"stepfun/step-3.5-flash"},"perceptron/perceptron-mk1":{"id":"perceptron/perceptron-mk1","name":"Perceptron Mk1","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":0.15,"output":1.5},"canonical_model_id":"perceptron/perceptron-mk1"},"perceptron/perceptron-mk1.5":{"id":"perceptron/perceptron-mk1.5","name":"Perceptron Mk1.5","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":36864,"output":8192},"cost":{"input":0.15,"output":1.5}},"openrouter/pareto-code":{"id":"openrouter/pareto-code","name":"Pareto Code Router","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":200000}},"openrouter/auto":{"id":"openrouter/auto","name":"Auto Router","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"auto","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2023-11-08","last_updated":"2023-11-08","modalities":{"input":["text","image","audio","pdf","video"],"output":["text","image"]},"open_weights":false,"limit":{"context":2000000,"output":2000000}},"openrouter/fusion":{"id":"openrouter/fusion","name":"Fusion","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000}},"openrouter/bodybuilder":{"id":"openrouter/bodybuilder","name":"Body Builder (beta)","description":"Preview model for early access evaluation, prototyping, and compatibility testing","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2025-12-05","last_updated":"2025-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000}},"openrouter/free":{"id":"openrouter/free","name":"Free Models Router","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-01","last_updated":"2026-02-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":8000},"cost":{"input":0,"output":0}},"mistralai/mistral-small-3.1-24b-instruct":{"id":"mistralai/mistral-small-3.1-24b-instruct","name":"Mistral Small 3.1 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":102400},"cost":{"input":0.351,"output":0.555},"canonical_model_id":"mistral/mistral-small-3.1-24b-instruct"},"mistralai/mistral-medium-3":{"id":"mistralai/mistral-medium-3","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/mistral-medium-3"},"mistralai/mistral-nemo":{"id":"mistralai/mistral-nemo","name":"Mistral Nemo","description":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.029,"output":0.03},"canonical_model_id":"mistral/mistral-nemo"},"mistralai/mistral-large":{"id":"mistralai/mistral-large","name":"Mistral Large","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11-30","release_date":"2024-02-26","last_updated":"2024-02-26","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":102400},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mistral-large"},"mistralai/mistral-small-3.2-24b-instruct":{"id":"mistralai/mistral-small-3.2-24b-instruct","name":"Mistral Small 3.2 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.09375,"output":0.25},"canonical_model_id":"mistral/mistral-small-3.2-24b-instruct"},"mistralai/mistral-saba":{"id":"mistralai/mistral-saba","name":"Saba","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-02-17","last_updated":"2025-02-17","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":26214},"cost":{"input":0.2,"output":0.6,"cache_read":0.02},"canonical_model_id":"mistral/mistral-saba"},"mistralai/ministral-8b-2512":{"id":"mistralai/ministral-8b-2512","name":"Ministral 3 8B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.15,"output":0.15,"cache_read":0.015},"canonical_model_id":"mistral/ministral-8b-2512"},"mistralai/devstral-2512":{"id":"mistralai/devstral-2512","name":"Devstral 2","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/devstral-2512"},"mistralai/mistral-medium-3.1":{"id":"mistralai/mistral-medium-3.1","name":"Mistral Medium 3.1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-30","release_date":"2025-08-12","last_updated":"2025-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/mistral-medium-3.1"},"mistralai/ministral-3b-2512":{"id":"mistralai/ministral-3b-2512","name":"Ministral 3 3B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":104857},"cost":{"input":0.1,"output":0.1,"cache_read":0.01},"canonical_model_id":"mistral/ministral-3b-2512"},"mistralai/mixtral-8x22b-instruct":{"id":"mistralai/mixtral-8x22b-instruct","name":"Mixtral 8x22B Instruct","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-01-31","release_date":"2024-04-17","last_updated":"2024-04-17","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":52428},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mixtral-8x22b-instruct"},"mistralai/mistral-medium-3-5":{"id":"mistralai/mistral-medium-3-5","name":"Mistral Medium 3.5","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":209715},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"mistral/mistral-medium-3-5"},"mistralai/mistral-large-2512":{"id":"mistralai/mistral-large-2512","name":"Mistral Large 3","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"mistral/mistral-large-2512"},"mistralai/mistral-small-2603":{"id":"mistralai/mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"mistral/mistral-small-2603"},"mistralai/mistral-large-4-0":{"id":"mistralai/mistral-large-4-0","name":"Mistral Large 4","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":262144},"cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"mistralai/voxtral-small-24b-2507":{"id":"mistralai/voxtral-small-24b-2507","name":"Voxtral Small 24B 2507","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":26214},"cost":{"input":0.1,"output":0.3,"cache_read":0.01},"canonical_model_id":"mistral/voxtral-small-24b-2507"},"mistralai/ministral-14b-2512":{"id":"mistralai/ministral-14b-2512","name":"Ministral 3 14B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.2,"output":0.2,"cache_read":0.02},"canonical_model_id":"mistral/ministral-14b-2512"},"mistralai/mistral-large-2407":{"id":"mistralai/mistral-large-2407","name":"Mistral Large 2407","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-03-31","release_date":"2024-07-24","last_updated":"2024-07-24","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mistral-large-2407"},"mistralai/codestral-2508":{"id":"mistralai/codestral-2508","name":"Codestral 25.08","description":"Mistral coding model for code completion, generation, and developer workflows","family":"codestral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":204800},"cost":{"input":0.3,"output":0.9,"cache_read":0.03},"canonical_model_id":"mistral/codestral-2508"},"mistralai/mistral-small-24b-instruct-2501":{"id":"mistralai/mistral-small-24b-instruct-2501","name":"Mistral Small 3","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-01-30","last_updated":"2025-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.05,"output":0.08},"canonical_model_id":"mistral/mistral-small-24b-instruct-2501"},"relace/relace-apply-3":{"id":"relace/relace-apply-3","name":"Relace Apply 3","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2025-09-26","last_updated":"2025-09-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.85,"output":1.25},"canonical_model_id":"relace/relace-apply-3"},"relace/relace-search":{"id":"relace/relace-search","name":"Relace Search","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":1,"output":3},"canonical_model_id":"relace/relace-search"},"minimax/minimax-m2-her":{"id":"minimax/minimax-m2-her","name":"MiniMax-M2 Her","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-01-23","last_updated":"2026-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":2048},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2-Her"},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.255,"output":1.02},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-m1":{"id":"minimax/minimax-m1","name":"MiniMax M1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06-30","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":40000},"cost":{"input":0.55,"output":2.2}},"minimax/minimax-m3.1-flash:free":{"id":"minimax/minimax-m3.1-flash:free","name":"MiniMax M3.1 Flash (free)","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0,"output":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax-M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-01":{"id":"minimax/minimax-01","name":"MiniMax-01","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-03-31","release_date":"2025-01-15","last_updated":"2025-01-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000192,"output":40000},"cost":{"input":0.2,"output":1.1}},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":32768},"cost":{"input":0.27,"output":0.95,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"prism-ml/ternary-bonsai-2-27b":{"id":"prism-ml/ternary-bonsai-2-27b","name":"Ternary Bonsai 2 27B","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.075,"output":0.5,"cache_read":0.0375}},"~google/gemini-pro-latest":{"id":"~google/gemini-pro-latest","name":"Gemini Pro Latest","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["audio","pdf","image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":2,"output":10,"reasoning":10,"cache_read":0.1,"cache_write":0.1875,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}}},"~google/gemini-flash-latest":{"id":"~google/gemini-flash-latest","name":"Gemini Flash Latest","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667}},"~anthropic/claude-opus-latest":{"id":"~anthropic/claude-opus-latest","name":"Claude Opus Latest","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5}},"~anthropic/claude-haiku-latest":{"id":"~anthropic/claude-haiku-latest","name":"Claude Haiku Latest","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]}},"~anthropic/claude-fable-latest":{"id":"~anthropic/claude-fable-latest","name":"Claude Fable Latest","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5}},"~anthropic/claude-sonnet-latest":{"id":"~anthropic/claude-sonnet-latest","name":"Claude Sonnet Latest","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5}},"elorian/glimpse":{"id":"elorian/glimpse","name":"Glimpse","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":256000},"cost":{"input":0.6,"output":3,"cache_read":0.06}},"liquid/lfm-2.5-2.6b:free":{"id":"liquid/lfm-2.5-2.6b:free","name":"LFM2.5-2.6B (free)","description":"Free provider route for experiments, demos, and cost-sensitive chat workloads","family":"liquid","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":8192},"cost":{"input":0,"output":0}},"liquid/lfm-2.5-24b-a2b":{"id":"liquid/lfm-2.5-24b-a2b","name":"LFM2.5-24B-A2B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"liquid","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.05,"output":0.7,"cache_read":0.025}},"nousresearch/hermes-3-llama-3.1-405b":{"id":"nousresearch/hermes-3-llama-3.1-405b","name":"Hermes 3 405B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-16","last_updated":"2024-08-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":1,"output":1},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-405b"},"nousresearch/hermes-4-405b":{"id":"nousresearch/hermes-4-405b","name":"Hermes 4 405B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"hermes","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":1,"output":3},"canonical_model_id":"nousresearch/hermes-4-405b"},"nousresearch/hermes-3-llama-3.1-70b":{"id":"nousresearch/hermes-3-llama-3.1-70b","name":"Hermes 3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-18","last_updated":"2024-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.7,"output":0.7},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-70b"},"~openai/gpt-luna-latest":{"id":"~openai/gpt-luna-latest","name":"GPT Luna Latest","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}}},"~openai/gpt-sol-latest":{"id":"~openai/gpt-sol-latest","name":"GPT Sol Latest","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}}},"~openai/gpt-astra-latest":{"id":"~openai/gpt-astra-latest","name":"GPT Astra Latest","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}}},"~openai/gpt-mini-latest":{"id":"~openai/gpt-mini-latest","name":"GPT Mini Latest","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075}},"~openai/gpt-terra-latest":{"id":"~openai/gpt-terra-latest","name":"GPT Terra Latest","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}}},"upstage/solar-pro-3":{"id":"upstage/solar-pro-3","name":"Solar Pro 3","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07","release_date":"2026-01-26","last_updated":"2026-03-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":117964},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"upstage/solar-pro3"},"upstage/solar-pro4":{"id":"upstage/solar-pro4","name":"Solar Pro 4","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.15,"output":0.6,"cache_read":0.03},"canonical_model_id":"upstage/solar-pro4"},"upstage/solar-mini4":{"id":"upstage/solar-mini4","name":"Solar Mini 4","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"solar-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.05,"output":0.2,"cache_read":0.005},"canonical_model_id":"upstage/solar-mini4"},"ibm-granite/granite-4.0-h-micro":{"id":"ibm-granite/granite-4.0-h-micro","name":"Granite 4.0 Micro","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"granite","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-10-20","last_updated":"2025-10-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":117900},"cost":{"input":0.017,"output":0.112}},"ibm-granite/granite-4.2-8b":{"id":"ibm-granite/granite-4.2-8b","name":"Granite 4.2 8B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"granite","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-25","last_updated":"2026-08-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.06,"output":0.25,"cache_read":0.015},"canonical_model_id":"ibm/granite-4.2-8b"},"cohere/command-a":{"id":"cohere/command-a","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a"},"cohere/command-r-08-2024":{"id":"cohere/command-r-08-2024","name":"Command R","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"cohere/command-r-08-2024"},"cohere/command-a-plus":{"id":"cohere/command-a-plus","name":"Command A+","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04-01","release_date":"2026-05-20","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":192000,"output":64000},"cost":{"input":0.3,"output":1.5,"cache_read":0.15},"canonical_model_id":"cohere/command-a-plus-05-2026"},"cohere/command-r7b-12-2024":{"id":"cohere/command-r7b-12-2024","name":"Command R7B","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-12-13","last_updated":"2024-12-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.0375,"output":0.15},"canonical_model_id":"cohere/command-r7b-12-2024"},"cohere/command-r-plus-08-2024":{"id":"cohere/command-r-plus-08-2024","name":"Command R+","description":"Cohere's RAG workhorse for long-context enterprise search and tool use","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-r-plus-08-2024"},"cohere/north-mini-code:free":{"id":"cohere/north-mini-code:free","name":"North Mini Code (free)","description":"Cohere coding model for practical software engineering and agentic edits","family":"north","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09-23","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0,"output":0},"canonical_model_id":"cohere/north-mini-code-1-0"},"perplexity/sonar-deep-research":{"id":"perplexity/sonar-deep-research","name":"Sonar Deep Research","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar-deep-research","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-07","last_updated":"2025-03-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":115200},"cost":{"input":2,"output":8,"reasoning":3}},"perplexity/sonar-pro":{"id":"perplexity/sonar-pro","name":"Sonar Pro","description":"Advanced Sonar search model for deeper research and cited synthesis","family":"sonar-pro","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-07","last_updated":"2025-03-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8000},"cost":{"input":3,"output":15}},"perplexity/sonar":{"id":"perplexity/sonar","name":"Sonar","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-01-27","last_updated":"2025-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":127072,"output":114364},"cost":{"input":1,"output":1}},"perplexity/sonar-reasoning-pro":{"id":"perplexity/sonar-reasoning-pro","name":"Sonar Reasoning Pro","description":"Web-grounded reasoning model for multi-step research and cited answers","family":"sonar-reasoning","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-07","last_updated":"2025-03-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":115200},"cost":{"input":2,"output":8}},"perplexity/sonar-pro-search":{"id":"perplexity/sonar-pro-search","name":"Sonar Pro Search","description":"Advanced Sonar search model for deeper research and cited synthesis","family":"sonar-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-10-30","last_updated":"2025-10-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8000},"cost":{"input":3,"output":15}},"xiaomi/mimo-v2.5-pro":{"id":"xiaomi/mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":131072},"status":"deprecated","cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.6-pro-ultraspeed":{"id":"xiaomi/mimo-v2.6-pro-ultraspeed","name":"MiMo-V2.6-Pro-UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.35,"output":8.7,"cache_read":0.036},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":131072},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2.6-flash":{"id":"xiaomi/mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"unbiased/pareto-26.9":{"id":"unbiased/pareto-26.9","name":"Pareto 26.9","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.25}},"unbiased/pareto-26.10":{"id":"unbiased/pareto-26.10","name":"Pareto 26.10","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.8,"output":3.2,"cache_read":0.03}},"moondream/moondream3-preview":{"id":"moondream/moondream3-preview","name":"Moondream 3 Preview","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.3,"output":2.5}},"moondream/moondream3.1-9b-a2b":{"id":"moondream/moondream3.1-9b-a2b","name":"Moondream 3.1","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.3,"output":2.5}},"x-ai/grok-4.20-multi-agent":{"id":"x-ai/grok-4.20-multi-agent","name":"Grok 4.20 Multi-Agent","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-09-01","release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":1800000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":900000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"x-ai/grok-4.7":{"id":"x-ai/grok-4.7","name":"Grok 4.7","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"x-ai/grok-build-0.1":{"id":"x-ai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":230400},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}},"canonical_model_id":"xai/grok-build-0.1"},"x-ai/grok-4.5":{"id":"x-ai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"x-ai/grok-4.20":{"id":"x-ai/grok-4.20","name":"Grok 4.20","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09-01","release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":1800000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"google/gemma-4-26b-a4b-it:free":{"id":"google/gemma-4-26b-a4b-it:free","name":"Gemma 4 26B A4B (free)","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-3.1-flash-image":{"id":"google/gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["image","text"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image"},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.05,"output":0.15},"canonical_model_id":"google/gemma-3-12b-it"},"google/gemini-3-pro-image-preview":{"id":"google/gemini-3-pro-image-preview","type":"image","name":"Nano Banana Pro Preview","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3-pro-image-preview"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"reasoning":3,"cache_read":0.05,"cache_write":0.083333},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-3.1-flash-image-preview":{"id":"google/gemini-3.1-flash-image-preview","type":"image","name":"Nano Banana 2 Preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["image","text"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":58982},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image-preview"},"google/gemma-4-31b-it:free":{"id":"google/gemma-4-31b-it:free","name":"Gemma 4 31B (free)","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.08,"output":0.35,"cache_read":0.04},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"status":"deprecated","cost":{"input":0.1,"output":0.4,"reasoning":0.4,"cache_read":0.01,"cache_write":0.083333},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":1.25,"output":10,"reasoning":10,"cache_read":0.125,"cache_write":0.375,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemma-2-27b-it":{"id":"google/gemma-2-27b-it","name":"Gemma 2 27B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-07-13","last_updated":"2024-07-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":2048},"cost":{"input":0.65,"output":0.65},"canonical_model_id":"google/gemma-2-27b-it"},"google/lyria-3-clip-preview":{"id":"google/lyria-3-clip-preview","name":"Lyria 3 Clip Preview","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"lyria","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-25","last_updated":"2026-03-25","modalities":{"input":["text","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/lyria-3-clip-preview"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.825,"output":4.125,"reasoning":4.125,"cache_read":0.0825,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-2.5-pro-preview":{"id":"google/gemini-2.5-pro-preview","name":"Gemini 2.5 Pro Preview 06-05","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01-31","release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["pdf","image","text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"reasoning":10,"cache_read":0.125,"cache_write":0.375,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro-preview"},"google/gemini-3-pro-image":{"id":"google/gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3-pro-image"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"reasoning":9,"cache_read":0.15,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-3.1-flash-lite-image":{"id":"google/gemini-3.1-flash-lite-image","type":"image","name":"Nano Banana 2 Lite","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":58982},"cost":{"input":0.25,"output":1.5},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"google/gemini-nano-banana-2.1":{"id":"google/gemini-nano-banana-2.1","type":"image","name":"Nano Banana 2.1","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["image","text"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":58982},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"google/gemini-nano-banana-2.1"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Gemini multimodal model for text, image, audio, video, and document tasks","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":2,"output":10,"reasoning":10,"cache_read":0.1,"cache_write":0.1875},"canonical_model_id":"google/gemini-4-argon"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.09,"output":0.3,"cache_read":0.05},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemma-3-4b-it":{"id":"google/gemma-3-4b-it","name":"Gemma 3 4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.04,"output":0.05,"cache_read":0.02},"canonical_model_id":"google/gemma-3-4b-it"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.2,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"status":"deprecated","cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.1-flash-lite-preview":{"id":"google/gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"google/gemini-2.5-flash-image":{"id":"google/gemini-2.5-flash-image","type":"image","name":"Nano Banana","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.083333},"canonical_model_id":"google/gemini-2.5-flash-image"},"google/lyria-3-pro-preview":{"id":"google/lyria-3-pro-preview","name":"Lyria 3 Pro Preview","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"lyria","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-25","last_updated":"2026-03-25","modalities":{"input":["text","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/lyria-3-pro-preview"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 preview","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy-mt2-1.8b":{"id":"tencent/hy-mt2-1.8b","name":"Hy-MT2-1.8B","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.044,"output":0.177},"canonical_model_id":"tencent/hy-mt2-1.8b"},"tencent/hy-mt2-7b":{"id":"tencent/hy-mt2-7b","name":"Hy-MT2-7B","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.074,"output":0.295},"canonical_model_id":"tencent/hy-mt2-7b"},"tencent/hy-mt2-30b-a3b":{"id":"tencent/hy-mt2-30b-a3b","name":"Hy-MT2-30B-A3B","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.074,"output":0.295},"canonical_model_id":"tencent/hy-mt2-30b-a3b"},"tencent/hy3-preview":{"id":"tencent/hy3-preview","name":"Hy3 preview","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.18,"output":0.6,"cache_read":0.06},"canonical_model_id":"tencent/hy3-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":128000},"cost":{"input":0.132,"output":0.528,"cache_read":0.033},"canonical_model_id":"tencent/hy3"},"tencent/hunyuan-a13b-instruct":{"id":"tencent/hunyuan-a13b-instruct","name":"Hunyuan A13B Instruct","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-07-08","last_updated":"2025-07-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.14,"output":0.57}},"sakana/fugu-max":{"id":"sakana/fugu-max","name":"Fugu Max","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"sakana/fugu-max"},"sakana/sakana-namazu":{"id":"sakana/sakana-namazu","name":"Sakana Namazu","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"sakana-namazu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.95,"output":4,"cache_read":0.15},"canonical_model_id":"sakana/sakana-namazu"},"sakana/fugu-ultra-v2":{"id":"sakana/fugu-ultra-v2","name":"Fugu Ultra v2","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-08-28","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra-v2"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra"},"cognitivecomputations/dolphin-mistral-24b-venice-edition":{"id":"cognitivecomputations/dolphin-mistral-24b-venice-edition","name":"Uncensored","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04-30","release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.2,"output":0.9},"canonical_model_id":"cognitivecomputations/dolphin-mistral-24b-venice-edition"},"rekaai/reka-edge":{"id":"rekaai/reka-edge","name":"Reka Edge","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"reka","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":14745},"cost":{"input":0.1,"output":0.1},"canonical_model_id":"rekaai/reka-edge"},"rekaai/reka-flash-3":{"id":"rekaai/reka-flash-3","name":"Reka Flash 3","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"reka","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01-31","release_date":"2025-03-10","last_updated":"2025-03-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":58982},"cost":{"input":0.1,"output":0.2},"canonical_model_id":"rekaai/reka-flash-3"},"openai/o3-mini-high":{"id":"openai/o3-mini-high","name":"o3 Mini High","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-10-31","release_date":"2025-02-12","last_updated":"2025-02-12","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55}},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4":{"id":"openai/gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8191,"output":4096},"cost":{"input":30,"output":60},"canonical_model_id":"openai/gpt-4"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-3.5-turbo-instruct":{"id":"openai/gpt-3.5-turbo-instruct","name":"GPT-3.5 Turbo Instruct","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2021-09-30","release_date":"2023-09-28","last_updated":"2023-09-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4095,"output":3685},"cost":{"input":1.5,"output":2}},"openai/gpt-5.6-luna-pro":{"id":"openai/gpt-5.6-luna-pro","name":"GPT-5.6 Luna Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-audio":{"id":"openai/gpt-audio","type":"chat","name":"GPT-Audio","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-01","release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10},"canonical_model_id":"openai/gpt-audio"},"openai/gpt-5.2-chat":{"id":"openai/gpt-5.2-chat","name":"GPT-5.2 Chat","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-chat-latest"},"openai/gpt-6-luna-pro":{"id":"openai/gpt-6-luna-pro","name":"GPT-6 Luna Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}}},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/o3-pro":{"id":"openai/o3-pro","name":"o3-pro","description":"High-effort o3 tier for difficult technical reasoning and careful answers","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","pdf","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":20,"output":80},"canonical_model_id":"openai/o3-pro"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5-turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-chat-latest":{"id":"openai/gpt-chat-latest","name":"GPT Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-05-05","last_updated":"2026-05-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5}},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-4o-2024-08-06":{"id":"openai/gpt-4o-2024-08-06","name":"GPT-4o (2024-08-06)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-08-06"},"openai/gpt-3.5-turbo-0613":{"id":"openai/gpt-3.5-turbo-0613","name":"GPT-3.5 Turbo (older v0613)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2021-09-30","release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4095,"output":3685},"cost":{"input":1,"output":2}},"openai/gpt-6-astra-pro":{"id":"openai/gpt-6-astra-pro","name":"GPT-6 Astra Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}}},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-6-sol-pro":{"id":"openai/gpt-6-sol-pro","name":"GPT-6 Sol Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}}},"openai/gpt-oss-safeguard-20b":{"id":"openai/gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.018,"output":0.09,"cache_read":0.009},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/gpt-3.5-turbo-16k":{"id":"openai/gpt-3.5-turbo-16k","name":"GPT-3.5 Turbo 16k","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2021-09-30","release_date":"2023-08-28","last_updated":"2023-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":3,"output":4}},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.03},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/gpt-5.6-terra-pro":{"id":"openai/gpt-5.6-terra-pro","name":"GPT-5.6 Terra Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5.1-codex"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2 Pro","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5-image":{"id":"openai/gpt-5-image","type":"image","name":"GPT-5 Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-10-01","release_date":"2025-10-14","last_updated":"2025-10-14","modalities":{"input":["image","text","pdf"],"output":["image","text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":10,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-5-image"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/gpt-4o-2024-05-13":{"id":"openai/gpt-4o-2024-05-13","name":"GPT-4o (2024-05-13)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-05-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":5,"output":15},"canonical_model_id":"openai/gpt-4o-2024-05-13"},"openai/o4-mini-high":{"id":"openai/o4-mini-high","name":"o4 Mini High","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-06-30","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275}},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.6-sol-pro":{"id":"openai/gpt-5.6-sol-pro","name":"GPT-5.6 Sol Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5-image-mini":{"id":"openai/gpt-5-image-mini","type":"image","name":"GPT-5 Image Mini","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-10-16","last_updated":"2025-10-16","modalities":{"input":["pdf","image","text"],"output":["image","text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":2.5,"output":2,"cache_read":0.25},"canonical_model_id":"openai/gpt-5-image-mini"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1-codex-max":{"id":"openai/gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex-max"},"openai/gpt-4o-mini-2024-07-18":{"id":"openai/gpt-4o-mini-2024-07-18","name":"GPT-4o-mini (2024-07-18)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/o1-pro":{"id":"openai/o1-pro","name":"o1-pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2025-03-19","last_updated":"2025-03-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":150,"output":600},"canonical_model_id":"openai/o1-pro"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-6.1-sol-pro":{"id":"openai/gpt-6.1-sol-pro","name":"GPT-6.1 Sol Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}}},"openai/gpt-4o-2024-11-20":{"id":"openai/gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"openai/gpt-5.4-image-2":{"id":"openai/gpt-5.4-image-2","type":"image","name":"GPT-5.4 Image 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":false,"structured_output":true,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["image","text","pdf"],"output":["image","text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":8,"output":15,"cache_read":2},"canonical_model_id":"openai/gpt-5.4-image-2"},"openai/gpt-audio-mini":{"id":"openai/gpt-audio-mini","type":"chat","name":"GPT-Audio Mini","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-01","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.6,"output":2.4},"canonical_model_id":"openai/gpt-audio-mini"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"morph/morph-v3-fast":{"id":"morph/morph-v3-fast","name":"Morph V3 Fast","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-07","last_updated":"2025-07-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.8,"output":1.2},"canonical_model_id":"morph/morph-v3-fast"},"morph/morph-v3-large":{"id":"morph/morph-v3-large","name":"Morph V3 Large","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-07","last_updated":"2025-07-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.9,"output":1.9},"canonical_model_id":"morph/morph-v3-large"},"arcee-ai/trinity-large-thinking":{"id":"arcee-ai/trinity-large-thinking","name":"Trinity Large Thinking","description":"Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use","family":"trinity","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":80000},"cost":{"input":0.25,"output":0.8,"cache_read":0.06},"canonical_model_id":"arcee-ai/trinity-large-thinking"}}},"tensorx":{"id":"tensorx","env":["TENSORX_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.tensorx.ai/v1","name":"TensorX","doc":"https://docs.tensorx.ai/","models":{"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.75,"output":3.5,"cache_read":0.4375,"cache_write":2.185},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.25,"output":0.3,"cache_read":0.06},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-r1-0528":{"id":"deepseek/deepseek-r1-0528","name":"DeepSeek R1-0528","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":8192},"cost":{"input":0.66,"output":2.6,"cache_read":0.165,"cache_write":0.825},"canonical_model_id":"deepseek/deepseek-r1"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":2,"output":4,"cache_read":0.5},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.3,"output":0.5,"cache_read":0.075,"cache_write":0.375}},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.5,"output":1.5,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":1.75,"output":4.5,"cache_read":0.44},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":202752},"cost":{"input":1.4,"output":4.4,"cache_read":0.35,"cache_write":1.75},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":202752},"cost":{"input":1,"output":3.2,"cache_read":0.25,"cache_write":1.25},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-5v-turbo":{"id":"z-ai/glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.3,"cache_write":1.5},"canonical_model_id":"zhipuai/glm-5v-turbo"},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.3,"cache_write":1.5},"canonical_model_id":"zhipuai/glm-5-turbo"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.5,"output":4.5,"cache_read":0.375},"canonical_model_id":"zhipuai/glm-5.2"},"qwen/qwen3.5-9b":{"id":"qwen/qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.15,"output":0.2,"cache_read":0.0375,"cache_write":0.1875},"canonical_model_id":"alibaba/qwen3.5-9b"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":3.5,"cache_read":0.125,"cache_write":0.625},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3-235b-a22b-2507":{"id":"qwen/qwen3-235b-a22b-2507","name":"Qwen3 235B-A22B-2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06-30","release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":262144},"cost":{"input":0.072,"output":0.464,"cache_read":0.018,"cache_write":0.09}},"qwen/qwen3.8-2.4t-a95b":{"id":"qwen/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":64000},"cost":{"input":2.5,"output":6,"cache_read":0.63},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"qwen/qwen3.8-flash-next":{"id":"qwen/qwen3.8-flash-next","name":"Qwen3.8 Flash Next","description":"Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":64000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.8-flash-next"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.4,"output":2.4,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3.8-27b"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1,"output":4,"cache_read":0.25,"cache_write":1.25},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.75},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.25,"output":4.5,"cache_read":0.3125},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":2.8,"cache_read":0.125,"cache_write":0.625},"canonical_model_id":"moonshotai/kimi-k2.5"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.4,"output":2,"cache_read":0.1},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":65536},"cost":{"input":0.3,"output":1.2,"cache_read":0.075,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5"}}},"atomic-chat":{"id":"atomic-chat","env":["ATOMIC_CHAT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"http://127.0.0.1:1337/v1","name":"Atomic Chat","doc":"https://atomic.chat","models":{"Qwen3_5-9B-Q4_K_M":{"id":"Qwen3_5-9B-Q4_K_M","name":"Qwen 3.5 9B (Q4_K_M)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-05","last_updated":"2026-04-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0}},"gemma-4-E4B-it-MLX-4bit":{"id":"gemma-4-E4B-it-MLX-4bit","name":"Gemma 4 E4B Instruct (MLX 4-bit)","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0}},"gemma-4-E4B-it-IQ4_XS":{"id":"gemma-4-E4B-it-IQ4_XS","name":"Gemma 4 E4B Instruct (IQ4_XS)","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0}},"Qwen3_5-9B-MLX-4bit":{"id":"Qwen3_5-9B-MLX-4bit","name":"Qwen 3.5 9B (MLX 4-bit)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-05","last_updated":"2026-04-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0}},"Meta-Llama-3_1-8B-Instruct-GGUF":{"id":"Meta-Llama-3_1-8B-Instruct-GGUF","name":"Meta Llama 3.1 8B Instruct (GGUF)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0,"output":0}}}},"openreason":{"id":"openreason","env":["OPENREASON_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.openreason.app/v1","name":"OpenReason","doc":"https://openreason.app/docs","models":{"deepseek-ai/deepseek-v4-flash-0731":{"id":"deepseek-ai/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.1371,"output":0.2743},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.0022,"output":4.22},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.1055,"output":0.422},"canonical_model_id":"openai/gpt-oss-120b"}}},"freemodel":{"id":"freemodel","env":["FREEMODEL_API_KEY"],"npm":"@ai-sdk/anthropic","api":"https://cc.freemodel.dev/v1","name":"FreeModel","doc":"https://freemodel.dev","models":{"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://api.freemodel.dev/v1"},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":5},"canonical_model_id":"openai/gpt-5.5"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://api.freemodel.dev/v1"},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.4"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://api.freemodel.dev/v1"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075,"cache_write":0.75},"canonical_model_id":"openai/gpt-5.4-mini"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://api.freemodel.dev/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.175,"cache_write":1.75},"canonical_model_id":"openai/gpt-5.3-codex"}}},"merge-gateway":{"id":"merge-gateway","env":["MERGE_GATEWAY_API_KEY"],"npm":"merge-gateway-ai-sdk-provider","api":"https://api-gateway.merge.dev/v1/ai-sdk","name":"Merge Gateway","doc":"https://docs.merge.dev/merge-gateway","models":{"anthropic/claude-3-7-sonnet-20250219":{"id":"anthropic/claude-3-7-sonnet-20250219","name":"Claude 3.7 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10-31","release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-3-7-sonnet-20250219"},"anthropic/claude-opus-4-1-20250805":{"id":"anthropic/claude-opus-4-1-20250805","name":"Claude Opus 4.1 (20250805)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4-5-20251101":{"id":"anthropic/claude-opus-4-5-20251101","name":"Claude Opus 4.5 (20251101)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024,"max":127999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-fable-5-1":{"id":"anthropic/claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-sonnet-4-5-20250929":{"id":"anthropic/claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5 (20250929)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"anthropic/claude-sonnet-4-20250514":{"id":"anthropic/claude-sonnet-4-20250514","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"anthropic/claude-opus-4-8":{"id":"anthropic/claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":128000}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-haiku-4-5-20251001":{"id":"anthropic/claude-haiku-4-5-20251001","name":"Claude Haiku 4.5 (20251001)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"anthropic/claude-sonnet-5-5":{"id":"anthropic/claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-haiku-5-5":{"id":"anthropic/claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-5-5":{"id":"anthropic/claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-opus-4-20250514":{"id":"anthropic/claude-opus-4-20250514","name":"Claude Opus 4 (20250514)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-20250514"},"xai/grok-4.3":{"id":"xai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"xai/grok-4.20-0309-reasoning":{"id":"xai/grok-4.20-0309-reasoning","name":"Grok 4.20","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"xai/grok-4.7":{"id":"xai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"xai/grok-build-0.1":{"id":"xai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2},"canonical_model_id":"xai/grok-build-0.1"},"xai/grok-4.5":{"id":"xai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.5"},"xai/grok-4.6":{"id":"xai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"xai/grok-4.20-0309-non-reasoning":{"id":"xai/grok-4.20-0309-non-reasoning","name":"Grok 4.20 Non-Reasoning","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"moonshot/kimi-k2.6":{"id":"moonshot/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshot/kimi-k3":{"id":"moonshot/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":2.9,"output":14,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshot/kimi-k2.7-code-highspeed":{"id":"moonshot/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshot/kimi-k2.7-code":{"id":"moonshot/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshot/kimi-k2.5":{"id":"moonshot/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"deepseek/deepseek-v4-flash-0423":{"id":"deepseek/deepseek-v4-flash-0423","name":"DeepSeek V4 Flash 0423","description":"Initial DeepSeek V4 Flash snapshot for economical reasoning, coding, and million-token agent workloads","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.139,"output":0.278},"canonical_model_id":"deepseek/deepseek-v4-flash-0423"},"deepseek/deepseek-v4-flash-0731-fast":{"id":"deepseek/deepseek-v4-flash-0731-fast","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.28,"output":0.56,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.035,"output":0.07,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-r1-0528":{"id":"deepseek/deepseek-r1-0528","name":"DeepSeek R1 (0528)","description":"May 2025 DeepSeek R1 update with improved mathematics, coding, and reasoning","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":32768},"cost":{"input":1.35,"output":5.4,"cache_read":0.945},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek/deepseek-v3":{"id":"deepseek/deepseek-v3","name":"DeepSeek V3","description":"Open DeepSeek MoE chat model for coding, math, and general reasoning","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":81920},"cost":{"input":0.27,"output":1.1},"canonical_model_id":"deepseek/deepseek-v3"},"deepseek/deepseek-r1":{"id":"deepseek/deepseek-r1","name":"DeepSeek R1","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":40960},"cost":{"input":1.35,"output":5.4},"canonical_model_id":"deepseek/deepseek-r1"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.74,"output":3.48,"cache_read":0.003625},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v3.1":{"id":"deepseek/deepseek-v3.1","name":"DeepSeek V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.56,"output":1.68,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006,"cache_write":0.1125},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4-pro-0423":{"id":"deepseek/deepseek-v4-pro-0423","name":"DeepSeek V4 Pro 0423","description":"DeepSeek V4 Pro initial snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":1.65,"output":3.3},"canonical_model_id":"deepseek/deepseek-v4-pro-0423"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.28,"output":0.4,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"nvidia/nemotron-3.5-lightning-30b-a3b":{"id":"nvidia/nemotron-3.5-lightning-30b-a3b","name":"Nemotron 3.5 Lightning 30B A3B","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-nano-9b-v2":{"id":"nvidia/nemotron-nano-9b-v2","name":"Nemotron Nano 9B","description":"Compact Nemotron model for efficient reasoning and deployable AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-09","release_date":"2025-08-18","last_updated":"2025-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.06,"output":0.23},"canonical_model_id":"nvidia/nemotron-nano-9b-v2"},"nvidia/nemotron-3-nano-30b-a3b":{"id":"nvidia/nemotron-3-nano-30b-a3b","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"cost":{"input":0.05,"output":0.2},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"meta/llama-3.1-8b-instruct":{"id":"meta/llama-3.1-8b-instruct","name":"Llama 3.1 8B","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":2048},"cost":{"input":0.22,"output":0.22},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta/llama-3.3-70b-instruct":{"id":"meta/llama-3.3-70b-instruct","name":"Llama 3.3 70B Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.22,"output":0.5,"cache_read":0.11},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":262144},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-open":{"id":"meta/muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"meta/llama-3.1-70b-instruct":{"id":"meta/llama-3.1-70b-instruct","name":"Llama 3.1 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":2048},"cost":{"input":0.99,"output":0.99},"canonical_model_id":"meta/llama-3.1-70b-instruct"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":262144},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"qwen/qwen3-vl-30b-a3b-instruct":{"id":"qwen/qwen3-vl-30b-a3b-instruct","name":"Qwen3-VL 30B-A3B Instruct","description":"Open vision-language MoE model for image and video understanding, documents, and visual agents","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.0216},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"qwen/qwen3.5-9b":{"id":"qwen/qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.09,"output":0.13},"canonical_model_id":"alibaba/qwen3.5-9b"},"qwen/qwen3-vl-235b-a22b-instruct":{"id":"qwen/qwen3-vl-235b-a22b-instruct","name":"Qwen3-VL 235B A22B Instruct","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.21,"output":1.9,"cache_read":0.105},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"qwen/qwen-plus":{"id":"qwen/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.115,"output":0.287,"cache_read":0.023,"cache_write":0.14375},"canonical_model_id":"alibaba/qwen-plus"},"qwen/qwen-flash":{"id":"qwen/qwen-flash","name":"Qwen Flash","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.022,"output":0.216,"cache_read":0.0044,"cache_write":0.0275},"canonical_model_id":"alibaba/qwen-flash"},"qwen/qwen3-coder-plus":{"id":"qwen/qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Hosted Qwen coder for software agents, repo edits, and long-context code","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.574,"output":2.294,"cache_read":0.1148,"cache_write":0.7175},"canonical_model_id":"alibaba/qwen3-coder-plus"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.115,"output":0.917,"cache_read":0.023},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3.5-plus":{"id":"qwen/qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.115,"output":0.688,"cache_read":0.0115,"cache_write":0.14375},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen/qwen3.6-35b-a3b":{"id":"qwen/qwen3.6-35b-a3b","name":"Qwen3.6 35B A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.248,"output":1.485,"cache_read":0.0496},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.276,"output":1.651,"cache_read":0.0276,"cache_write":0.345},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen/qwen3-235b-a22b-instruct-2507":{"id":"qwen/qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.0574},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.165,"output":0.99,"cache_read":0.0165,"cache_write":0.20625},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3.5-flash":{"id":"qwen/qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.029,"output":0.287,"cache_read":0.0029,"cache_write":0.03625},"canonical_model_id":"alibaba/qwen3.5-flash"},"qwen/qwen3-235b-a22b":{"id":"qwen/qwen3-235b-a22b","name":"Qwen3 235B A22B","description":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.287,"output":1.147,"cache_read":0.0574},"canonical_model_id":"alibaba/qwen3-235b-a22b"},"qwen/qwen3-coder-flash":{"id":"qwen/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.144,"output":0.574,"cache_read":0.0288,"cache_write":0.18},"canonical_model_id":"alibaba/qwen3-coder-flash"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":977000,"output":128000},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3-30b-a3b":{"id":"qwen/qwen3-30b-a3b","name":"Qwen3 30B A3B","description":"Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.108,"output":1.076,"cache_read":0.0216},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"qwen/qwen3-vl-235b-a22b-thinking":{"id":"qwen/qwen3-vl-235b-a22b-thinking","name":"Qwen3-VL 235B A22B Thinking","description":"Qwen vision-language thinking model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.287,"output":2.867,"cache_read":0.0574},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-thinking"},"qwen/qwen3.8-2.4t-a95b":{"id":"qwen/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":1010000},"cost":{"input":2.5,"output":6.25,"cache_read":0.5},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.4,"output":1.6},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3-next-80b-a3b-instruct":{"id":"qwen/qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.1,"output":1.1,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.086,"output":0.688,"cache_read":0.0172},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3-next-80b-a3b-thinking":{"id":"qwen/qwen3-next-80b-a3b-thinking","name":"Qwen3-Next 80B-A3B (Thinking)","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.057,"output":0.459,"cache_read":0.020357},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3-coder-480b-a35b-instruct":{"id":"qwen/qwen3-coder-480b-a35b-instruct","name":"Qwen3 Coder 480B A35B Instruct","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.22,"output":1.8},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"qwen/qwen3-32b":{"id":"qwen/qwen3-32b","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-32b"},"qwen/qwen3-vl-flash":{"id":"qwen/qwen3-vl-flash","name":"Qwen3-VL Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.022,"output":0.215,"cache_read":0.0044,"cache_write":0.0275},"canonical_model_id":"alibaba/qwen3-vl-flash"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.172,"output":1.032,"cache_read":0.0344},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3.6-max-preview":{"id":"qwen/qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"cost":{"input":1.31,"output":7.88},"canonical_model_id":"alibaba/qwen3.6-max-preview"},"qwen/qwen3-max-preview":{"id":"qwen/qwen3-max-preview","name":"Qwen3 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-05","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"cost":{"input":1.08,"output":4.8},"canonical_model_id":"alibaba/qwen3-max-preview"},"qwen/qwen3-coder-next":{"id":"qwen/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.15,"output":0.8,"cache_read":0.075},"canonical_model_id":"alibaba/qwen3-coder-next"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.4126,"output":2.4754},"canonical_model_id":"alibaba/qwen3.6-27b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.825,"output":2.4755,"cache_read":0.165,"cache_write":1.03125},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3-vl-plus":{"id":"qwen/qwen3-vl-plus","name":"Qwen3-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.143,"output":1.434,"cache_read":0.0286,"cache_write":0.17875},"canonical_model_id":"alibaba/qwen3-vl-plus"},"qwen/qwen3.8-max":{"id":"qwen/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen/qwen3-max":{"id":"qwen/qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.359,"output":1.434,"cache_read":0.0718,"cache_write":0.44875},"canonical_model_id":"alibaba/qwen3-max"},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.6,"output":2.5},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"writer/palmyra-x5":{"id":"writer/palmyra-x5","name":"Palmyra X5","description":"Enterprise multimodal model for writing, analysis, and tool-assisted workflows","family":"palmyra","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.6,"output":6}},"writer/palmyra-x4":{"id":"writer/palmyra-x4","name":"Palmyra X4","description":"Enterprise language model for writing, analysis, and tool-assisted workflows","family":"palmyra","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2024-10-09","last_updated":"2024-10-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000},"cost":{"input":2.5,"output":10}},"bytedance/dola-seed-2.0-code":{"id":"bytedance/dola-seed-2.0-code","name":"Seed 2.0 Code","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"seed","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.4,"output":2.4},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"bytedance/dola-seed-2.0-code-preview":{"id":"bytedance/dola-seed-2.0-code-preview","name":"Dola Seed 2.0 Code (preview)","description":"Preview coding model for repository understanding, refactors, and engineering tasks","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-28","last_updated":"2026-03-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3}},"bytedance/dola-seed-2.0-mini":{"id":"bytedance/dola-seed-2.0-mini","name":"Seed 2.0 Mini","description":"Low-cost Seed model for general chat, extraction, and lightweight production tasks","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.1,"output":0.4}},"bytedance/dola-seed-2.0-lite":{"id":"bytedance/dola-seed-2.0-lite","name":"Seed 2.0 Lite","description":"Efficient Seed model for general chat, analysis, and lightweight production tasks","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-28","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.25,"output":2}},"bytedance/dola-seed-2.0-pro":{"id":"bytedance/dola-seed-2.0-pro","name":"Seed 2.0 Pro","description":"Higher-capability Seed model for complex chat, analysis, and production tasks","family":"seed","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-28","last_updated":"2026-03-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3}},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048000,"output":32000},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-m2.5-highspeed":{"id":"minimax/minimax-m2.5-highspeed","name":"MiniMax M2.5 Highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":8192},"cost":{"input":0.6,"output":2.4,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5-highspeed"},"minimax/minimax-m3.1-flash":{"id":"minimax/minimax-m3.1-flash","name":"MiniMax M3.1 Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax M2.7 Highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":8192},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":128000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5"},"cohere/command-a-03-2025":{"id":"cohere/command-a-03-2025","name":"Command A 03-2025","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a-03-2025"},"cohere/command-r-08-2024":{"id":"cohere/command-r-08-2024","name":"Command R 08-2024","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"cohere/command-r-08-2024"},"cohere/command-r7b-12-2024":{"id":"cohere/command-r7b-12-2024","name":"Command R7B 12-2024","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-12-13","last_updated":"2024-12-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.0375,"output":0.15},"canonical_model_id":"cohere/command-r7b-12-2024"},"cohere/command-r-plus-08-2024":{"id":"cohere/command-r-plus-08-2024","name":"Command R+ 08-2024","description":"Cohere's RAG workhorse for long-context enterprise search and tool use","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-r-plus-08-2024"},"zai/glm-5.3-flash":{"id":"zai/glm-5.3-flash","name":"GLM-5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai/glm-4.5v":{"id":"zai/glm-4.5v","name":"Glm 4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.6,"output":1.8,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5v"},"zai/glm-4.5-air":{"id":"zai/glm-4.5-air","name":"GLM-4.5 Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":98304},"cost":{"input":0.2,"output":1.1,"cache_read":0.03,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-air"},"zai/glm-5.3":{"id":"zai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.7,"output":2.2,"cache_read":0.13},"canonical_model_id":"zhipuai/glm-5.3"},"zai/glm-4.7":{"id":"zai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7"},"zai/glm-5.1":{"id":"zai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"zai/glm-4.7-flash":{"id":"zai/glm-4.7-flash","name":"GLM 4.7 Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.07,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"zai/glm-5":{"id":"zai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2,"cache_write":0},"canonical_model_id":"zhipuai/glm-5"},"zai/glm-5-turbo":{"id":"zai/glm-5-turbo","name":"GLM-5 Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24,"cache_write":0},"canonical_model_id":"zhipuai/glm-5-turbo"},"zai/glm-4.6":{"id":"zai/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6"},"zai/glm-5.2":{"id":"zai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":50000}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.05,"output":3.3,"cache_read":0.195},"canonical_model_id":"zhipuai/glm-5.2"},"zai/glm-4.5":{"id":"zai/glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5"},"zai/glm-4.7-flashx":{"id":"zai/glm-4.7-flashx","name":"GLM-4.7 FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.07,"output":0.4,"cache_read":0.01,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"google/gemini-3.1-flash-image":{"id":"google/gemini-3.1-flash-image","type":"image","name":"Gemini 3.1 Flash Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image"},"google/gemini-flash-lite-latest":{"id":"google/gemini-flash-lite-latest","name":"Gemini Flash-Lite Latest","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-flash-lite-latest"},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.09,"output":0.29},"canonical_model_id":"google/gemma-3-12b-it"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-2.5-computer-use-preview-10-2025":{"id":"google/gemini-2.5-computer-use-preview-10-2025","name":"Gemini 2.5 Computer Use Preview (10-2025)","description":"Specialized Gemini 2.5 model for browser-control agents that automate UI tasks","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-10-07","last_updated":"2025-10-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"cost":{"input":1.25,"output":10},"canonical_model_id":"google/gemini-2.5-computer-use-preview-10-2025"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.08,"output":0.45,"cache_read":0.04},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemini-embedding-001":{"id":"google/gemini-embedding-001","type":"embedding","name":"Gemini Embedding 001","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-05","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2048,"output":4096},"cost":{"input":0.15,"output":0},"canonical_model_id":"google/gemini-embedding-001"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash-Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-embedding-2":{"id":"google/gemini-embedding-2","type":"embedding","name":"Gemini Embedding 2","description":"Multimodal embedding model mapping text, images, video, audio, and PDFs into a unified embedding space","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-11","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0.2,"output":0},"canonical_model_id":"google/gemini-embedding-2"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3-pro-image":{"id":"google/gemini-3-pro-image","type":"image","name":"Gemini 3 Pro Image","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":65536},"cost":{"input":2,"output":12},"canonical_model_id":"google/gemini-3-pro-image"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"google/gemini-4-argon"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.13,"output":0.4},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-flash-latest":{"id":"google/gemini-flash-latest","name":"Gemini Flash Latest","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-flash-latest"},"google/gemma-3-4b-it":{"id":"google/gemma-3-4b-it","name":"Gemma 3 4B","description":"Open multimodal Gemma instruction model for efficient text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.04,"output":0.08},"canonical_model_id":"google/gemma-3-4b-it"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash-Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B It","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.1-flash-lite-preview":{"id":"google/gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"google/gemini-2.5-flash-image":{"id":"google/gemini-2.5-flash-image","type":"image","name":"Gemini 2.5 Flash Image","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":2.5},"canonical_model_id":"google/gemini-2.5-flash-image"},"google/gemini-3-pro-preview":{"id":"google/gemini-3-pro-preview","name":"Gemini 3 Pro Preview","description":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3-pro-preview"},"sakana/fugu-max":{"id":"sakana/fugu-max","name":"Fugu Max","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"sakana/fugu-max"},"sakana/sakana-namazu":{"id":"sakana/sakana-namazu","name":"Sakana Namazu","description":"Japanese-specialized reasoning model based on Kimi K2.6 and tuned for Japanese language, culture, and business workflows","family":"sakana-namazu","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.95,"output":4,"cache_read":0.15},"canonical_model_id":"sakana/sakana-namazu"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"sakana/fugu-ultra"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4 Mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o Mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4":{"id":"openai/gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":30,"output":60},"canonical_model_id":"openai/gpt-4"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-5.3-chat-latest":{"id":"openai/gpt-5.3-chat-latest","name":"GPT-5.3 Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-chat-latest"},"openai/gpt-live-1":{"id":"openai/gpt-live-1","type":"realtime","name":"GPT-Live 1","description":"Full-duplex voice model for natural conversations that delegates reasoning and tools to a backend agent","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-07-31","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-live-1"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 Nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 Mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-audio":{"id":"openai/gpt-audio","type":"chat","name":"GPT Audio","description":"Audio model for speech input and output with the Chat Completions API","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-01","release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10},"canonical_model_id":"openai/gpt-audio"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/whisper-1":{"id":"openai/whisper-1","type":"transcription","name":"Whisper","description":"General-purpose speech recognition model for transcription and translation","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2023-03-01","last_updated":"2023-03-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":448000,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"openai/whisper-1"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3 Mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5-search-api":{"id":"openai/gpt-5-search-api","type":"chat","name":"GPT-5 Search API","description":"GPT-5 model trained for web search in the Chat Completions API","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-10-14","last_updated":"2025-10-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":1.25,"output":10},"canonical_model_id":"openai/gpt-5-search-api"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-4o-2024-08-06":{"id":"openai/gpt-4o-2024-08-06","name":"GPT-4o (2024-08-06)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-08-06"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5.1-chat-latest":{"id":"openai/gpt-5.1-chat-latest","name":"GPT-5.1 Chat Latest","description":"Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-chat-latest"},"openai/gpt-oss-safeguard-20b":{"id":"openai/gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":0.07,"output":0.2,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT-OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.04,"output":0.15,"cache_read":0.02},"canonical_model_id":"openai/gpt-oss-20b"},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/gpt-4o-2024-05-13":{"id":"openai/gpt-4o-2024-05-13","name":"GPT-4o (2024-05-13)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-05-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":5,"output":15},"canonical_model_id":"openai/gpt-4o-2024-05-13"},"openai/gpt-oss-safeguard-120b":{"id":"openai/gpt-oss-safeguard-120b","name":"GPT OSS Safeguard 120B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-safeguard-120b"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":5},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.2-chat-latest":{"id":"openai/gpt-5.2-chat-latest","name":"GPT-5.2 Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-chat-latest"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT-OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.05,"output":0.25},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":1.5,"output":9,"cache_read":0.15},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-4o-2024-11-20":{"id":"openai/gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"openai/gpt-5-chat-latest":{"id":"openai/gpt-5-chat-latest","name":"GPT-5 Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5-chat-latest"},"openai/gpt-audio-mini":{"id":"openai/gpt-audio-mini","type":"chat","name":"GPT Audio Mini","description":"Cost-efficient audio model for speech input and output with the Chat Completions API","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-01","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.6,"output":2.4},"canonical_model_id":"openai/gpt-audio-mini"},"mistral/devstral-small-2507":{"id":"mistral/devstral-small-2507","name":"Devstral Small","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-07-10","last_updated":"2025-07-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/devstral-small-2507"},"mistral/magistral-medium-latest":{"id":"mistral/magistral-medium-latest","name":"Magistral Medium (latest)","description":"Mistral reasoning model for transparent analysis, math, and complex decisions","family":"magistral-medium","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-03-17","last_updated":"2025-03-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2,"output":5},"canonical_model_id":"mistral/magistral-medium-latest"},"mistral/codestral-latest":{"id":"mistral/codestral-latest","name":"Codestral (latest)","description":"Mistral code model for completions, refactors, and developer IDE workflows","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-05-29","last_updated":"2025-01-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":4096},"cost":{"input":0.3,"output":0.9},"canonical_model_id":"mistral/codestral-latest"},"mistral/mistral-small-latest":{"id":"mistral/mistral-small-latest","name":"Mistral Small (latest)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"mistral/mistral-small-latest"},"mistral/devstral-2512":{"id":"mistral/devstral-2512","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/devstral-2512"},"mistral/mistral-large-2512":{"id":"mistral/mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"mistral/mistral-large-2512"},"mistral/mistral-medium-2505":{"id":"mistral/mistral-medium-2505","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/mistral-medium-2505"},"mistral/devstral-medium-latest":{"id":"mistral/devstral-medium-latest","name":"Devstral 2 (latest)","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/devstral-medium-latest"},"mistral/pixtral-large-latest":{"id":"mistral/pixtral-large-latest","name":"Pixtral Large (latest)","description":"Mistral's larger vision model for document-heavy image understanding and chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2024-11-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/pixtral-large-latest"},"mistral/mistral-large-2411":{"id":"mistral/mistral-large-2411","name":"Mistral Large 2.1","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-18","last_updated":"2024-11-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/mistral-large-2411"},"mistral/devstral-medium-2507":{"id":"mistral/devstral-medium-2507","name":"Devstral Medium","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-07-10","last_updated":"2025-07-10","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/devstral-medium-2507"},"mistral/mistral-large-latest":{"id":"mistral/mistral-large-latest","name":"Mistral Large (latest)","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/mistral-large-latest"},"mistral/mistral-medium-latest":{"id":"mistral/mistral-medium-latest","name":"Mistral Medium (latest)","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/mistral-medium-latest"}}},"ai-router":{"id":"ai-router","env":["AI_ROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.ai-router.dev/v1","name":"AI-ROUTER","doc":"https://ai-router.dev/openai-compatible-api-gateway/","models":{"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":1,"output":6,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"openai/gpt-5.6-luna"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"openai/gpt-5.6-sol"}}},"siliconflow-cn":{"id":"siliconflow-cn","env":["SILICONFLOW_CN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.siliconflow.cn/v1","name":"SiliconFlow (China)","doc":"https://cloud.siliconflow.com/models","models":{"inclusionAI/Ling-flash-2.0":{"id":"inclusionAI/Ling-flash-2.0","name":"inclusionAI/Ling-flash-2.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-18","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":131000},"cost":{"input":0.14,"output":0.57}},"deepseek-ai/DeepSeek-V3.2":{"id":"deepseek-ai/DeepSeek-V3.2","name":"deepseek-ai/DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-03","last_updated":"2025-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":164000,"output":164000},"cost":{"input":0.27,"output":0.42}},"deepseek-ai/DeepSeek-V4-Flash":{"id":"deepseek-ai/DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-ai/DeepSeek-OCR":{"id":"deepseek-ai/DeepSeek-OCR","name":"deepseek-ai/DeepSeek-OCR","description":"OCR model for extracting structured text from documents and screenshots","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-10-20","last_updated":"2025-10-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0}},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"deepseek-ai/DeepSeek-V4-Pro","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1049000,"output":393000},"cost":{"input":1.74,"output":3.48,"cache_read":0.145}},"deepseek-ai/DeepSeek-V3.1-Terminus":{"id":"deepseek-ai/DeepSeek-V3.1-Terminus","name":"deepseek-ai/DeepSeek-V3.1-Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-29","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":164000,"output":164000},"cost":{"input":0.27,"output":1}},"deepseek-ai/DeepSeek-R1":{"id":"deepseek-ai/DeepSeek-R1","name":"deepseek-ai/DeepSeek-R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":164000,"output":164000},"cost":{"input":0.5,"output":2.18}},"deepseek-ai/DeepSeek-V3":{"id":"deepseek-ai/DeepSeek-V3","name":"deepseek-ai/DeepSeek-V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-12-26","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":164000,"output":164000},"cost":{"input":0.25,"output":1}},"stepfun-ai/Step-3.5-Flash":{"id":"stepfun-ai/Step-3.5-Flash","name":"stepfun-ai/Step-3.5-Flash","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","family":"step","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.1,"output":0.3}},"ByteDance-Seed/Seed-OSS-36B-Instruct":{"id":"ByteDance-Seed/Seed-OSS-36B-Instruct","name":"ByteDance-Seed/Seed-OSS-36B-Instruct","description":"Tool-capable chat model for instruction following and agentic application workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-04","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.21,"output":0.57}},"Qwen/Qwen3.5-35B-A3B":{"id":"Qwen/Qwen3.5-35B-A3B","name":"Qwen/Qwen3.5-35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-25","last_updated":"2026-02-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.23,"output":1.86}},"Qwen/Qwen3.6-35B-A3B":{"id":"Qwen/Qwen3.6-35B-A3B","name":"Qwen/Qwen3.6-35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.23,"output":1.86}},"Qwen/Qwen3.5-4B":{"id":"Qwen/Qwen3.5-4B","name":"Qwen/Qwen3.5-4B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0}},"Qwen/Qwen3-235B-A22B-Thinking-2507":{"id":"Qwen/Qwen3-235B-A22B-Thinking-2507","name":"Qwen/Qwen3-235B-A22B-Thinking-2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.13,"output":0.6}},"Qwen/Qwen3.5-122B-A10B":{"id":"Qwen/Qwen3.5-122B-A10B","name":"Qwen/Qwen3.5-122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.29,"output":2.32}},"Qwen/Qwen3.5-9B":{"id":"Qwen/Qwen3.5-9B","name":"Qwen/Qwen3.5-9B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.22,"output":1.74}},"Qwen/Qwen3-32B":{"id":"Qwen/Qwen3-32B","name":"Qwen/Qwen3-32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-30","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":131000},"cost":{"input":0.14,"output":0.57}},"Qwen/Qwen3-8B":{"id":"Qwen/Qwen3-8B","name":"Qwen/Qwen3-8B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-30","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":131000},"cost":{"input":0.06,"output":0.06}},"Qwen/Qwen3.5-27B":{"id":"Qwen/Qwen3.5-27B","name":"Qwen/Qwen3.5-27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-25","last_updated":"2026-02-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.26,"output":2.09}},"Qwen/Qwen3.5-397B-A17B":{"id":"Qwen/Qwen3.5-397B-A17B","name":"Qwen/Qwen3.5-397B-A17B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.29,"output":1.74}},"Qwen/Qwen3-14B":{"id":"Qwen/Qwen3-14B","name":"Qwen/Qwen3-14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-30","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":131000},"cost":{"input":0.07,"output":0.28}},"Qwen/Qwen3-30B-A3B-Instruct-2507":{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507","name":"Qwen/Qwen3-30B-A3B-Instruct-2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.09,"output":0.3}},"Qwen/Qwen3-VL-32B-Instruct":{"id":"Qwen/Qwen3-VL-32B-Instruct","name":"Qwen/Qwen3-VL-32B-Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-21","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.2,"output":0.6}},"Qwen/Qwen2.5-7B-Instruct":{"id":"Qwen/Qwen2.5-7B-Instruct","name":"Qwen/Qwen2.5-7B-Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-09-18","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":33000,"output":4000},"cost":{"input":0.05,"output":0.05}},"Qwen/Qwen3-Coder-30B-A3B-Instruct":{"id":"Qwen/Qwen3-Coder-30B-A3B-Instruct","name":"Qwen/Qwen3-Coder-30B-A3B-Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-01","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.07,"output":0.28}},"Qwen/Qwen3-VL-30B-A3B-Instruct":{"id":"Qwen/Qwen3-VL-30B-A3B-Instruct","name":"Qwen/Qwen3-VL-30B-A3B-Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-05","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.29,"output":1}},"Qwen/Qwen3-Coder-480B-A35B-Instruct":{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct","name":"Qwen/Qwen3-Coder-480B-A35B-Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-31","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.25,"output":1}},"Qwen/Qwen2.5-72B-Instruct":{"id":"Qwen/Qwen2.5-72B-Instruct","name":"Qwen/Qwen2.5-72B-Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-09-18","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":33000,"output":4000},"cost":{"input":0.59,"output":0.59}},"Qwen/Qwen3-VL-32B-Thinking":{"id":"Qwen/Qwen3-VL-32B-Thinking","name":"Qwen/Qwen3-VL-32B-Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-21","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.2,"output":1.5}},"Qwen/Qwen3-VL-30B-A3B-Thinking":{"id":"Qwen/Qwen3-VL-30B-A3B-Thinking","name":"Qwen/Qwen3-VL-30B-A3B-Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-11","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.29,"output":1}},"Qwen/Qwen3-VL-8B-Instruct":{"id":"Qwen/Qwen3-VL-8B-Instruct","name":"Qwen/Qwen3-VL-8B-Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.18,"output":0.68}},"PaddlePaddle/PaddleOCR-VL-1.5":{"id":"PaddlePaddle/PaddleOCR-VL-1.5","name":"PaddlePaddle/PaddleOCR-VL-1.5","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-29","last_updated":"2026-01-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":0,"output":0}},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1049000,"output":262000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-4.5-Air":{"id":"zai-org/GLM-4.5-Air","name":"zai-org/GLM-4.5-Air","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-air","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":131000},"cost":{"input":0.14,"output":0.86}},"Pro/deepseek-ai/DeepSeek-V3":{"id":"Pro/deepseek-ai/DeepSeek-V3","name":"Pro/deepseek-ai/DeepSeek-V3","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-12-26","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":164000,"output":164000},"cost":{"input":0.25,"output":1}},"Pro/deepseek-ai/DeepSeek-V3.2":{"id":"Pro/deepseek-ai/DeepSeek-V3.2","name":"Pro/deepseek-ai/DeepSeek-V3.2","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-03","last_updated":"2025-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":164000,"output":164000},"cost":{"input":0.27,"output":0.42}},"Pro/deepseek-ai/DeepSeek-V3.1-Terminus":{"id":"Pro/deepseek-ai/DeepSeek-V3.1-Terminus","name":"Pro/deepseek-ai/DeepSeek-V3.1-Terminus","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-29","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":164000,"output":164000},"cost":{"input":0.27,"output":1}},"Pro/deepseek-ai/DeepSeek-R1":{"id":"Pro/deepseek-ai/DeepSeek-R1","name":"Pro/deepseek-ai/DeepSeek-R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":164000,"output":164000},"cost":{"input":0.5,"output":2.18}},"Pro/moonshotai/Kimi-K2.6":{"id":"Pro/moonshotai/Kimi-K2.6","name":"Pro/moonshotai/Kimi-K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.95,"output":4,"cache_read":0.16}},"Pro/moonshotai/Kimi-K2.5":{"id":"Pro/moonshotai/Kimi-K2.5","name":"Pro/moonshotai/Kimi-K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.45,"output":2.25,"cache_read":0.07}},"Pro/MiniMaxAI/MiniMax-M2.5":{"id":"Pro/MiniMaxAI/MiniMax-M2.5","name":"Pro/MiniMaxAI/MiniMax-M2.5","description":"Frontier MiniMax model for engineering, office tasks, and agentic reasoning","family":"minimax","attachment":false,"reasoning":false,"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":192000,"output":131000},"cost":{"input":0.3,"output":1.22}},"Pro/zai-org/GLM-5.1":{"id":"Pro/zai-org/GLM-5.1","name":"Pro/zai-org/GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-04-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":205000,"output":205000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0}},"Pro/zai-org/GLM-5":{"id":"Pro/zai-org/GLM-5","name":"Pro/zai-org/GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":205000,"output":205000},"cost":{"input":1,"output":3.2}},"tencent/Hunyuan-A13B-Instruct":{"id":"tencent/Hunyuan-A13B-Instruct","name":"tencent/Hunyuan-A13B-Instruct","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-06-30","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":131000},"cost":{"input":0.14,"output":0.57}}}},"bothub":{"id":"bothub","env":["BOTHUB_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://openai.bothub.ru/v1","name":"Bothub","doc":"https://bothub.ru/models","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.12,"output":0.44},"canonical_model_id":"zhipuai/glm-5.3-flash"},"muse-spark-1.3-contributor":{"id":"muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.1,"output":0.2},"canonical_model_id":"meta/muse-spark-1.3"},"gemma-4-31b-it:free":{"id":"gemma-4-31b-it:free","name":"Gemma 4 31B IT (free)","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-31b-it"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.1,"output":0.28},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.61,"output":4.84},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.06,"output":0.37},"canonical_model_id":"openai/gpt-5.6-luna"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.72,"output":5.41},"canonical_model_id":"zhipuai/glm-5.3"},"nemotron-3-ultra-550b-a55b:free":{"id":"nemotron-3-ultra-550b-a55b:free","name":"Nemotron 3 Ultra (free)","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"}}},"abliteration-ai":{"id":"abliteration-ai","env":["ABLIT_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.abliteration.ai/v1","name":"abliteration.ai","doc":"https://docs.abliteration.ai/models","models":{"abliterated-model":{"id":"abliterated-model","name":"Abliterated Model","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]},{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-01-06","last_updated":"2026-07-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":150000,"input":150000,"output":8192},"cost":{"input":3,"output":3,"cache_read":0.3}},"abliterated-model-large-v2":{"id":"abliterated-model-large-v2","name":"Abliterated Model Large V2","description":"GLM-5.3 model abliterated and finetuned for cyber, ML red teaming, and agent testing","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-29","last_updated":"2026-08-31","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":999990},"cost":{"input":5,"output":5,"cache_read":0.5}},"abliterated-model-large":{"id":"abliterated-model-large","name":"Abliterated Model Large","description":"GLM-5.2 model abliterated and finetuned for cyber, ML red teaming, and agent testing","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]},{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-25","last_updated":"2026-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":999990},"cost":{"input":5,"output":5,"cache_read":0.5}}}},"vultr":{"id":"vultr","env":["VULTR_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.vultrinference.com/v1","name":"Vultr","doc":"https://api.vultrinference.com/","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.35},"canonical_model_id":"zhipuai/glm-5.3-flash"},"nemotron-3.5-content-safety":{"id":"nemotron-3.5-content-safety","name":"Nemotron 3.5 Content Safety","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.05,"output":0.15},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.1,"output":0.25},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"muse-glimmer-30b":{"id":"muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.25,"output":1},"canonical_model_id":"meta/muse-glimmer-30b"},"mimo-v2.6-flash-rl":{"id":"mimo-v2.6-flash-rl","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.25},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.75,"output":3},"canonical_model_id":"zhipuai/glm-5.3"},"glm-5.x-menthol":{"id":"glm-5.x-menthol","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.4,"output":1.75},"canonical_model_id":"zhipuai/glm-5"},"qwen3.8-flash-next":{"id":"qwen3.8-flash-next","name":"Qwen3.8 Flash Next","description":"Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.1,"output":0.2},"canonical_model_id":"alibaba/qwen3.8-flash-next"},"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.15,"output":1},"canonical_model_id":"alibaba/qwen3.8-27b"},"nemotron-3-nano-omni-30b-a3b-reasoning":{"id":"nemotron-3-nano-omni-30b-a3b-reasoning","name":"Nemotron 3 Nano Omni 30B A3B Reasoning","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.1,"output":0.25},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":512000},"cost":{"input":0.2,"output":0.9},"canonical_model_id":"minimax/MiniMax-M3"},"mimo-v2.6-pro-rl":{"id":"mimo-v2.6-pro-rl","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.4,"output":0.8},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.75,"output":3},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"laguna-s-2.1":{"id":"laguna-s-2.1","name":"Laguna S 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.09,"output":0.18},"canonical_model_id":"poolside/laguna-s-2.1"}}},"nan":{"id":"nan","env":["NAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.nan.builders/v1","name":"NaN","doc":"https://nan.builders/docs/models","models":{"qwen3.6":{"id":"qwen3.6","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen-image-2.1":{"id":"qwen-image-2.1","type":"image","name":"Qwen-Image-2.1","description":"Unified text-to-image generation and image editing model with native RGBA transparency and reference-image editing","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-14","last_updated":"2026-09-21","modalities":{"input":["text"],"output":["image"]},"open_weights":true,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen-image-2.1"},"glm5.3-flash":{"id":"glm5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"gemma4":{"id":"gemma4","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.8-flash"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"glm5.3":{"id":"glm5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3"}}},"sensenova":{"id":"sensenova","env":["SENSENOVA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://token.sensenova.cn/v1","name":"SenseNova (China)","doc":"https://platform.sensenova.cn/docs","models":{"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k3"},"sensenova-6.8-flash-lite":{"id":"sensenova-6.8-flash-lite","name":"SenseNova 6.8 Flash Lite","description":"SenseNova lightweight multimodal agent model for real-world complex tasks, data analysis, and complex information presentation","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-11","last_updated":"2026-08-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0,"cache_read":0}},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash is a high-performance conversational model by DeepSeek, supporting thinking/non-thinking modes, 1M context window, and tool calling.","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"}}},"vercel":{"id":"vercel","env":["AI_GATEWAY_API_KEY"],"npm":"@ai-sdk/gateway","name":"Vercel AI Gateway","doc":"https://github.com/vercel/ai/tree/5eb85cc45a259553501f535b8ac79a77d0e79223/packages/gateway","models":{"meituan/longcat-2.5-preview":{"id":"meituan/longcat-2.5-preview","name":"LongCat 2.5 Preview","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"longcat","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":false,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"meituan/longcat-2.5-preview"},"spacexai/grok-4.20-multi-agent":{"id":"spacexai/grok-4.20-multi-agent","name":"Grok 4.20 Multi-Agent","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"release_date":"2026-03-10","last_updated":"2026-03-10","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"spacexai/grok-4.20-multi-agent-beta":{"id":"spacexai/grok-4.20-multi-agent-beta","name":"Grok 4.20 Multi Agent Beta","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"spacexai/grok-imagine-video":{"id":"spacexai/grok-imagine-video","name":"Grok Imagine","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-01-28","last_updated":"2026-01-28","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"spacexai/grok-4.20-non-reasoning-beta":{"id":"spacexai/grok-4.20-non-reasoning-beta","name":"Grok 4.20 Beta Non-Reasoning","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.4,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"spacexai/grok-4.20-non-reasoning":{"id":"spacexai/grok-4.20-non-reasoning","name":"Grok 4.20 Non-Reasoning","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"spacexai/grok-4.1-fast-non-reasoning":{"id":"spacexai/grok-4.1-fast-non-reasoning","name":"Grok 4.1 Fast Non-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast"},"spacexai/grok-4.3":{"id":"spacexai/grok-4.3","name":"Grok 4.3","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"spacexai/grok-4.1-fast-reasoning":{"id":"spacexai/grok-4.1-fast-reasoning","name":"Grok 4.1 Fast Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast-reasoning"},"spacexai/grok-stt":{"id":"spacexai/grok-stt","name":"Grok STT","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"grok","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0}},"spacexai/grok-imagine-video-1.5":{"id":"spacexai/grok-imagine-video-1.5","type":"video","name":"Grok Imagine Video 1.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-05-30","last_updated":"2026-05-30","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"xai/grok-imagine-video-1.5"},"spacexai/grok-voice-think-fast-2.0":{"id":"spacexai/grok-voice-think-fast-2.0","name":"Grok Voice Think Fast 2.0","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"grok","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"spacexai/grok-4.20-reasoning-beta":{"id":"spacexai/grok-4.20-reasoning-beta","name":"Grok 4.20 Beta Reasoning","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"spacexai/grok-tts":{"id":"spacexai/grok-tts","type":"speech","name":"Grok TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"xai/grok-tts"},"spacexai/grok-4.7":{"id":"spacexai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"spacexai/grok-build-0.1":{"id":"spacexai/grok-build-0.1","name":"Grok Build 0.1","description":"Grok coding model for agentic engineering, edits, and codebase workflows","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}},"canonical_model_id":"xai/grok-build-0.1"},"spacexai/grok-4.5":{"id":"spacexai/grok-4.5","name":"Grok 4.5","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"spacexai/grok-imagine-image":{"id":"spacexai/grok-imagine-image","type":"image","name":"Grok Imagine Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-28","last_updated":"2026-01-28","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"xai/grok-imagine-image"},"spacexai/grok-imagine-image-2.0":{"id":"spacexai/grok-imagine-image-2.0","type":"image","name":"Grok Imagine Image 2.0","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-08-07","last_updated":"2026-08-07","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"xai/grok-imagine-image-2.0"},"spacexai/grok-4.20-reasoning":{"id":"spacexai/grok-4.20-reasoning","name":"Grok 4.20 Reasoning","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"spacexai/grok-4.6":{"id":"spacexai/grok-4.6","name":"Grok 4.6","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"spacexai/grok-imagine-video-1.5-lite":{"id":"spacexai/grok-imagine-video-1.5-lite","type":"video","name":"Grok Imagine Video 1.5 Lite","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"xai/grok-imagine-video-1.5-lite"},"spacexai/grok-voice-think-fast-1.0":{"id":"spacexai/grok-voice-think-fast-1.0","name":"Grok Voice Think Fast 1.0","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"grok","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"inception/mercury-2":{"id":"inception/mercury-2","name":"Mercury 2","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":50000},"cost":{"input":0.25,"output":0.75,"cache_read":0.025},"canonical_model_id":"inception/mercury-2"},"inception/mercury-2.5":{"id":"inception/mercury-2.5","name":"Mercury 2.5","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":260000,"output":65536},"cost":{"input":0.04,"output":0.15,"cache_read":0.004},"canonical_model_id":"inception/mercury-2.5"},"inception/mercury-3":{"id":"inception/mercury-3","name":"Mercury 3","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1,"output":0.35,"cache_read":0.01}},"inception/mercury-coder-small":{"id":"inception/mercury-coder-small","name":"Mercury Coder Small Beta","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"mercury","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-02-26","last_updated":"2025-02-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":16384},"cost":{"input":0.25,"output":1}},"anthropic/claude-opus-5.5-fast":{"id":"anthropic/claude-opus-5.5-fast","name":"Claude Opus 5.5 (Fast)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":8,"output":40,"cache_read":0.4,"cache_write":10},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4.8-fast":{"id":"anthropic/claude-opus-4.8-fast","name":"Claude Opus 4.8 (Fast)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100001}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-opus-5-fast":{"id":"anthropic/claude-opus-5-fast","name":"Claude Opus 5 (Fast)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"amazon/titan-embed-text-v2":{"id":"amazon/titan-embed-text-v2","name":"Titan Text Embeddings V2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"titan-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-04-30","last_updated":"2024-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536}},"amazon/nova-2-lite":{"id":"amazon/nova-2-lite","name":"Nova 2 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.3,"output":2.5,"cache_read":0.075},"canonical_model_id":"amazon/nova-2-lite"},"amazon/nova-pro":{"id":"amazon/nova-pro","name":"Nova Pro","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.8,"output":3.2,"cache_read":0.2,"cache_write":0.8},"canonical_model_id":"amazon/nova-pro"},"amazon/nova-lite":{"id":"amazon/nova-lite","name":"Nova Lite","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.06,"output":0.24,"cache_read":0.015,"cache_write":0.06},"canonical_model_id":"amazon/nova-lite"},"amazon/nova-micro":{"id":"amazon/nova-micro","name":"Nova Micro","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.035,"output":0.14,"cache_read":0.00875,"cache_write":0.035},"canonical_model_id":"amazon/nova-micro"},"topaz/starlight-precise-2.6":{"id":"topaz/starlight-precise-2.6","name":"Starlight Precise 2.6","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"topaz/wonder-3.5":{"id":"topaz/wonder-3.5","name":"Wonder 3.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"topaz/proteus":{"id":"topaz/proteus","name":"Proteus","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"fish-audio/s2.1-pro":{"id":"fish-audio/s2.1-pro","name":"S2.1 Pro","description":"Speech generation model for controllable voice, narration, and audio delivery","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-07-28","last_updated":"2026-07-28","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"fish-audio/s1":{"id":"fish-audio/s1","name":"S1","description":"Speech generation model for controllable voice, narration, and audio delivery","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2025-10-20","last_updated":"2025-10-20","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"fish-audio/drama-3":{"id":"fish-audio/drama-3","name":"Drama 3","description":"Speech generation model for controllable voice, narration, and audio delivery","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"fish-audio/s2-pro":{"id":"fish-audio/s2-pro","name":"S2 Pro","description":"Speech generation model for controllable voice, narration, and audio delivery","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"fish-audio/transcribe-1":{"id":"fish-audio/transcribe-1","name":"Transcribe-1","description":"Speech transcription model for accurate audio-to-text and captioning workflows","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-03-01","last_updated":"2026-03-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0}},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.076,"output":0.153,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-v3.2-thinking":{"id":"deepseek/deepseek-v3.2-thinking","name":"DeepSeek V3.2 Thinking","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8000},"cost":{"input":0.62,"output":1.85},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4.1-flash-1017-fast":{"id":"deepseek/deepseek-v4.1-flash-1017-fast","name":"DeepSeek V4.1 Flash 1017 Fast","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0.6,"output":2.4,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.066},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v3.1":{"id":"deepseek/deepseek-v3.1","name":"DeepSeek-V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":128000},"cost":{"input":0.25,"output":0.95,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.2156,"output":0.6468,"cache_read":0.0068},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.13,"output":0.26,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8000},"cost":{"input":0.62,"output":1.85},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.075,"output":0.3,"cache_read":0.0015},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.075,"output":0.3,"cache_read":0.0015},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4.1-flash-fast":{"id":"deepseek/deepseek-v4.1-flash-fast","name":"DeepSeek V4.1 Flash Fast","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.6,"output":2.4,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-r1":{"id":"deepseek/deepseek-r1","name":"DeepSeek-R1","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"status":"deprecated","cost":{"input":1.35,"output":5.4},"canonical_model_id":"deepseek/deepseek-r1"},"microsoft/mai-voice-2-flash":{"id":"microsoft/mai-voice-2-flash","name":"MAI-Voice-2-Flash","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mai","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"microsoft/mai-image-2.6-flash":{"id":"microsoft/mai-image-2.6-flash","name":"MAI-Image-2.6-Flash","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"mai","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":1.75,"output":19}},"microsoft/mai-voice-2.1":{"id":"microsoft/mai-voice-2.1","name":"MAI-Voice-2.1","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mai","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"microsoft/mai-voice-2.1-flash":{"id":"microsoft/mai-voice-2.1-flash","name":"MAI-Voice-2.1-Flash","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mai","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"microsoft/mai-transcribe-2":{"id":"microsoft/mai-transcribe-2","name":"MAI-Transcribe 2","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"mai","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-09-03","last_updated":"2026-09-03","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0}},"microsoft/mai-image-2.6":{"id":"microsoft/mai-image-2.6","name":"MAI-Image-2.6","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"mai","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":38}},"microsoft/mai-voice-2":{"id":"microsoft/mai-voice-2","name":"MAI-Voice-2","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mai","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"microsoft/mai-transcribe-2-streaming":{"id":"microsoft/mai-transcribe-2-streaming","name":"MAI-Transcribe-2-Streaming","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"mai","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0}},"nvidia/nemotron-3.5-lightning":{"id":"nvidia/nemotron-3.5-lightning","name":"Nemotron 3.5 Lightning 30B","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.05,"output":0.2,"cache_read":0.01},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-3.5-super-vl":{"id":"nvidia/nemotron-3.5-super-vl","name":"NVIDIA Nemotron 3.5 Super VL","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.2,"output":0.5,"cache_read":0.04},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/nemotron-3-ultra-550b-a55b":{"id":"nvidia/nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65000},"cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/nemotron-nano-9b-v2":{"id":"nvidia/nemotron-nano-9b-v2","name":"Nvidia Nemotron Nano 9B V2","description":"Compact Nemotron model for efficient reasoning and deployable AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2025-08-18","last_updated":"2025-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.06,"output":0.23},"canonical_model_id":"nvidia/nemotron-nano-9b-v2"},"nvidia/nemotron-3.5-lightning-free":{"id":"nvidia/nemotron-3.5-lightning-free","name":"Nemotron 3.5 Lightning Free","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65000},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-3-nano-30b-a3b":{"id":"nvidia/nemotron-3-nano-30b-a3b","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.05,"output":0.2,"cache_read":0.025},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/nemotron-3-super-120b-a12b":{"id":"nvidia/nemotron-3-super-120b-a12b","name":"NVIDIA Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"cost":{"input":0.15,"output":0.65},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/nemotron-nano-12b-v2-vl":{"id":"nvidia/nemotron-nano-12b-v2-vl","name":"Nvidia Nemotron Nano 12B V2 VL","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-10-28","last_updated":"2025-10-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.2,"output":0.6},"canonical_model_id":"nvidia/nemotron-nano-12b-v2-vl"},"inclusionai/ling-3.0-flash-sante":{"id":"inclusionai/ling-3.0-flash-sante","name":"Ling 3.0 Flash Sante","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash-sante"},"inclusionai/ling-3.0-flash-vl":{"id":"inclusionai/ling-3.0-flash-vl","name":"Ling 3.0 Flash VL","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash-vl"},"inclusionai/ling-3.0-flash":{"id":"inclusionai/ling-3.0-flash","name":"Ling 3.0 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"cost":{"input":0.021,"output":0.063,"cache_read":0.0042},"canonical_model_id":"inclusionai/ling-3.0-flash"},"inclusionai/ling-3.0-flash-fin":{"id":"inclusionai/ling-3.0-flash-fin","name":"Ling 3.0 Flash Fin","description":"Finance-enhanced model for financial research, multi-step investment workflows, and long-horizon planning and execution","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-09-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash-fin"},"inclusionai/ling-3.1-flash":{"id":"inclusionai/ling-3.1-flash","name":"Ling 3.1 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.06},"canonical_model_id":"inclusionai/ling-3.1-flash"},"meta/muse-spark-1.3-contributor":{"id":"meta/muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/muse-image-1.0":{"id":"meta/muse-image-1.0","name":"Muse Image 1.0","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"muse","attachment":true,"reasoning":false,"tool_call":false,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"meta/llama-3.1-8b":{"id":"meta/llama-3.1-8b","name":"Llama 3.1 8B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.22,"output":0.22},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta/muse-glimmer-30b":{"id":"meta/muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.35,"output":1.5,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"meta/muse-spark-1.2-contributor":{"id":"meta/muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"meta/llama-3.1-70b":{"id":"meta/llama-3.1-70b","name":"Llama 3.1 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-3.1-70b-instruct"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-open":{"id":"meta/muse-spark-open","name":"Muse Spark Open","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"meta/muse-wonder-1.0":{"id":"meta/muse-wonder-1.0","name":"Muse Wonder","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"meta/llama-3.3-70b":{"id":"meta/llama-3.3-70b","name":"Llama-3.3-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0}},"meta/llama-4-scout":{"id":"meta/llama-4-scout","name":"Llama-4-Scout-17B-16E-Instruct-FP8","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0}},"meta/llama-4-maverick":{"id":"meta/llama-4-maverick","name":"Llama-4-Maverick-17B-128E-Instruct-FP8","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0}},"heygen/heygen-video-1":{"id":"heygen/heygen-video-1","name":"HeyGen Video","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"release_date":"2026-10-09","last_updated":"2026-10-09","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"inference-net/schematron-v2-small":{"id":"inference-net/schematron-v2-small","name":"Schematron V2 Small","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":0.05,"output":0.23,"cache_read":0.05},"canonical_model_id":"inference-net/schematron-v2-small"},"inference-net/schematron-v2-turbo":{"id":"inference-net/schematron-v2-turbo","name":"Schematron V2 Turbo","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.03,"output":0.15,"cache_read":0.03},"canonical_model_id":"inference-net/schematron-v2-turbo"},"moonshotai/kimi-k3-fast":{"id":"moonshotai/kimi-k3-fast","name":"Kimi K3 Fast","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":4.5,"output":22.5,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2.7-code-highspeed":{"id":"moonshotai/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code High Speed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshotai/kimi-k2":{"id":"moonshotai/kimi-k2","name":"Kimi K2 Instruct","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-12-31","release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.57,"output":2.3},"canonical_model_id":"moonshotai/kimi-k2"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.6,"output":3},"canonical_model_id":"moonshotai/kimi-k2.5"},"poolside/laguna-s-2.1-free":{"id":"poolside/laguna-s-2.1-free","name":"Laguna S 2.1 Free","description":"Free provider route for experiments, demos, and cost-sensitive chat workloads","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-s-2.1"},"poolside/laguna-s-2.1":{"id":"poolside/laguna-s-2.1","name":"Laguna S 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.09,"output":0.18,"cache_read":0.009},"canonical_model_id":"poolside/laguna-s-2.1"},"fireworks/ember-1":{"id":"fireworks/ember-1","name":"Ember-1","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"fireworks/ember-1"},"bytedance/seedance-v1.0-pro-fast":{"id":"bytedance/seedance-v1.0-pro-fast","name":"Seedance v1.0 Pro Fast","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-10-24","last_updated":"2025-10-31","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"bytedance/seedance-v1.5-pro":{"id":"bytedance/seedance-v1.5-pro","name":"Seedance v1.5 Pro","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"bytedance/seedance-2.0-fast":{"id":"bytedance/seedance-2.0-fast","name":"Seedance 2.0 Fast","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"seed","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-04-14","last_updated":"2026-04-14","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"bytedance/seedance-2.0":{"id":"bytedance/seedance-2.0","type":"video","name":"Seedance 2.0","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"seed","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-10-04","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"bytedance-seed/seedance-2.0"},"bytedance/seed-1.8":{"id":"bytedance/seed-1.8","name":"Seed 1.8","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-28","last_updated":"2025-12-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32768},"cost":{"input":0.25,"output":2,"cache_read":0.05,"tiers":[{"input":0.5,"output":4,"cache_read":0.05,"tier":{"type":"context","size":128001}}]},"canonical_model_id":"bytedance-seed/seed-1-8"},"bytedance/seedream-4.5":{"id":"bytedance/seedream-4.5","name":"Seedream 4.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-03","last_updated":"2025-11-28","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"bytedance/seed-2.1-turbo":{"id":"bytedance/seed-2.1-turbo","name":"Seed 2.1 Turbo","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":2.5,"cache_read":0.1},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"bytedance/seed-1.6":{"id":"bytedance/seed-1.6","name":"Seed 1.6","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.25,"output":2,"cache_read":0.05,"tiers":[{"input":0.5,"output":4,"cache_read":0.05,"tier":{"type":"context","size":128001}}]},"canonical_model_id":"bytedance-seed/seed-1-6"},"bytedance/seedream-4.0":{"id":"bytedance/seedream-4.0","name":"Seedream 4.0","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-09-09","last_updated":"2025-08-28","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"bytedance/seedance-2.5":{"id":"bytedance/seedance-2.5","type":"video","name":"Seedance 2.5","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-07-31","last_updated":"2026-10-04","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"bytedance-seed/seedance-2.5"},"bytedance/seedance-v1.0-pro":{"id":"bytedance/seedance-v1.0-pro","name":"Seedance v1.0 Pro","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-06-11","last_updated":"2025-06-11","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"bytedance/seedream-5.0-lite":{"id":"bytedance/seedream-5.0-lite","name":"Seedream 5.0 Lite","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-01-28","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"bytedance/seedream-5.0-pro":{"id":"bytedance/seedream-5.0-pro","name":"Seedream 5.0 Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-11","last_updated":"2026-07-11","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"bytedance/seedance-2.0-mini":{"id":"bytedance/seedance-2.0-mini","name":"Seedance 2.0 Mini","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"seed","attachment":true,"reasoning":false,"tool_call":false,"release_date":"2026-06-22","last_updated":"2026-06-22","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":1000000},"cost":{"input":0.45,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"reflection/beam-free":{"id":"reflection/beam-free","name":"Beam 1.0 (Free)","description":"Free provider route for experiments, demos, and cost-sensitive chat workloads","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"reflection/beam":{"id":"reflection/beam","name":"Beam 1.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"stepfun/step-3.7-flash":{"id":"stepfun/step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"stepfun/step-5-preview":{"id":"stepfun/step-5-preview","name":"Step 5 Preview","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1,"output":2.7,"cache_read":0.1},"canonical_model_id":"stepfun/step-5-preview"},"quiverai/arrow-2-telos":{"id":"quiverai/arrow-2-telos","name":"Arrow 2 Telos","description":"High-fidelity SVG generation model for complex vector work and long-context refinement","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"release_date":"2026-09-16","last_updated":"2026-09-16","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":6,"output":30,"cache_read":0.6,"cache_write":7.5},"canonical_model_id":"quiverai/arrow-2-telos"},"quiverai/arrow-2":{"id":"quiverai/arrow-2","name":"Arrow 2","description":"Fast SVG generation model for creation, vectorization, editing, and animation","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"release_date":"2026-09-16","last_updated":"2026-09-16","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5},"canonical_model_id":"quiverai/arrow-2"},"bfl/flux-3-image":{"id":"bfl/flux-3-image","name":"FLUX 3 Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"bfl/flux-kontext-pro":{"id":"bfl/flux-kontext-pro","name":"FLUX.1 Kontext Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-05-29","last_updated":"2025-06","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0}},"bfl/flux-pro-1.1":{"id":"bfl/flux-pro-1.1","name":"FLUX1.1 [pro]","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-10-02","last_updated":"2024-10","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0}},"bfl/flux-3-image-klein":{"id":"bfl/flux-3-image-klein","name":"FLUX 3 Image Klein","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"bfl/flux-2-klein-9b":{"id":"bfl/flux-2-klein-9b","name":"FLUX.2 [klein] 9B","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-15","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"bfl/flux-2-flex":{"id":"bfl/flux-2-flex","name":"FLUX.2 [flex]","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-11-25","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"bfl/flux-pro-1.0-fill":{"id":"bfl/flux-pro-1.0-fill","name":"FLUX.1 Fill [pro]","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-10-01","last_updated":"2024-10","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0}},"bfl/flux-3-video":{"id":"bfl/flux-3-video","name":"Flux 3","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-08-04","last_updated":"2026-08-04","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"bfl/flux-pro-1.1-ultra":{"id":"bfl/flux-pro-1.1-ultra","name":"FLUX1.1 [pro] Ultra","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-11-01","last_updated":"2024-11","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0}},"bfl/flux-2-max":{"id":"bfl/flux-2-max","name":"FLUX.2 [max]","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-16","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":67300,"output":67300}},"bfl/flux-kontext-max":{"id":"bfl/flux-kontext-max","name":"FLUX.1 Kontext Max","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-05-29","last_updated":"2025-06","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0}},"bfl/flux-2-klein-4b":{"id":"bfl/flux-2-klein-4b","name":"FLUX.2 [klein] 4B","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-15","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"bfl/flux-2-pro":{"id":"bfl/flux-2-pro","name":"FLUX.2 [pro]","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-11-25","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":67300,"output":67300}},"prodia/flux-fast-schnell":{"id":"prodia/flux-fast-schnell","name":"Flux Schnell","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-08-02","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0}},"minimax/minimax-m2.1-lightning":{"id":"minimax/minimax-m2.1-lightning","name":"MiniMax M2.1 Lightning","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-12-23","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":2.4,"cache_read":0.03,"cache_write":0.375}},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"Minimax M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":205000,"output":196608},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-h3-max":{"id":"minimax/minimax-h3-max","name":"MiniMax H3 Max","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"minimax","attachment":true,"reasoning":false,"tool_call":false,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-m2.5-highspeed":{"id":"minimax/minimax-m2.5-highspeed","name":"MiniMax M2.5 High Speed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131000},"cost":{"input":0.6,"output":2.4,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5-highspeed"},"minimax/minimax-m3.1-flash":{"id":"minimax/minimax-m3.1-flash","name":"MiniMax-M3.1-Flash","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0,"output":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax M2.7 High Speed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-h3":{"id":"minimax/minimax-h3","type":"video","name":"MiniMax H3","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"minimax","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-07-31","last_updated":"2026-10-04","modalities":{"input":["text","image"],"output":["video"]},"open_weights":true,"limit":{"context":0,"output":0},"canonical_model_id":"minimax/MiniMax-H3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131000},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5"},"elorian/glimpse":{"id":"elorian/glimpse","name":"Glimpse","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":false,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":220000,"output":4096},"cost":{"input":0.6,"output":3,"cache_read":0.06}},"callstack/apex":{"id":"callstack/apex","name":"Apex","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"release_date":"2026-10-08","last_updated":"2026-10-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.5,"output":3,"cache_read":0.2}},"liquid/lfm2.5-24b-a2b":{"id":"liquid/lfm2.5-24b-a2b","name":"LFM2.5-24B-A2B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"liquid","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8200},"cost":{"input":0.05,"output":0.7,"cache_read":0.025}},"klingai/kling-v2.6-i2v":{"id":"klingai/kling-v2.6-i2v","name":"Kling v2.6 Image-to-Video","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"ling","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-03","last_updated":"2025-12-21","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"klingai/kling-v2.6-motion-control":{"id":"klingai/kling-v2.6-motion-control","name":"Kling v2.6 Motion Control","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"ling","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-21","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"klingai/kling-v2.5-turbo-i2v":{"id":"klingai/kling-v2.5-turbo-i2v","name":"Kling v2.5 Turbo Image-to-Video","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"ling","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"klingai/kling-v2.5-turbo-t2v":{"id":"klingai/kling-v2.5-turbo-t2v","name":"Kling v2.5 Turbo Text-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"ling","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"klingai/kling-v2.6-t2v":{"id":"klingai/kling-v2.6-t2v","name":"Kling v2.6 Text-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"ling","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-03","last_updated":"2025-12-21","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"klingai/kling-v3.0-t2v":{"id":"klingai/kling-v3.0-t2v","name":"Kling v3.0 Text-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"ling","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"klingai/kling-v3.0-motion-control":{"id":"klingai/kling-v3.0-motion-control","name":"Kling v3.0 Motion Control","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"ling","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-03-04","last_updated":"2026-03-04","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"klingai/kling-v3.0-i2v":{"id":"klingai/kling-v3.0-i2v","name":"Kling v3.0 Image-to-Video","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"ling","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"cohere/rerank-v3.5":{"id":"cohere/rerank-v3.5","type":"reranking","name":"Cohere Rerank 3.5","description":"Reranking model for improving retrieval quality in search and recommendation systems","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-12-02","last_updated":"2024-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4096,"output":4096},"canonical_model_id":"cohere/rerank-v3.5"},"cohere/rerank-v4-pro":{"id":"cohere/rerank-v4-pro","type":"reranking","name":"Cohere Rerank 4 Pro","description":"Reranking model for improving retrieval quality in search and recommendation systems","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"canonical_model_id":"cohere/rerank-v4.0-pro"},"cohere/embed-v5.0-fast":{"id":"cohere/embed-v5.0-fast","type":"embedding","name":"Embed 5 Fast","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"cohere-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-30","last_updated":"2026-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":2048},"canonical_model_id":"cohere/embed-v5.0-fast"},"cohere/command-a":{"id":"cohere/command-a","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a"},"cohere/embed-v4.0":{"id":"cohere/embed-v4.0","type":"embedding","name":"Embed v4.0","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"cohere-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":1536},"canonical_model_id":"cohere/embed-v4.0"},"cohere/embed-v5.0-pro":{"id":"cohere/embed-v5.0-pro","type":"embedding","name":"Embed 5 Pro","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"cohere-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-30","last_updated":"2026-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":2048},"canonical_model_id":"cohere/embed-v5.0-pro"},"cohere/rerank-v4-fast":{"id":"cohere/rerank-v4-fast","type":"reranking","name":"Cohere Rerank 4 Fast","description":"Reranking model for improving retrieval quality in search and recommendation systems","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"canonical_model_id":"cohere/rerank-v4.0-fast"},"perplexity/pplx-embed-v1-4b":{"id":"perplexity/pplx-embed-v1-4b","name":"Embed v1 4b","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0}},"perplexity/pplx-embed-v1-0.6b":{"id":"perplexity/pplx-embed-v1-0.6b","name":"Embed v1 0.6b","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"v0","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0}},"perplexity/sonar":{"id":"perplexity/sonar","name":"Sonar","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-02","release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":127000,"output":8000},"cost":{"input":0.25,"output":2.5,"cache_write":0.0625}},"xiaomi/mimo-v2.5-pro":{"id":"xiaomi/mimo-v2.5-pro","name":"MiMo V2.5 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":131000},"cost":{"input":0.3045,"output":0.609,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.6-pro-ultraspeed":{"id":"xiaomi/mimo-v2.6-pro-ultraspeed","name":"MiMo V2.6 Pro UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.35,"output":8.7,"cache_read":0.036},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo M2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":131100},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2.6-flash":{"id":"xiaomi/mimo-v2.6-flash","name":"MiMo V2.6 Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo V2.6 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"mixedbread/toast-1":{"id":"mixedbread/toast-1","name":"Toast 1","description":"Specialized search model for knowledge-intensive questions, multi-step retrieval, and evidence synthesis","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":4000},"cost":{"input":0.3,"output":0.72,"cache_read":0.036},"canonical_model_id":"mixedbread/toast-1"},"interfaze/interfaze-beta":{"id":"interfaze/interfaze-beta","name":"Interfaze Beta","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"temperature":true,"release_date":"2025-10-07","last_updated":"2026-04-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":1.5,"output":3.5}},"zai/glm-5.3-flash":{"id":"zai/glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131000},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai/glm-5.3-fast":{"id":"zai/glm-5.3-fast","name":"GLM 5.3 Fast","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":2.1,"output":6.6,"cache_read":0.21},"canonical_model_id":"zhipuai/glm-5.3"},"zai/glm-5.2-fast":{"id":"zai/glm-5.2-fast","name":"GLM 5.2 Fast","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":2.8,"output":8.8,"cache_read":0.56},"canonical_model_id":"zhipuai/glm-5.2"},"zai/glm-4.5v":{"id":"zai/glm-4.5v","name":"GLM 4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":66000,"output":16000},"cost":{"input":0.6,"output":1.8,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5v"},"zai/glm-4.5-air":{"id":"zai/glm-4.5-air","name":"GLM 4.5 Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":96000},"cost":{"input":0.2,"output":1.1,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-4.5-air"},"zai/glm-5.3":{"id":"zai/glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.3"},"zai/glm-4.7":{"id":"zai/glm-4.7","name":"GLM 4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":120000},"cost":{"input":0.6,"output":2.2},"canonical_model_id":"zhipuai/glm-4.7"},"zai/glm-5.1":{"id":"zai/glm-5.1","name":"GLM 5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":64000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai/glm-4.7-flash":{"id":"zai/glm-4.7-flash","name":"GLM 4.7 Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131000},"cost":{"input":0.07,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"zai/glm-5":{"id":"zai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":131100},"cost":{"input":1,"output":3.2},"canonical_model_id":"zhipuai/glm-5"},"zai/glm-5v-turbo":{"id":"zai/glm-5v-turbo","name":"GLM 5V Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"zai/glm-5.3-flashx":{"id":"zai/glm-5.3-flashx","name":"GLM 5.3 FlashX","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.37,"output":1.25,"cache_read":0.075},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai/glm-5-turbo":{"id":"zai/glm-5-turbo","name":"GLM 5 Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202800,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"zai/glm-4.6":{"id":"zai/glm-4.6","name":"GLM 4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":96000},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"zai/glm-5.2":{"id":"zai/glm-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.8,"output":2.55,"cache_read":0.16},"canonical_model_id":"zhipuai/glm-5.2"},"zai/glm-4.5":{"id":"zai/glm-4.5","name":"GLM 4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":96000},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5"},"zai/glm-4.7-flashx":{"id":"zai/glm-4.7-flashx","name":"GLM 4.7 FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.06,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"voyage/voyage-code-2":{"id":"voyage/voyage-code-2","name":"voyage-code-2","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-01-01","last_updated":"2024-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536}},"voyage/voyage-finance-2":{"id":"voyage/voyage-finance-2","name":"voyage-finance-2","description":"General-purpose chat model for instruction following, writing, and analysis","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-06-03","last_updated":"2024-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536}},"voyage/voyage-4-large":{"id":"voyage/voyage-4-large","name":"voyage-4-large","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-15","last_updated":"2026-03-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0}},"voyage/voyage-4-lite":{"id":"voyage/voyage-4-lite","name":"voyage-4-lite","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-15","last_updated":"2026-03-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0}},"voyage/voyage-4":{"id":"voyage/voyage-4","name":"voyage-4","description":"General-purpose chat model for instruction following, writing, and analysis","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-15","last_updated":"2026-03-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0}},"voyage/rerank-2.5":{"id":"voyage/rerank-2.5","name":"Voyage Rerank 2.5","description":"Reranking model for improving retrieval quality in search and recommendation systems","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000}},"voyage/rerank-2.5-lite":{"id":"voyage/rerank-2.5-lite","name":"Voyage Rerank 2.5 Lite","description":"Reranking model for improving retrieval quality in search and recommendation systems","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000}},"voyage/rerank-3-lite":{"id":"voyage/rerank-3-lite","name":"Voyage Rerank 3 Lite","description":"Reranking model for improving retrieval quality in search and recommendation systems","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-09-30","last_updated":"2026-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000}},"voyage/voyage-law-2":{"id":"voyage/voyage-law-2","name":"voyage-law-2","description":"General-purpose chat model for instruction following, writing, and analysis","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-04-15","last_updated":"2024-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536}},"voyage/voyage-code-3":{"id":"voyage/voyage-code-3","name":"voyage-code-3","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-12-04","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536}},"voyage/rerank-3":{"id":"voyage/rerank-3","name":"Voyage Rerank 3","description":"Reranking model for improving retrieval quality in search and recommendation systems","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-09-30","last_updated":"2026-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000}},"voyage/voyage-3.5":{"id":"voyage/voyage-3.5","name":"voyage-3.5","description":"General-purpose chat model for instruction following, writing, and analysis","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536}},"voyage/voyage-3.5-lite":{"id":"voyage/voyage-3.5-lite","name":"voyage-3.5-lite","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536}},"voyage/voyage-3-large":{"id":"voyage/voyage-3-large","name":"voyage-3-large","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-01-07","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536}},"google/text-multilingual-embedding-002":{"id":"google/text-multilingual-embedding-002","name":"Text Multilingual Embedding 002","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-03-01","last_updated":"2024-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536}},"google/gemini-3.8-live":{"id":"google/gemini-3.8-live","type":"realtime","name":"Gemini 3.8 Live","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gemini-flash","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-09-15","last_updated":"2026-09-15","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0.75,"output":4.5},"canonical_model_id":"google/gemini-3.8-live"},"google/gemini-3-flash":{"id":"google/gemini-3-flash","name":"Gemini 3 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65000},"cost":{"input":0.5,"output":3,"cache_read":0.05}},"google/gemini-3.1-flash-image":{"id":"google/gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3.1-flash-image"},"google/gemini-3.1-flash-image-preview":{"id":"google/gemini-3.1-flash-image-preview","type":"image","name":"Gemini 3.1 Flash Image Preview (Nano Banana 2)","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3.1-flash-image-preview"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-3.5-transcribe-live":{"id":"google/gemini-3.5-transcribe-live","type":"transcription","name":"Gemini 3.5 Transcribe Live","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"google/gemini-3.5-transcribe-live"},"google/veo-3.0-generate-001":{"id":"google/veo-3.0-generate-001","name":"Veo 3.0","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-05-20","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"google/gemini-embedding-001":{"id":"google/gemini-embedding-001","type":"embedding","name":"Gemini Embedding 001","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536},"canonical_model_id":"google/gemini-embedding-001"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65000},"cost":{"input":0.25,"output":1.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.8-flash-tts":{"id":"google/gemini-3.8-flash-tts","type":"speech","name":"Gemini 3.8 Flash TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gemini-flash","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0.5,"output":9},"canonical_model_id":"google/gemini-3.8-flash-tts"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-embedding-2":{"id":"google/gemini-embedding-2","type":"embedding","name":"Gemini Embedding 2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-11","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"google/gemini-embedding-2"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.8-flash"},"google/veo-3.0-fast-generate-001":{"id":"google/veo-3.0-fast-generate-001","name":"Veo 3.0 Fast Generate","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-07-31","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"google/gemini-3-pro-image":{"id":"google/gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"google/gemini-3-pro-image"},"google/gemini-3.8-live-extended-thinking":{"id":"google/gemini-3.8-live-extended-thinking","type":"realtime","name":"Gemini 3.8 Live Extended Thinking","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gemini-flash","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-09-15","last_updated":"2026-09-15","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0.75,"output":4.5},"canonical_model_id":"google/gemini-3.8-live-extended-thinking"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"google/gemini-3.5-flash"},"google/veo-3.1-generate-001":{"id":"google/veo-3.1-generate-001","name":"Veo 3.1","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-10-15","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"google/gemini-3.1-flash-lite-image":{"id":"google/gemini-3.1-flash-lite-image","type":"image","name":"Gemini 3.1 Flash Lite Image (Nano Banana 2 Lite)","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":4096},"cost":{"input":0.25,"output":1.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"google/gemini-nano-banana-2.1":{"id":"google/gemini-nano-banana-2.1","type":"image","name":"Gemini Nano Banana 2.1","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-nano-banana-2.1"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Gemini multimodal model for text, image, audio, video, and document tasks","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"google/gemini-4-argon"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/text-embedding-005":{"id":"google/text-embedding-005","name":"Text Embedding 005","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-08-01","last_updated":"2024-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536}},"google/gemini-3.8-flash-lite-tts":{"id":"google/gemini-3.8-flash-lite-tts","type":"speech","name":"Gemini 3.8 Flash-Lite TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gemini-flash-lite","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0.5,"output":6},"canonical_model_id":"google/gemini-3.8-flash-lite-tts"},"google/veo-3.1-lite-generate-001":{"id":"google/veo-3.1-lite-generate-001","name":"Veo 3.1 Lite Generate","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65000},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-3.5-transcribe":{"id":"google/gemini-3.5-transcribe","type":"transcription","name":"Gemini 3.5 Transcribe","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":2,"output":12},"canonical_model_id":"google/gemini-3.5-transcribe"},"google/veo-3.1-fast-generate-001":{"id":"google/veo-3.1-fast-generate-001","name":"Veo 3.1 Fast Generate","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-10-15","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-omni-flash-preview":{"id":"google/gemini-omni-flash-preview","type":"video","name":"Gemini Omni Flash Preview","description":"Omni-modal model for text, vision, audio, and multimodal agent tasks","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":false,"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":57920},"cost":{"input":1.5,"output":9},"canonical_model_id":"google/gemini-omni-flash-preview"},"google/gemini-2.5-flash-image":{"id":"google/gemini-2.5-flash-image","type":"image","name":"Nano Banana (Gemini 2.5 Flash Image)","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":65535},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-2.5-flash-image"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Tencent Hy4 Preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.14,"output":0.58,"cache_read":0.035},"canonical_model_id":"tencent/hy3"},"tencent/hy-mt2-lite":{"id":"tencent/hy-mt2-lite","name":"Tencent Hy-MT2-Lite","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":4000},"cost":{"input":0.044,"output":0.177}},"tencent/hy-mt2-plus":{"id":"tencent/hy-mt2-plus","name":"Tencent Hy-MT2-Plus","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":4000},"cost":{"input":0.074,"output":0.295}},"tencent/hy-mt2-pro":{"id":"tencent/hy-mt2-pro","name":"Tencent Hy-MT2-Pro","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":4000},"cost":{"input":0.074,"output":0.295}},"sakana/fugu-max":{"id":"sakana/fugu-max","name":"Fugu Max","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"sakana/fugu-max"},"sakana/namazu":{"id":"sakana/namazu","name":"Sakana Namazu","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"sakana-namazu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.95,"output":4,"cache_read":0.15},"canonical_model_id":"sakana/sakana-namazu"},"sakana/fugu-ultra-v2":{"id":"sakana/fugu-ultra-v2","name":"Fugu Ultra v2","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-08-28","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra-v2"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra"},"openai/gpt-5.2-fast":{"id":"openai/gpt-5.2-fast","name":"GPT 5.2 (Fast)","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":3.5,"output":28,"cache_read":0.35},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT 5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":872000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-live-1":{"id":"openai/gpt-live-1","type":"realtime","name":"GPT-Live 1","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-07-31","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"openai/gpt-live-1"},"openai/gpt-image-2.5-sunburst":{"id":"openai/gpt-image-2.5-sunburst","type":"image","name":"GPT Image 2.5 Sunburst","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-sunburst"},"openai/tts-1":{"id":"openai/tts-1","type":"speech","name":"TTS-1","description":"Speech generation model for controllable voice, narration, and audio delivery","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"openai/tts-1"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT 5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-image-2.5-flare":{"id":"openai/gpt-image-2.5-flare","type":"image","name":"GPT Image 2.5 Flare","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-flare"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":128000,"output":272000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/o3-fast":{"id":"openai/o3-fast","name":"o3 (Fast)","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":100000,"output":100000},"cost":{"input":3.5,"output":14,"cache_read":0.875},"canonical_model_id":"openai/o3"},"openai/text-embedding-3-large":{"id":"openai/text-embedding-3-large","type":"embedding","name":"text-embedding-3-large","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"input":6656,"output":1536},"canonical_model_id":"openai/text-embedding-3-large"},"openai/gpt-6.1-sol-fast":{"id":"openai/gpt-6.1-sol-fast","name":"GPT-6.1 Sol (Fast)","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.4,"cache_write":10,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.4,"cache_write":10}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/o3-pro":{"id":"openai/o3-pro","name":"o3 Pro","description":"High-effort o3 tier for difficult technical reasoning and careful answers","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":100000,"output":100000},"cost":{"input":20,"output":80},"canonical_model_id":"openai/o3-pro"},"openai/gpt-realtime-2.1":{"id":"openai/gpt-realtime-2.1","type":"realtime","name":"gpt-realtime-2.1","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gpt","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-09-30","release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":128000,"input":96000,"output":32000},"cost":{"input":4,"output":24,"cache_read":0.4},"canonical_model_id":"openai/gpt-realtime-2.1"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT 5.6 Terra","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/whisper-1":{"id":"openai/whisper-1","type":"transcription","name":"Whisper","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2023-03-01","last_updated":"2023-03-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"openai/whisper-1"},"openai/gpt-5.1-thinking":{"id":"openai/gpt-5.1-thinking","name":"GPT 5.1 Thinking","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-11-12","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"openai/gpt-5.4-mini-fast":{"id":"openai/gpt-5.4-mini-fast","name":"GPT 5.4 Mini (Fast)","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.5-fast":{"id":"openai/gpt-5.5-fast","name":"GPT 5.5 (Fast)","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":872000,"output":128000},"cost":{"input":12.5,"output":75,"cache_read":1.25},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"input":12289,"output":4096},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-image-1":{"id":"openai/gpt-image-1","type":"image","name":"GPT Image 1","description":"OpenAI image model for production generation, edits, and brand-safe visual workflows","family":"gpt-image","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-04-24","last_updated":"2025-04-24","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":40,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-1"},"openai/gpt-5-codex":{"id":"openai/gpt-5-codex","name":"GPT-5-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5-codex"},"openai/gpt-image-1.5":{"id":"openai/gpt-image-1.5","type":"image","name":"GPT Image 1.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":32,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-1.5"},"openai/tts-1-hd":{"id":"openai/tts-1-hd","type":"speech","name":"TTS-1 HD","description":"Speech generation model for controllable voice, narration, and audio delivery","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"openai/tts-1-hd"},"openai/gpt-4o-transcribe":{"id":"openai/gpt-4o-transcribe","type":"transcription","name":"GPT-4o Transcribe","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-03-20","last_updated":"2025-03-20","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":2.5,"output":10},"canonical_model_id":"openai/gpt-4o-transcribe"},"openai/gpt-5.4-fast":{"id":"openai/gpt-5.4-fast","name":"GPT 5.4 (Fast)","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-5.3-codex-fast":{"id":"openai/gpt-5.3-codex-fast","name":"GPT 5.3 Codex (Fast)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":3.5,"output":28,"cache_read":0.35},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/gpt-5.6-terra-fast":{"id":"openai/gpt-5.6-terra-fast","name":"GPT 5.6 Terra (Fast)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":24,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":36,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":36,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT 5.6 Luna","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5.1-thinking-fast":{"id":"openai/gpt-5.1-thinking-fast","name":"GPT 5.1 Thinking (Fast)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"release_date":"2025-11-12","last_updated":"2025-11-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":2.5,"output":20,"cache_read":0.25}},"openai/gpt-oss-safeguard-20b":{"id":"openai/gpt-oss-safeguard-20b","name":"gpt-oss-safeguard-20b","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":112000,"output":16000},"cost":{"input":0.07,"output":0.2},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"openai/gpt-image-2":{"id":"openai/gpt-image-2","type":"image","name":"GPT Image 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT 5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-5-mini-fast":{"id":"openai/gpt-5-mini-fast","name":"GPT-5 mini (Fast)","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.45,"output":3.6,"cache_read":0.045},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-image-1-mini":{"id":"openai/gpt-image-1-mini","name":"GPT Image 1 Mini","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":2,"output":8,"cache_read":0.2}},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":122880,"output":8192},"cost":{"input":0.03,"output":0.14},"canonical_model_id":"openai/gpt-oss-20b"},"openai/text-embedding-ada-002":{"id":"openai/text-embedding-ada-002","type":"embedding","name":"text-embedding-ada-002","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2022-12-15","last_updated":"2022-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"input":6656,"output":1536},"canonical_model_id":"openai/text-embedding-ada-002"},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2-Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.03},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/gpt-realtime-whisper":{"id":"openai/gpt-realtime-whisper","type":"transcription","name":"gpt-realtime-whisper","description":"Streaming speech-to-text model for low-latency transcript deltas from live audio","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"openai/gpt-realtime-whisper"},"openai/o4-mini-fast":{"id":"openai/o4-mini-fast","name":"o4-mini (Fast)","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":100000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o4-mini"},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT-5.1-Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5.1-codex"},"openai/gpt-4.1-nano-fast":{"id":"openai/gpt-4.1-nano-fast","name":"GPT-4.1 nano (Fast)","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"input":1014808,"output":32768},"cost":{"input":0.2,"output":0.8,"cache_read":0.05},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-6-astra-fast":{"id":"openai/gpt-6-astra-fast","name":"GPT-6 Astra (Fast)","description":"Fast variant of GPT-6 Astra for low-latency assistance and high-volume workloads.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":20,"output":100,"cache_read":2,"cache_write":25,"tiers":[{"input":40,"output":150,"cache_read":4,"cache_write":50,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":40,"output":150,"cache_read":4,"cache_write":50}},"canonical_model_id":"openai/gpt-6-astra-fast"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT 5.2 ","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5.6-sol-fast":{"id":"openai/gpt-5.6-sol-fast","name":"GPT 5.6 Sol (Fast)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":8,"output":40,"cache_read":0.8,"cache_write":10,"tiers":[{"input":16,"output":60,"cache_read":1.6,"cache_write":20,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":16,"output":60,"cache_read":1.6,"cache_write":20}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6-luna-fast":{"id":"openai/gpt-6-luna-fast","name":"GPT-6 Luna (Fast)","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.5,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":0.4,"output":1.5,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-4o-mini-fast":{"id":"openai/gpt-4o-mini-fast","name":"GPT-4o mini (Fast)","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":111616,"output":16384},"cost":{"input":0.25,"output":1,"cache_read":0.125},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-oss-safeguard-120b":{"id":"openai/gpt-oss-safeguard-120b","name":"GPT OSS Safeguard 120B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":112000,"output":16000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-safeguard-120b"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT 5.6 Sol","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-6-sol-fast":{"id":"openai/gpt-6-sol-fast","name":"GPT-6 Sol (Fast)","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-realtime-2":{"id":"openai/gpt-realtime-2","type":"realtime","name":"gpt-realtime-2","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-09-30","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":4,"output":24,"cache_read":0.4},"canonical_model_id":"openai/gpt-realtime-2"},"openai/gpt-4.1-mini-fast":{"id":"openai/gpt-4.1-mini-fast","name":"GPT-4.1 mini (Fast)","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"input":1014808,"output":32768},"cost":{"input":0.7,"output":2.8,"cache_read":0.175},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-4o-fast":{"id":"openai/gpt-4o-fast","name":"GPT-4o (Fast)","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":111616,"output":16384},"cost":{"input":4.25,"output":17,"cache_read":2.125},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT 5.4 Nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-realtime-mini":{"id":"openai/gpt-realtime-mini","type":"realtime","name":"GPT-Realtime mini","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gpt-mini","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-10-01","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0.6,"output":2.4,"cache_read":0.06},"canonical_model_id":"openai/gpt-realtime-mini"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.1,"output":0.5,"cache_read":0.1},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai/text-embedding-3-small":{"id":"openai/text-embedding-3-small","type":"embedding","name":"text-embedding-3-small","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"input":6656,"output":1536},"canonical_model_id":"openai/text-embedding-3-small"},"openai/gpt-5.1-codex-max":{"id":"openai/gpt-5.1-codex-max","name":"GPT 5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex-max"},"openai/gpt-4o-mini-transcribe":{"id":"openai/gpt-4o-mini-transcribe","type":"transcription","name":"GPT-4o mini Transcribe","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"gpt-mini","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-03-20","last_updated":"2025-03-20","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":1.25,"output":5},"canonical_model_id":"openai/gpt-4o-mini-transcribe"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT 5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT 5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":872000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-realtime-1.5":{"id":"openai/gpt-realtime-1.5","type":"realtime","name":"GPT-Realtime-1.5","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-09-30","release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":4,"output":16,"cache_read":0.4},"canonical_model_id":"openai/gpt-realtime-1.5"},"openai/gpt-4.1-fast":{"id":"openai/gpt-4.1-fast","name":"GPT-4.1 (Fast)","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"input":1014808,"output":32768},"cost":{"input":3.5,"output":14,"cache_read":0.875},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5-fast":{"id":"openai/gpt-5-fast","name":"GPT-5 (Fast)","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":2.5,"output":20,"cache_read":0.25},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT 5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/gpt-5.6-luna-fast":{"id":"openai/gpt-5.6-luna-fast","name":"GPT 5.6 Luna (Fast)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.4,"output":2.4,"cache_read":0.04,"cache_write":0.5,"tiers":[{"input":0.8,"output":3.6,"cache_read":0.08,"cache_write":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.8,"output":3.6,"cache_read":0.08,"cache_write":1}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"flex":{"cost":{"input":0.025,"output":0.2,"cache_read":0.0025},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":0.05,"output":0.4,"cache_read":0.005}},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":10,"output":30}},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":15,"output":60,"cache_read":7.5}},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":14,"cache_read":0.875},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":1,"output":4,"cache_read":0.25},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":2,"output":8,"cache_read":0.5}},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"experimental":{"modes":{"fast":{"cost":{"input":4.25,"output":17,"cache_read":2.125},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":10,"cache_read":1.25}},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":0.45,"output":3.6,"cache_read":0.045},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.125,"output":1,"cache_read":0.0125},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":14,"cache_read":0.875},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2,"output":8,"cache_read":0.5}},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.55}},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":2.5,"output":20,"cache_read":0.25},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.625,"output":5,"cache_read":0.0625},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"experimental":{"modes":{"fast":{"cost":{"input":0.7,"output":2.8,"cache_read":0.175},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.4,"output":1.6,"cache_read":0.1}},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","experimental":{"modes":{"fast":{"cost":{"input":0.2,"output":0.8,"cache_read":0.05},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.1,"output":0.4,"cache_read":0.025}},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"experimental":{"modes":{"fast":{"cost":{"input":0.25,"output":1,"cache_read":0.125},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.15,"output":0.6,"cache_read":0.075}},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","experimental":{"modes":{"fast":{"cost":{"input":2,"output":8,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.55,"output":2.2,"cache_read":0.1375},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":1.1,"output":4.4,"cache_read":0.275}},"morph/morph-v3-large":{"id":"morph/morph-v3-large","name":"Morph v3 Large","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-08-15","last_updated":"2024-08-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"cost":{"input":0.9,"output":1.9}},"morph/morph-v3-fast":{"id":"morph/morph-v3-fast","name":"Morph v3 Fast","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-08-15","last_updated":"2024-08-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":16000},"cost":{"input":0.8,"output":1.2}},"stealth/glyph-cluster":{"id":"stealth/glyph-cluster","name":"Glyph Cluster","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0,"output":0}},"recraft/recraft-v4.1-flash":{"id":"recraft/recraft-v4.1-flash","name":"Recraft V4.1 Flash","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"recraft/recraft-v4.1":{"id":"recraft/recraft-v4.1","name":"Recraft V4.1","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-05-14","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"recraft/recraft-v4":{"id":"recraft/recraft-v4","name":"Recraft V4","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"recraft/recraft-v4.1-pro":{"id":"recraft/recraft-v4.1-pro","name":"Recraft V4.1 Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-05-14","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"recraft/recraft-v2":{"id":"recraft/recraft-v2","name":"Recraft V2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-03-13","last_updated":"2024-03","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0}},"recraft/recraft-v4.1-utility":{"id":"recraft/recraft-v4.1-utility","name":"Recraft V4.1 Utility","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-05-14","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"recraft/recraft-v4-pro":{"id":"recraft/recraft-v4-pro","name":"Recraft V4 Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"recraft/recraft-v3":{"id":"recraft/recraft-v3","name":"Recraft V3","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-10-30","last_updated":"2024-10","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0}},"recraft/recraft-v4.1-utility-pro":{"id":"recraft/recraft-v4.1-utility-pro","name":"Recraft V4.1 Utility Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-05-14","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"alibaba/qwen3.7-flash":{"id":"alibaba/qwen3.7-flash","name":"Qwen 3.7 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":64000},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.038,"tiers":[{"input":0.1,"output":0.4,"cache_read":0.02,"cache_write":0.125,"tier":{"type":"context","size":32000}},{"input":0.2,"output":0.8,"cache_read":0.04,"cache_write":0.25,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"alibaba/wan-v2.6-i2v":{"id":"alibaba/wan-v2.6-i2v","name":"Wan v2.6 Image-to-Video","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"alibaba/qwen3-embedding-4b":{"id":"alibaba/qwen3-embedding-4b","type":"embedding","name":"Qwen3 Embedding 4B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"canonical_model_id":"alibaba/qwen3-embedding-4b"},"alibaba/wan-v2.7-r2v":{"id":"alibaba/wan-v2.7-r2v","name":"Wan v2.7 Reference-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"alibaba/qwen-3-14b":{"id":"alibaba/qwen-3-14b","name":"Qwen3-14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.24}},"alibaba/qwen3-coder-30b-a3b":{"id":"alibaba/qwen3-coder-30b-a3b","name":"Qwen 3 Coder 30B A3B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"alibaba/qwen3.5-plus":{"id":"alibaba/qwen3.5-plus","name":"Qwen 3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1,"max":81920}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":2.4,"cache_read":0.04,"cache_write":0.5,"tiers":[{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":256001}}],"context_over_200k":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625}},"canonical_model_id":"alibaba/qwen3.5-plus"},"alibaba/qwen3.8-omni-flash":{"id":"alibaba/qwen3.8-omni-flash","name":"Qwen 3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"alibaba/qwen3-embedding-0.6b":{"id":"alibaba/qwen3-embedding-0.6b","name":"Qwen3 Embedding 0.6B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-11-14","last_updated":"2025-11-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768}},"alibaba/qwen3.6-plus":{"id":"alibaba/qwen3.6-plus","name":"Qwen 3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1,"max":131072}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625,"tiers":[{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5}},"canonical_model_id":"alibaba/qwen3.6-plus"},"alibaba/qwen3.5-flash":{"id":"alibaba/qwen3.5-flash","name":"Qwen 3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1,"max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"alibaba/qwen3.5-flash"},"alibaba/qwen3.8-flash":{"id":"alibaba/qwen3.8-flash","name":"Qwen 3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":128000},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"alibaba/wan-v2.6-t2v":{"id":"alibaba/wan-v2.6-t2v","name":"Wan v2.6 Text-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"alibaba/qwen3.8-max-prime":{"id":"alibaba/qwen3.8-max-prime","name":"Qwen 3.8 Max Prime","description":"High-throughput edition of Qwen3.8 Max for coding, professional work, multimodal understanding, and long-running agent workflows","family":"qwen3.8-max","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":4,"output":12,"cache_read":0.5,"cache_write":5},"canonical_model_id":"alibaba/qwen3.8-max-prime"},"alibaba/qwen3.8-2.4t-a95b":{"id":"alibaba/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"alibaba/qwen3.7-plus":{"id":"alibaba/qwen3.7-plus","name":"Qwen 3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.24,"cache_write":1.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.24,"cache_write":1.5}},"canonical_model_id":"alibaba/qwen3.7-plus"},"alibaba/qwen3-next-80b-a3b-instruct":{"id":"alibaba/qwen3-next-80b-a3b-instruct","name":"Qwen3 Next 80B A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262114,"output":262114},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"alibaba/wan-v3.0-video-prime":{"id":"alibaba/wan-v3.0-video-prime","name":"Wan v3.0 Video Prime","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"alibaba/qwen3-next-80b-a3b-thinking":{"id":"alibaba/qwen3-next-80b-a3b-thinking","name":"Qwen3 Next 80B A3B Thinking","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"alibaba/qwen3-embedding-8b":{"id":"alibaba/qwen3-embedding-8b","type":"embedding","name":"Qwen3 Embedding 8B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"canonical_model_id":"alibaba/qwen3-embedding-8b"},"alibaba/wan-v2.7-t2v":{"id":"alibaba/wan-v2.7-t2v","name":"Wan v2.7 Text-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"alibaba/wan-v2.5-t2v-preview":{"id":"alibaba/wan-v2.5-t2v-preview","name":"Wan v2.5 Text-to-Video Preview","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-09-24","last_updated":"2025-09-24","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"alibaba/qwen3.8-27b":{"id":"alibaba/qwen3.8-27b","name":"Qwen3.8 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.5,"output":3,"cache_read":0.1,"cache_write":0.625},"canonical_model_id":"alibaba/qwen3.8-27b"},"alibaba/wan-v2.6-r2v":{"id":"alibaba/wan-v2.6-r2v","name":"Wan v2.6 Reference-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"alibaba/qwen3-max-thinking":{"id":"alibaba/qwen3-max-thinking","name":"Qwen 3 Max Thinking","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1,"max":81920}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-23","last_updated":"2025-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":1.2,"output":6,"cache_read":0.24,"tiers":[{"input":2.4,"output":12,"cache_read":0.48,"tier":{"type":"context","size":32001}},{"input":3,"output":15,"cache_read":0.6,"tier":{"type":"context","size":128001}}]}},"alibaba/wan-v2.6-r2v-flash":{"id":"alibaba/wan-v2.6-r2v-flash","name":"Wan v2.6 Reference-to-Video Flash","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"alibaba/wan-v3.0-video":{"id":"alibaba/wan-v3.0-video","name":"Wan v3.0 Video","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"release_date":"2026-08-23","last_updated":"2026-08-23","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"alibaba/qwen-3-32b":{"id":"alibaba/qwen-3-32b","name":"Qwen 3.32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-32b"},"alibaba/qwen-3-30b":{"id":"alibaba/qwen-3-30b","name":"Qwen3-30B-A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.5},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"alibaba/qwen3-coder-next":{"id":"alibaba/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.5,"output":1.2},"canonical_model_id":"alibaba/qwen3-coder-next"},"alibaba/qwen3.6-27b":{"id":"alibaba/qwen3.6-27b","name":"Qwen 3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":131072}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.6-27b"},"alibaba/qwen3.7-max":{"id":"alibaba/qwen3.7-max","name":"Qwen 3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":64000},"cost":{"input":2.5,"output":7.5,"cache_read":0.5,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.7-max"},"alibaba/qwen3-coder":{"id":"alibaba/qwen3-coder","name":"Qwen3 Coder 480B A35B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.22,"output":1.8,"cache_read":0.022,"tiers":[{"input":2.7,"output":13.5,"cache_read":0.54,"tier":{"type":"context","size":32001}},{"input":4.5,"output":22.5,"cache_read":0.9,"tier":{"type":"context","size":128001}}]},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"alibaba/wan-v2.6-i2v-flash":{"id":"alibaba/wan-v2.6-i2v-flash","name":"Wan v2.6 Image-to-Video Flash","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"alibaba/qwen-3-235b":{"id":"alibaba/qwen-3-235b","name":"Qwen3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.22,"output":0.88},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"alibaba/qwen3.8-max":{"id":"alibaba/qwen3.8-max","name":"Qwen 3.8 Max","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"temperature":true,"release_date":"2026-07-19","last_updated":"2026-07-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"alibaba/qwen3.8-max-0902":{"id":"alibaba/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"alibaba/qwen3-vl-instruct":{"id":"alibaba/qwen3-vl-instruct","name":"Qwen3 VL Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.2,"output":0.88,"cache_read":0.11}},"arcee-ai/trinity-large-thinking":{"id":"arcee-ai/trinity-large-thinking","name":"Trinity Large Thinking","description":"Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use","family":"trinity","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262100,"output":80000},"cost":{"input":0.25,"output":0.8},"canonical_model_id":"arcee-ai/trinity-large-thinking"},"mistral/mistral-small":{"id":"mistral/mistral-small","name":"Mistral Small (latest)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":4000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"mistral/mistral-small-latest"},"mistral/mistral-medium-3.5":{"id":"mistral/mistral-medium-3.5","name":"Mistral Medium Latest","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":256000},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"mistral/mistral-medium-3-5"},"mistral/mistral-nemo":{"id":"mistral/mistral-nemo","name":"Mistral Nemo","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.029,"output":0.03},"canonical_model_id":"mistral/mistral-nemo"},"mistral/ministral-3b":{"id":"mistral/ministral-3b","name":"Ministral 3B (latest)","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-10-16","last_updated":"2024-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4000},"cost":{"input":0.1,"output":0.1,"cache_read":0.01}},"mistral/ministral-8b":{"id":"mistral/ministral-8b","name":"Ministral 8B (latest)","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-10-16","last_updated":"2024-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":4000},"cost":{"input":0.15,"output":0.15,"cache_read":0.015}},"mistral/mistral-embed":{"id":"mistral/mistral-embed","type":"embedding","name":"Mistral Embed","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"mistral-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2023-12-11","last_updated":"2023-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536},"canonical_model_id":"mistral/mistral-embed"},"mistral/codestral-embed":{"id":"mistral/codestral-embed","type":"embedding","name":"Codestral Embed","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"codestral-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536},"canonical_model_id":"mistral/codestral-embed-2505"},"mistral/mistral-large-4":{"id":"mistral/mistral-large-4","name":"Mistral Large 4","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":262144},"cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"mistral/mistral-large-3":{"id":"mistral/mistral-large-3","name":"Mistral Large 3","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"mistral/mistral-large-2512"},"mistral/ministral-14b":{"id":"mistral/ministral-14b","name":"Ministral 14B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.2,"output":0.2,"cache_read":0.02},"canonical_model_id":"mistral/ministral-14b"},"mistral/codestral":{"id":"mistral/codestral","name":"Codestral (latest)","description":"Mistral code model for completions, refactors, and developer IDE workflows","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-05-29","last_updated":"2025-01-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.3,"output":0.9,"cache_read":0.03},"canonical_model_id":"mistral/codestral-latest"}}},"volcengine-coding-plan":{"id":"volcengine-coding-plan","env":["ARK_CODING_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://ark.cn-beijing.volces.com/api/coding/v3","name":"Volcengine Ark Coding Plan","doc":"https://www.volcengine.com/docs/82379/1928261","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-pro"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k3"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.3"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"minimax/MiniMax-M3"},"doubao-seed-2.0-lite":{"id":"doubao-seed-2.0-lite","name":"Seed 2.0 Lite","description":"Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"doubao-seed-evolving":{"id":"doubao-seed-evolving","name":"Seed Evolving","description":"Rolling ByteDance Seed model for rapidly updated reasoning, coding, and agent capabilities","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"bytedance-seed/seed-evolving"},"doubao-seed-2.1-turbo":{"id":"doubao-seed-2.1-turbo","name":"Seed 2.1 Turbo","description":"Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"}}},"minimax":{"id":"minimax","env":["MINIMAX_API_KEY"],"npm":"@ai-sdk/anthropic","api":"https://api.minimax.io/anthropic/v1","name":"MiniMax (minimax.io)","doc":"https://platform.minimax.io/docs/guides/quickstart","models":{"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":524288},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"tiers":[{"input":0.6,"output":2.4,"cache_read":0.12,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":0.6,"output":2.4,"cache_read":0.12}}},"MiniMax-M2.7-highspeed":{"id":"MiniMax-M2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375}},"MiniMax-M3.1-Flash":{"id":"MiniMax-M3.1-Flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"MiniMax-M2":{"id":"MiniMax-M2","name":"MiniMax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3,"output":1.2}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375}},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375}},"MiniMax-M2.1":{"id":"MiniMax-M2.1","name":"MiniMax-M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375}},"MiniMax-M2.5-highspeed":{"id":"MiniMax-M2.5-highspeed","name":"MiniMax-M2.5-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.6,"output":2.4,"cache_read":0.03,"cache_write":0.375}}}},"iflowcn":{"id":"iflowcn","env":["IFLOW_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://apis.iflow.cn/v1","name":"iFlow","doc":"https://platform.iflow.cn/en/docs","models":{"qwen3-235b-a22b-thinking-2507":{"id":"qwen3-235b-a22b-thinking-2507","name":"Qwen3-235B-A22B-Thinking","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-01","last_updated":"2025-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0,"output":0}},"qwen3-235b":{"id":"qwen3-235b","name":"Qwen3-235B-A22B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-01","last_updated":"2024-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0,"output":0}},"qwen3-coder-plus":{"id":"qwen3-coder-plus","name":"Qwen3-Coder-Plus","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-01","last_updated":"2025-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0,"output":0}},"qwen3-235b-a22b-instruct":{"id":"qwen3-235b-a22b-instruct","name":"Qwen3-235B-A22B-Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-01","last_updated":"2025-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0,"output":0}},"kimi-k2-0905":{"id":"kimi-k2-0905","name":"Kimi-K2-0905","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":0,"output":0}},"deepseek-v3":{"id":"deepseek-v3","name":"DeepSeek-V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0,"output":0}},"kimi-k2":{"id":"kimi-k2","name":"Kimi-K2","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-01","last_updated":"2024-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"cost":{"input":0,"output":0}},"deepseek-r1":{"id":"deepseek-r1","name":"DeepSeek-R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-20","last_updated":"2025-01-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0,"output":0}},"qwen3-32b":{"id":"qwen3-32b","name":"Qwen3-32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-01","last_updated":"2024-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0,"output":0}},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek-V3.2-Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":64000},"cost":{"input":0,"output":0}},"qwen3-max-preview":{"id":"qwen3-max-preview","name":"Qwen3-Max-Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0,"output":0}},"glm-4.6":{"id":"glm-4.6","name":"GLM-4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-01","last_updated":"2025-11-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":0,"output":0}},"qwen3-vl-plus":{"id":"qwen3-vl-plus","name":"Qwen3-VL-Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0,"output":0}},"qwen3-max":{"id":"qwen3-max","name":"Qwen3-Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0,"output":0}}}},"mixlayer":{"id":"mixlayer","env":["MIXLAYER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://models.mixlayer.ai/v1","name":"Mixlayer","doc":"https://docs.mixlayer.com","models":{"qwen/qwen3.5-9b":{"id":"qwen/qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.4}},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B A10B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.4,"output":3.2}},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":2.4}},"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.25,"output":1.3}},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B A17B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3.6}}}},"volcengine":{"id":"volcengine","env":["ARK_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://ark.cn-beijing.volces.com/api/v3","name":"Volcengine Ark","doc":"https://www.volcengine.com/docs/82379/1330310","models":{"doubao-seed-2-0-code-preview-260215":{"id":"doubao-seed-2-0-code-preview-260215","name":"Seed 2.0 Code","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.47499,"output":2.37494,"cache_read":0.095,"tiers":[{"input":0.71248,"output":3.56241,"cache_read":0.1425,"tier":{"type":"context","size":32000}},{"input":1.42496,"output":7.12482,"cache_read":0.28499,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"doubao-seed-2-0-lite-260428":{"id":"doubao-seed-2-0-lite-260428","name":"Seed 2.0 Lite","description":"Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":131072},"cost":{"input":0.08906,"output":0.53436,"cache_read":0.01781,"tiers":[{"input":0.13359,"output":0.80154,"cache_read":0.02672,"tier":{"type":"context","size":32000}},{"input":0.26718,"output":1.60308,"cache_read":0.05344,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"glm-5-2-260617":{"id":"glm-5-2-260617","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.18747,"output":4.15615,"cache_read":0.29687},"canonical_model_id":"zhipuai/glm-5.2"},"doubao-seed-2-0-mini-260428":{"id":"doubao-seed-2-0-mini-260428","name":"Seed 2.0 Mini","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":131072},"cost":{"input":0.02969,"output":0.29687,"cache_read":0.00594,"tiers":[{"input":0.05937,"output":0.59374,"cache_read":0.01187,"tier":{"type":"context","size":32000}},{"input":0.11875,"output":1.18747,"cache_read":0.02375,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"doubao-seed-1-6-flash-250828":{"id":"doubao-seed-1-6-flash-250828","name":"Seed 1.6 Flash","description":"Low-latency ByteDance Seed model for high-throughput chat, extraction, and lightweight tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.02227,"output":0.22265,"cache_read":0.00445,"tiers":[{"input":0.04453,"output":0.4453,"cache_read":0.00445,"tier":{"type":"context","size":32000}},{"input":0.08906,"output":0.8906,"cache_read":0.00445,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-6-flash"},"deepseek-v4-flash-ga-260731":{"id":"deepseek-v4-flash-ga-260731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.4453,"output":1.3359,"cache_read":0.01484},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"doubao-seed-2-0-pro-260215":{"id":"doubao-seed-2-0-pro-260215","name":"Seed 2.0 Pro","description":"Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.47499,"output":2.37494,"cache_read":0.095,"tiers":[{"input":0.71248,"output":3.56241,"cache_read":0.1425,"tier":{"type":"context","size":32000}},{"input":1.42496,"output":7.12482,"cache_read":0.28499,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-pro"},"glm-5-3-flash-260828":{"id":"glm-5-3-flash-260828","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.11875,"output":0.41563,"cache_read":0.03414},"canonical_model_id":"zhipuai/glm-5.3-flash"},"doubao-seed-1-8-251228":{"id":"doubao-seed-1-8-251228","name":"Seed 1.8","description":"ByteDance Seed model for multimodal reasoning, long-context analysis, and agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-12-28","last_updated":"2025-12-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":0.11875,"output":1.18747,"cache_read":0.02375,"tiers":[{"input":0.17812,"output":2.37494,"cache_read":0.02375,"tier":{"type":"context","size":32000}},{"input":0.35624,"output":3.56241,"cache_read":0.02375,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-8"},"doubao-seed-1-6-251015":{"id":"doubao-seed-1-6-251015","name":"Seed 1.6","description":"ByteDance Seed model for long-context reasoning, instruction following, and tool-assisted tasks","family":"seed","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":0.11875,"output":1.18747,"cache_read":0.02375,"tiers":[{"input":0.17812,"output":2.37494,"cache_read":0.02375,"tier":{"type":"context","size":32000}},{"input":0.35624,"output":3.56241,"cache_read":0.02375,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-6"},"deepseek-v4-pro-ga-260813":{"id":"deepseek-v4-pro-ga-260813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.3359,"output":4.00771,"cache_read":0.04453},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"doubao-seed-1-6-vision-250815":{"id":"doubao-seed-1-6-vision-250815","name":"Seed 1.6 Vision","description":"ByteDance Seed multimodal model for image understanding, visual reasoning, and tool-assisted tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-08-15","last_updated":"2025-08-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.11875,"output":1.18747,"cache_read":0.02375,"tiers":[{"input":0.17812,"output":2.37494,"cache_read":0.02375,"tier":{"type":"context","size":32000}},{"input":0.35624,"output":3.56241,"cache_read":0.02375,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-6-vision"},"doubao-seed-2-1-turbo-260628":{"id":"doubao-seed-2-1-turbo-260628","name":"Seed 2.1 Turbo","description":"Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.4453,"output":2.22651,"cache_read":0.08906},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"doubao-seed-2-1-pro-260628":{"id":"doubao-seed-2-1-pro-260628","name":"Seed 2.1 Pro","description":"Flagship ByteDance Seed 2.1 model for complex multimodal reasoning, coding, and agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.8906,"output":4.45301,"cache_read":0.17812},"canonical_model_id":"bytedance-seed/seed-2.1-pro"},"doubao-seed-character-260628":{"id":"doubao-seed-character-260628","name":"Seed Character","description":"ByteDance Seed model optimized for character-driven dialogue and consistent conversational behavior","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.11875,"output":0.29687,"cache_read":0.02375,"tiers":[{"input":0.17812,"output":0.8906,"cache_read":0.02375,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"bytedance-seed/seed-character"},"doubao-seed-evolving":{"id":"doubao-seed-evolving","name":"Seed Evolving","description":"Rolling ByteDance Seed model for rapidly updated reasoning, coding, and agent capabilities","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.8906,"output":4.45301,"cache_read":0.17812},"canonical_model_id":"bytedance-seed/seed-evolving"}}},"iteracompute":{"id":"iteracompute","env":["ITERACOMPUTE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.iteracompute.com/v1","name":"IteraCompute","doc":"https://iteracompute.com/docs.html","models":{"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731 (legacy alias)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.29,"output":1.16,"cache_read":0.0058},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.89,"output":2.69,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.29,"output":1.16,"cache_read":0.0058},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.099,"output":0.35,"cache_read":0.025},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.99,"output":3.45,"cache_read":0.245},"canonical_model_id":"zhipuai/glm-5.3"},"qwen/qwen3.8-2.4t-a95b":{"id":"qwen/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.95,"output":5.7,"cache_read":0.245},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"alibaba/qwen3.8-27b"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":999999},"cost":{"input":2.39,"output":11.99,"cache_read":0.24},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.69,"output":3.29,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":524288},"cost":{"input":0.299,"output":1.19,"cache_read":0.045},"canonical_model_id":"minimax/MiniMax-M3"}}},"io-net":{"id":"io-net","env":["IOINTELLIGENCE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.intelligence.io.solutions/api/v1","name":"IO.NET","doc":"https://io.net/docs/guides/intelligence/io-intelligence","models":{"deepseek-ai/DeepSeek-R1-0528":{"id":"deepseek-ai/DeepSeek-R1-0528","name":"DeepSeek R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":2,"output":8.75,"cache_read":1,"cache_write":4}},"meta-llama/Llama-3.3-70B-Instruct":{"id":"meta-llama/Llama-3.3-70B-Instruct","name":"Llama 3.3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.13,"output":0.38,"cache_read":0.065,"cache_write":0.26}},"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8":{"id":"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8","name":"Llama 4 Maverick 17B 128E Instruct","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-15","last_updated":"2025-01-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":430000,"output":4096},"cost":{"input":0.15,"output":0.6,"cache_read":0.075,"cache_write":0.3}},"meta-llama/Llama-3.2-90B-Vision-Instruct":{"id":"meta-llama/Llama-3.2-90B-Vision-Instruct","name":"Llama 3.2 90B Vision Instruct","description":"Open Llama multimodal model for image understanding and text reasoning","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4096},"cost":{"input":0.35,"output":0.4,"cache_read":0.175,"cache_write":0.7}},"moonshotai/Kimi-K2-Instruct-0905":{"id":"moonshotai/Kimi-K2-Instruct-0905","name":"Kimi K2 Instruct","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2024-09-05","last_updated":"2024-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":4096},"cost":{"input":0.39,"output":1.9,"cache_read":0.195,"cache_write":0.78}},"moonshotai/Kimi-K2-Thinking":{"id":"moonshotai/Kimi-K2-Thinking","name":"Kimi K2 Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2024-11-01","last_updated":"2024-11-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":4096},"cost":{"input":0.55,"output":2.25,"cache_read":0.275,"cache_write":1.1}},"mistralai/Devstral-Small-2505":{"id":"mistralai/Devstral-Small-2505","name":"Devstral Small 2505","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-05-01","last_updated":"2025-05-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":0.05,"output":0.22,"cache_read":0.025,"cache_write":0.1}},"mistralai/Magistral-Small-2506":{"id":"mistralai/Magistral-Small-2506","name":"Magistral Small 2506","description":"Mistral reasoning model for transparent analysis, math, and complex decisions","family":"magistral-small","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-01","last_updated":"2025-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":0.5,"output":1.5,"cache_read":0.25,"cache_write":1}},"mistralai/Mistral-Large-Instruct-2411":{"id":"mistralai/Mistral-Large-Instruct-2411","name":"Mistral Large Instruct 2411","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-11-01","last_updated":"2024-11-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":2,"output":6,"cache_read":1,"cache_write":4}},"mistralai/Mistral-Nemo-Instruct-2407":{"id":"mistralai/Mistral-Nemo-Instruct-2407","name":"Mistral Nemo Instruct 2407","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-05","release_date":"2024-07-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.02,"output":0.04,"cache_read":0.01,"cache_write":0.04}},"Qwen/Qwen3-Next-80B-A3B-Instruct":{"id":"Qwen/Qwen3-Next-80B-A3B-Instruct","name":"Qwen 3 Next 80B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-10","last_updated":"2025-01-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":4096},"cost":{"input":0.1,"output":0.8,"cache_read":0.05,"cache_write":0.2}},"Qwen/Qwen2.5-VL-32B-Instruct":{"id":"Qwen/Qwen2.5-VL-32B-Instruct","name":"Qwen 2.5 VL 32B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2024-11-01","last_updated":"2024-11-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.05,"output":0.22,"cache_read":0.025,"cache_write":0.1}},"Qwen/Qwen3-235B-A22B-Thinking-2507":{"id":"Qwen/Qwen3-235B-A22B-Thinking-2507","name":"Qwen 3 235B Thinking","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-07-01","last_updated":"2025-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":4096},"cost":{"input":0.11,"output":0.6,"cache_read":0.055,"cache_write":0.22}},"zai-org/GLM-4.6":{"id":"zai-org/GLM-4.6","name":"GLM 4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-11-15","last_updated":"2024-11-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":4096},"cost":{"input":0.4,"output":1.75,"cache_read":0.2,"cache_write":0.8}},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT-OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-01","last_updated":"2024-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":4096},"cost":{"input":0.03,"output":0.14,"cache_read":0.015,"cache_write":0.06}},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT-OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-01","last_updated":"2024-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.04,"output":0.4,"cache_read":0.02,"cache_write":0.08}},"Intel/Qwen3-Coder-480B-A35B-Instruct-int4-mixed-ar":{"id":"Intel/Qwen3-Coder-480B-A35B-Instruct-int4-mixed-ar","name":"Qwen 3 Coder 480B","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-15","last_updated":"2025-01-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":106000,"output":4096},"cost":{"input":0.22,"output":0.95,"cache_read":0.11,"cache_write":0.44}}}},"azure":{"id":"azure","env":["AZURE_RESOURCE_NAME","AZURE_API_KEY"],"npm":"@ai-sdk/azure","name":"Azure","doc":"https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models","models":{"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"gpt-4o-mini":{"id":"gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"gpt-3.5-turbo-instruct":{"id":"gpt-3.5-turbo-instruct","name":"GPT-3.5 Turbo Instruct","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2021-08","release_date":"2023-09-21","last_updated":"2023-09-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4096,"output":4096},"status":"deprecated","cost":{"input":1.5,"output":2}},"claude-opus-4-1":{"id":"claude-opus-4-1","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25}},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"phi-4-mini":{"id":"phi-4-mini","name":"Phi-4-mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"phi","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.075,"output":0.3}},"ministral-3b":{"id":"ministral-3b","name":"Ministral 3B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-03","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.04,"output":0.04}},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"deepseek-v3.2-speciale":{"id":"deepseek-v3.2-speciale","name":"DeepSeek-V3.2-Speciale","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.58,"output":1.68}},"gpt-image-2.5-sunburst":{"id":"gpt-image-2.5-sunburst","type":"image","name":"GPT Image 2.5 Sunburst","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-sunburst"},"gpt-5.4-pro":{"id":"gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek-V4-Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":1.74,"output":3.48},"canonical_model_id":"deepseek/deepseek-v4-pro"},"gpt-image-2.5-flare":{"id":"gpt-image-2.5-flare","type":"image","name":"GPT Image 2.5 Flare","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-flare"},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":28,"cache_read":0.35},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"gpt-5-pro":{"id":"gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"beta","provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"gpt-3.5-turbo-0125":{"id":"gpt-3.5-turbo-0125","name":"GPT-3.5 Turbo 0125","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2021-08","release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":4096},"status":"deprecated","cost":{"input":0.5,"output":1.5}},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.71,"output":0.71}},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.95,"output":4},"canonical_model_id":"moonshotai/kimi-k2.6"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"text-embedding-3-large":{"id":"text-embedding-3-large","name":"text-embedding-3-large","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":3072},"cost":{"input":0.13,"output":0}},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":4,"output":24,"cache_read":0.4,"cache_write":5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"o3-mini":{"id":"o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-07-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}}},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"gpt-image-1":{"id":"gpt-image-1","type":"image","name":"GPT-Image-1","description":"OpenAI image model for production generation, edits, and brand-safe visual workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-24","last_updated":"2025-04-24","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":40,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-1"},"gpt-5-codex":{"id":"gpt-5-codex","name":"GPT-5-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"gpt-image-1.5":{"id":"gpt-image-1.5","type":"image","name":"GPT-Image-1.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":32,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-1.5"},"phi-4-multimodal":{"id":"phi-4-multimodal","name":"Phi-4-multimodal","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"phi","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.08,"output":0.32,"input_audio":4}},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"gpt-chat-latest":{"id":"gpt-chat-latest","name":"GPT Chat Latest","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-05-05","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"beta","cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5-instant"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"llama-4-maverick-17b-128e-instruct-fp8":{"id":"llama-4-maverick-17b-128e-instruct-fp8","name":"Llama 4 Maverick 17B 128E Instruct FP8","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":16384},"cost":{"input":0.25,"output":1}},"deepseek-r1":{"id":"deepseek-r1","name":"DeepSeek-R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-01-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"status":"deprecated","cost":{"input":1.35,"output":5.4}},"phi-4-reasoning":{"id":"phi-4-reasoning","name":"Phi-4-reasoning","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"phi","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.125,"output":0.5}},"model-router":{"id":"model-router","name":"Model Router","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"model-router","attachment":true,"reasoning":false,"tool_call":true,"release_date":"2025-05-19","last_updated":"2025-11-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":16384},"provider":{"npm":"@ai-sdk/azure","shape":"completions"},"cost":{"input":0.14,"output":0}},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"gpt-image-2":{"id":"gpt-image-2","type":"image","name":"GPT-Image-2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-31","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"anthropic/claude-opus-4-8"},"gpt-3.5-turbo-1106":{"id":"gpt-3.5-turbo-1106","name":"GPT-3.5 Turbo 1106","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2021-08","release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":4096},"status":"deprecated","cost":{"input":1,"output":2}},"text-embedding-ada-002":{"id":"text-embedding-ada-002","name":"text-embedding-ada-002","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2022-12-15","last_updated":"2022-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536},"cost":{"input":0.1,"output":0}},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-01-14","last_updated":"2026-01-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"o3":{"id":"o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"llama-4-scout-17b-16e-instruct":{"id":"llama-4-scout-17b-16e-instruct","name":"Llama 4 Scout 17B 16E Instruct","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.2,"output":0.78}},"o1":{"id":"o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":15,"output":60,"cache_read":7.5}},"gpt-5.1-codex-mini":{"id":"gpt-5.1-codex-mini","name":"GPT-5.1 Codex Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-14","last_updated":"2025-11-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"codestral-2501":{"id":"codestral-2501","name":"Codestral 25.01","description":"Mistral coding model for code completion, generation, and developer workflows","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-03","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.3,"output":0.9}},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-14","last_updated":"2025-11-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"status":"beta","canonical_model_id":"xai/grok-4.7"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek-V4-Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.19,"output":0.51},"canonical_model_id":"deepseek/deepseek-v4-flash"},"gpt-4-turbo":{"id":"gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"claude-mythos-5":{"id":"claude-mythos-5","name":"Claude Mythos 5","description":"Restricted Claude model for advanced cybersecurity and biology research workflows","family":"claude-mythos","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"beta","provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-mythos-5"},"mistral-medium-2505":{"id":"mistral-medium-2505","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.4,"output":2}},"grok-4-20-reasoning":{"id":"grok-4-20-reasoning","name":"Grok 4.20 (Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-04-08","last_updated":"2026-04-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":8192},"status":"beta","cost":{"input":2,"output":6}},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":8,"output":40,"cache_read":0.8,"cache_write":10},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"cohere-embed-v3-english":{"id":"cohere-embed-v3-english","name":"Embed v3 English","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"cohere-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-11-07","last_updated":"2023-11-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":1024},"cost":{"input":0.1,"output":0}},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.58,"output":1.68}},"grok-4-20-non-reasoning":{"id":"grok-4-20-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-04-08","last_updated":"2026-04-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":8192},"status":"beta","cost":{"input":2,"output":6}},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"status":"beta","cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"phi-4-reasoning-plus":{"id":"phi-4-reasoning-plus","name":"Phi-4-reasoning-plus","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"phi","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.125,"output":0.5}},"cohere-embed-v3-multilingual":{"id":"cohere-embed-v3-multilingual","name":"Embed v3 Multilingual","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"cohere-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-11-07","last_updated":"2023-11-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":1024},"cost":{"input":0.1,"output":0}},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-14","last_updated":"2025-11-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":2.5,"output":20,"cache_read":0.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"grok-4-1-fast-non-reasoning":{"id":"grok-4-1-fast-non-reasoning","name":"Grok 4.1 Fast (Non-Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-06-27","last_updated":"2025-06-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}},"text-embedding-3-small":{"id":"text-embedding-3-small","name":"text-embedding-3-small","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536},"cost":{"input":0.02,"output":0}},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"gpt-5.1-codex-max":{"id":"gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex-max"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":1.5,"output":9,"cache_read":0.15},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek-V4.1-Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"status":"beta","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"cohere-command-a":{"id":"cohere-command-a","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":2.5,"output":10}},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.6,"output":3},"canonical_model_id":"moonshotai/kimi-k2.5"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-11-24","last_updated":"2025-08-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005}},"cohere-embed-v-4-0":{"id":"cohere-embed-v-4-0","name":"Embed v4","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"cohere-embed","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":1536},"cost":{"input":0.12,"output":0}},"grok-4-1-fast-reasoning":{"id":"grok-4-1-fast-reasoning","name":"Grok 4.1 Fast (Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-06-27","last_updated":"2025-06-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}},"gpt-4-turbo-vision":{"id":"gpt-4-turbo-vision","name":"GPT-4 Turbo Vision","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":10,"output":30}},"phi-4-mini-reasoning":{"id":"phi-4-mini-reasoning","name":"Phi-4-mini-reasoning","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"phi","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.075,"output":0.3}},"codex-mini":{"id":"codex-mini","name":"Codex Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-04","release_date":"2025-05-16","last_updated":"2025-05-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.5,"output":6,"cache_read":0.375}},"phi-4":{"id":"phi-4","name":"Phi-4","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"phi","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.125,"output":0.5}},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"mistral-small-2503":{"id":"mistral-small-2503","name":"Mistral Small 3.1","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2025-03-01","last_updated":"2025-03-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":0.1,"output":0.3}}}},"github-copilot":{"id":"github-copilot","env":["GITHUB_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"https://api.githubcopilot.com","name":"GitHub Copilot","doc":"https://docs.github.com/en/copilot","models":{"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"mai-code-1.1-flash":{"id":"mai-code-1.1-flash","name":"MAI-Code-1.1-Flash","description":"Microsoft coding model with native vision support, optimized for fast and efficient software development","family":"mai","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":128000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02},"canonical_model_id":"microsoft/mai-code-1.1-flash"},"claude-opus-4.8":{"id":"claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":168000,"output":64000},"experimental":{"modes":{"fast":{"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"mai-code-1-flash-picker":{"id":"mai-code-1-flash-picker","name":"MAI-Code-1-Flash","description":"Microsoft coding model built for fast, efficient assistance in everyday developer workflows","family":"mai","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2026-06-02","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":128000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"microsoft/mai-code-1-flash"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"claude-opus-4.7":{"id":"claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":168000,"output":32000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"claude-opus-5.5":{"id":"claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":264000,"input":128000,"output":64000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":936000,"output":64000},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":936000,"output":64000},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.8-flash"},"claude-sonnet-5.5":{"id":"claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]},{"type":"budget_tokens","min":256,"max":24000}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":128000,"output":64000},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"google/gemini-4-argon"},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"input":372000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"claude-haiku-5.5":{"id":"claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"input":372000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.5"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":224000,"output":32000},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":936000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"input":372000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"claude-haiku-4.5":{"id":"claude-haiku-4.5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":32000}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":136000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]},{"type":"budget_tokens","min":256,"max":32000}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":936000,"output":64000},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.6-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"claude-sonnet-4.6":{"id":"claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024,"max":32000}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":168000,"output":32000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"claude-fable-5.1":{"id":"claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"}}},"pareto":{"id":"pareto","env":["PARETO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.paretoinference.com/v1","name":"Pareto Inference","doc":"https://docs.paretoinference.com/","models":{"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.03,"output":0.1,"cache_read":0.006},"canonical_model_id":"zhipuai/glm-5.3-flash"}}},"echo":{"id":"echo","env":["ECHO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://echo.tracerml.ai/v1","name":"Echo","doc":"https://echo.tracerml.ai/docs/api","models":{"echo":{"id":"echo","name":"Echo","description":"Adaptive model for coding, reasoning, and tool-driven agent workflows through one OpenAI-compatible endpoint","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-19","last_updated":"2026-08-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"status":"beta","cost":{"input":10,"output":50}}}},"lynkr":{"id":"lynkr","env":["LYNKR_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"http://127.0.0.1:8081/v1","name":"Lynkr","doc":"https://github.com/Fast-Editor/Lynkr","models":{"lynkr-auto":{"id":"lynkr-auto","name":"Lynkr Auto (complexity routing)","description":"Virtual model: Lynkr scores each request on complexity and routes it to the tier model the user configured (local Ollama/llama.cpp for simple requests, configured cloud providers for complex ones).","family":"auto","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-12-03","last_updated":"2026-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0}}}},"empiriolabs":{"id":"empiriolabs","env":["EMPIRIOLABS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.empiriolabs.ai/v1","name":"EmpirioLabs AI","doc":"https://docs.empiriolabs.ai","models":{"mimo-v2-6-pro":{"id":"mimo-v2-6-pro","name":"MiMo V2.6 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.435},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"kimi-k2-6":{"id":"kimi-k2-6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16000},"cost":{"input":0.8939,"output":3.7131,"cache_read":0.1788},"canonical_model_id":"moonshotai/kimi-k2.6"},"qwen3-7-flash":{"id":"qwen3-7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":131072}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":65536},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"tiers":[{"input":0.1,"output":0.4,"cache_read":0.02,"tier":{"type":"context","size":32000}},{"input":0.2,"output":0.8,"cache_read":0.04,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"muse-spark-1-2":{"id":"muse-spark-1-2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":1},"canonical_model_id":"meta/muse-spark-1.2"},"fugu-max":{"id":"fugu-max","name":"Fugu Max","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"sakana/fugu-max"},"gemma-3-27b":{"id":"gemma-3-27b","name":"Gemma 3 27B","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-3-27b-it"},"step-3-5-flash":{"id":"step-3-5-flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":131072},"cost":{"input":0.1,"output":0.3,"cache_read":0.02},"canonical_model_id":"stepfun/step-3.5-flash"},"step-3-7-flash":{"id":"step-3-7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":131072},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"qwen3-5-27b":{"id":"qwen3-5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":80000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.086,"output":0.688,"cache_read":0.086,"tiers":[{"input":0.258,"output":2.064,"cache_read":0.258,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-27b"},"gemma-4-26b-a4b":{"id":"gemma-4-26b-a4b","name":"Gemma 4 26B-A4B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.05,"output":0.29,"cache_read":0.025},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":393216}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":1.65,"output":3.3,"cache_read":1.65},"canonical_model_id":"deepseek/deepseek-v4-pro"},"seed-2-0-lite":{"id":"seed-2-0-lite","name":"Seed 2.0 Lite","description":"Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.31,"output":2.5,"cache_read":0.31,"tiers":[{"input":0.62,"output":5,"cache_read":0.62,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"qwen3-5-9b":{"id":"qwen3-5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1024,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.09,"output":0.13,"cache_read":0.045},"canonical_model_id":"alibaba/qwen3.5-9b"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":393216}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.424,"output":1.272,"cache_read":0.424},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"mimo-v2-6-pro-ultraspeed":{"id":"mimo-v2-6-pro-ultraspeed","name":"MiMo V2.6 Pro UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":4.35,"output":8.7,"cache_read":4.35},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"mimo-v2-6-flash":{"id":"mimo-v2-6-flash","name":"MiMo V2.6 Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.14},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"glm-4-7-flash":{"id":"glm-4-7-flash","name":"GLM 4.7 Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"kimi-k2-7-code-highspeed":{"id":"kimi-k2-7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":131072},"cost":{"input":1.9,"output":8,"cache_read":1.9},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"qwen3-6-27b":{"id":"qwen3-6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":80000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.412564,"output":2.475384,"cache_read":0.412564},"canonical_model_id":"alibaba/qwen3.6-27b"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3,"output":15,"cache_read":3},"canonical_model_id":"moonshotai/kimi-k3"},"fugu-ultra-v1-1":{"id":"fugu-ultra-v1-1","name":"Fugu Ultra v1.1","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra"},"deepseek-v4-1-flash":{"id":"deepseek-v4-1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":393216}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.3},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen3-8-max":{"id":"qwen3-8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":2},"canonical_model_id":"alibaba/qwen3.8-max"},"muse-glimmer-30b":{"id":"muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.2,"output":0.8,"cache_read":0.05},"canonical_model_id":"meta/muse-glimmer-30b"},"minimax-m2-7":{"id":"minimax-m2-7","name":"MiniMax M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.7"},"qwen3-8-max-0902":{"id":"qwen3-8-max-0902","name":"Qwen3.8 Max 0902","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":2},"canonical_model_id":"alibaba/qwen3.8-max"},"glm-5-1":{"id":"glm-5-1","name":"GLM 5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":38912}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202000,"output":128000},"cost":{"input":0.825,"output":3.301,"cache_read":0.165,"tiers":[{"input":1.1,"output":3.851,"cache_read":0.22,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5.1"},"seed-2-1-turbo":{"id":"seed-2-1-turbo","name":"Seed 2.1 Turbo","description":"Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"cost":{"input":0.63,"output":3.13,"cache_read":0.63},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"qwen3-5-122b-a10b":{"id":"qwen3-5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":80000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.115,"output":0.917,"cache_read":0.115,"tiers":[{"input":0.287,"output":2.294,"cache_read":0.287,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"muse-spark-1-3":{"id":"muse-spark-1-3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":1},"canonical_model_id":"meta/muse-spark-1.3"},"qwen3-6-flash":{"id":"qwen3-6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":64000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.25,"tiers":[{"input":1,"output":4,"cache_read":1,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1,"output":4,"cache_read":1}},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen3-5-plus":{"id":"qwen3-5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.36,"output":2.21,"cache_read":0.36,"tiers":[{"input":1.08,"output":6.62,"cache_read":1.08,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.08,"output":6.62,"cache_read":1.08}},"canonical_model_id":"alibaba/qwen3.5-plus"},"deepseek-v4-1-flash-1017":{"id":"deepseek-v4-1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":393216}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.3},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"mistral-small-4":{"id":"mistral-small-4","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.15,"output":0.6,"cache_read":0.15},"canonical_model_id":"mistral/mistral-small-2603"},"qwen3-5-flash":{"id":"qwen3-5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.09,"output":0.368,"cache_read":0.09},"canonical_model_id":"alibaba/qwen3.5-flash"},"qwen3-5-35b-a3b":{"id":"qwen3-5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":80000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.057,"output":0.459,"cache_read":0.057,"tiers":[{"input":0.229,"output":1.835,"cache_read":0.229,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":393216}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":1.32,"output":3.96,"cache_read":1.32},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"qwen3-8-flash":{"id":"qwen3-8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.16,"output":0.47,"cache_read":0.16},"canonical_model_id":"alibaba/qwen3.8-flash"},"kimi-k2-7-code":{"id":"kimi-k2-7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":131072},"cost":{"input":0.95,"output":4,"cache_read":0.95},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"qwen3-5-397b-a17b":{"id":"qwen3-5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":80000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.172,"output":1.032,"cache_read":0.172,"tiers":[{"input":0.43,"output":2.58,"cache_read":0.43,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"glm-4-6v-flash":{"id":"glm-4-6v-flash","name":"GLM 4.6V Flash","description":"Lightweight GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.6v-flash"},"minimax-m3-1-flash":{"id":"minimax-m3-1-flash","name":"MiniMax M3.1 Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0.1125,"output":0.45,"cache_read":0.01125},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"step-3-5-flash-2603":{"id":"step-3-5-flash-2603","name":"Step 3.5 Flash 2603","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":131072},"cost":{"input":0.1,"output":0.3,"cache_read":0.02},"canonical_model_id":"stepfun/step-3.5-flash-2603"},"qwen3-5-4b":{"id":"qwen3-5-4b","name":"Qwen3.5 4B","description":"Qwen3.5 4B is a low-cost multimodal reasoning model with 256K context, image and video input, function tools, and structured output.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1024,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-02","last_updated":"2026-03-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.04,"output":0.07,"cache_read":0.02}},"qwen3-8-27b":{"id":"qwen3-8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.17,"output":0.5,"cache_read":0.08},"canonical_model_id":"alibaba/qwen3.8-27b"},"glm-5-3":{"id":"glm-5-3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":1.4},"canonical_model_id":"zhipuai/glm-5.3"},"muse-spark-1-1":{"id":"muse-spark-1-1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":1},"canonical_model_id":"meta/muse-spark-1.1"},"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":131072},"cost":{"input":1,"output":2.7,"cache_read":0.05},"canonical_model_id":"stepfun/step-5-preview"},"glm-5-3-flash":{"id":"glm-5-3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.075,"output":0.25,"cache_read":0.075},"canonical_model_id":"zhipuai/glm-5.3-flash"},"qwen3-6-35b-a3b":{"id":"qwen3-6-35b-a3b","name":"Qwen3.6 35B A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.07,"output":0.42,"cache_read":0.035},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"seed-2-0-pro":{"id":"seed-2-0-pro","name":"Seed 2.0 Pro","description":"Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.63,"output":3.79,"cache_read":0.63,"tiers":[{"input":1.26,"output":7.58,"cache_read":1.26,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-pro"},"qwen3-8-omni-flash":{"id":"qwen3-8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.3,"output":0.94,"cache_read":0.3},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"fugu-ultra-v1-0":{"id":"fugu-ultra-v1-0","name":"Fugu Ultra v1.0","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":7.5,"output":45,"cache_read":1.5,"tiers":[{"input":15,"output":67.5,"cache_read":3,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":15,"output":67.5,"cache_read":3}},"canonical_model_id":"sakana/fugu-ultra"},"glm-5-2":{"id":"glm-5-2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":1.4},"canonical_model_id":"zhipuai/glm-5.2"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":524288},"cost":{"input":0.225,"output":0.9,"cache_read":0.045,"tiers":[{"input":0.45,"output":1.8,"cache_read":0.09,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":0.45,"output":1.8,"cache_read":0.09}},"canonical_model_id":"minimax/MiniMax-M3"},"glm-4-5-flash":{"id":"glm-4-5-flash","name":"GLM 4.5 Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":98304},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.5-flash"},"seed-2-0-code":{"id":"seed-2-0-code","name":"Seed 2.0 Code","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.4,"output":2.4,"cache_read":0.4,"tiers":[{"input":0.8,"output":4.8,"cache_read":0.8,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"seed-2-0-mini":{"id":"seed-2-0-mini","name":"Seed 2.0 Mini","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.12,"output":0.5,"cache_read":0.12,"tiers":[{"input":0.24,"output":1,"cache_read":0.24,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"fugu-ultra-v2-0":{"id":"fugu-ultra-v2-0","name":"Fugu Ultra v2.0","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra"},"qwen3-6-plus":{"id":"qwen3-6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.5,"tiers":[{"input":2,"output":6,"cache_read":2,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":2}},"canonical_model_id":"alibaba/qwen3.6-plus"},"minimax-m2-7-highspeed":{"id":"minimax-m2-7-highspeed","name":"MiniMax M2.7 Highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32768},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"qwen3-7-max":{"id":"qwen3-7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":64000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":2.5,"output":7.5,"cache_read":2.5},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen3-7-plus":{"id":"qwen3-7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":256000}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.4,"output":1.6,"cache_read":0.4,"tiers":[{"input":1.2,"output":4.8,"cache_read":1.2,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":1.2}},"canonical_model_id":"alibaba/qwen3.7-plus"}}},"zeldoc":{"id":"zeldoc","env":["ZELDOC_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.zeldoc.ai/v1","name":"Zeldoc","doc":"https://docs.zeldoc.ai","models":{"zdev":{"id":"zdev","name":"ZDev","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0}}}},"google-vertex":{"id":"google-vertex","env":["GOOGLE_VERTEX_PROJECT","GOOGLE_VERTEX_LOCATION","GOOGLE_APPLICATION_CREDENTIALS"],"npm":"@ai-sdk/google-vertex","name":"Vertex","doc":"https://cloud.google.com/vertex-ai/generative-ai/docs/models","models":{"claude-opus-5-5@default":{"id":"claude-opus-5-5@default","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"mistral-medium-3":{"id":"mistral-medium-3","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"provider":{"npm":"@ai-sdk/mistral","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/publishers/mistralai/models"},"cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/mistral-medium-2505"},"claude-sonnet-5-5@default":{"id":"claude-sonnet-5-5@default","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gemini-3.1-flash-image":{"id":"gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","video","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":60,"cache_read":0.05},"canonical_model_id":"google/gemini-3.1-flash-image"},"claude-sonnet-5@default":{"id":"claude-sonnet-5@default","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"beta","cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"claude-sonnet-4-5@20250929":{"id":"claude-sonnet-4-5@20250929","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"claude-opus-4-8@default":{"id":"claude-opus-4-8@default","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"gemini-3.1-pro-preview-customtools":{"id":"gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"beta","cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"gemini-embedding-001":{"id":"gemini-embedding-001","type":"embedding","name":"Gemini Embedding 001","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-05","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2048,"output":1},"cost":{"input":0.15,"output":0},"canonical_model_id":"google/gemini-embedding-001"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.7-flash"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"beta","cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"claude-opus-4-5@20251101":{"id":"claude-opus-4-5@20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.8-flash"},"claude-fable-5@default":{"id":"claude-fable-5@default","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"claude-opus-4-6@default":{"id":"claude-opus-4-6@default","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"claude-opus-5@default":{"id":"claude-opus-5@default","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"gemini-3-pro-image":{"id":"gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":120,"cache_read":0.2},"canonical_model_id":"google/gemini-3-pro-image"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"claude-sonnet-4-6@default":{"id":"claude-sonnet-4-6@default","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gemini-3.1-flash-lite-image":{"id":"gemini-3.1-flash-lite-image","type":"image","name":"Nano Banana 2 Lite","description":"Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","video","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":4096},"cost":{"input":0.25,"output":30,"cache_read":0.025},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"claude-sonnet-4@20250514":{"id":"claude-sonnet-4@20250514","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"gemini-nano-banana-2.1":{"id":"gemini-nano-banana-2.1","type":"image","name":"Nano Banana 2.1","description":"Image model for prompt-driven generation, multi-turn editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image","video","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":1.5,"output":30,"cache_read":0.15},"canonical_model_id":"google/gemini-nano-banana-2.1"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1,"input_audio":2},"canonical_model_id":"google/gemini-4-argon"},"gemini-2.5-flash-tts":{"id":"gemini-2.5-flash-tts","type":"speech","name":"Gemini 2.5 Flash TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gemini-flash","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-09-30","last_updated":"2025-12-10","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":32768,"input":8192,"output":16384},"cost":{"input":0.5,"output":10},"canonical_model_id":"google/gemini-2.5-flash-tts"},"gemini-2.5-pro-tts":{"id":"gemini-2.5-pro-tts","type":"speech","name":"Gemini 2.5 Pro TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gemini-pro","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-09-30","last_updated":"2025-12-10","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":32768,"input":8192,"output":16384},"cost":{"input":1,"output":20},"canonical_model_id":"google/gemini-2.5-pro-tts"},"claude-opus-4-7@default":{"id":"claude-opus-4-7@default","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"claude-opus-4@20250514":{"id":"claude-opus-4@20250514","name":"Claude Opus 4","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-20250514"},"claude-fable-5-1@default":{"id":"claude-fable-5-1@default","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"claude-haiku-4-5@20251001":{"id":"claude-haiku-4-5@20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1}},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"codestral-2":{"id":"codestral-2","name":"Codestral 2","description":"Mistral code model for low-latency completion, fill-in-the-middle, and code generation","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"provider":{"npm":"@ai-sdk/mistral","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/publishers/mistralai/models"},"cost":{"input":0.3,"output":0.9},"canonical_model_id":"mistral/codestral-2508"},"claude-haiku-5-5@default":{"id":"claude-haiku-5-5@default","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-opus-4-1@20250805":{"id":"claude-opus-4-1@20250805","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"gemini-2.5-flash-image":{"id":"gemini-2.5-flash-image","type":"image","name":"Nano Banana","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":30},"canonical_model_id":"google/gemini-2.5-flash-image"},"mistral-small-2503":{"id":"mistral-small-2503","name":"Mistral Small 3.1 24B","description":"Efficient multimodal model for instruction following, coding, reasoning, and function calling","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"provider":{"npm":"@ai-sdk/mistral","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/publishers/mistralai/models"},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/mistral-small-3-1-24b-instruct-2503"},"xai/grok-4.20-non-reasoning":{"id":"xai/grok-4.20-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"xai/grok-4.1-fast-non-reasoning":{"id":"xai/grok-4.1-fast-non-reasoning","name":"Grok 4.1 Fast","description":"Fast Grok model for responsive chat, tool-assisted work, and low-latency responses","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":30000},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast"},"xai/grok-4.3":{"id":"xai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":30000},"status":"beta","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"xai/grok-4.1-fast-reasoning":{"id":"xai/grok-4.1-fast-reasoning","name":"Grok 4.1 Fast (Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":30000},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast-reasoning"},"xai/grok-4.7":{"id":"xai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":500000},"status":"beta","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"xai/grok-4.20-reasoning":{"id":"xai/grok-4.20-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"xai/grok-4.6":{"id":"xai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":500000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"meta/llama-4-scout-17b-16e-instruct-maas":{"id":"meta/llama-4-scout-17b-16e-instruct-maas","name":"Llama 4 Scout 17B Instruct","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1310720,"output":8192},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.25,"output":0.7},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"meta/llama-3.3-70b-instruct-maas":{"id":"meta/llama-3.3-70b-instruct-maas","name":"Llama 3.3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.72,"output":0.72}},"meta/llama-4-maverick-17b-128e-instruct-maas":{"id":"meta/llama-4-maverick-17b-128e-instruct-maas","name":"Llama 4 Maverick 17B 128E Instruct","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":8192},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.35,"output":1.15}},"deepseek-ai/deepseek-v3.1-maas":{"id":"deepseek-ai/deepseek-v3.1-maas","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.6,"output":1.7,"cache_read":0.06}},"deepseek-ai/deepseek-v3.2-maas":{"id":"deepseek-ai/deepseek-v3.2-maas","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-12-17","last_updated":"2026-04-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":65536},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.56,"output":1.68,"cache_read":0.056}},"qwen/qwen3-235b-a22b-instruct-2507-maas":{"id":"qwen/qwen3-235b-a22b-instruct-2507-maas","name":"Qwen3 235B A22B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-13","last_updated":"2025-08-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.22,"output":0.88}},"moonshotai/kimi-k2-thinking-maas":{"id":"moonshotai/kimi-k2-thinking-maas","name":"Kimi K2 Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.6,"output":2.5,"cache_read":0.06}},"zai-org/glm-5.2-maas":{"id":"zai-org/glm-5.2-maas","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":64000},"status":"beta","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/glm-5-maas":{"id":"zai-org/glm-5-maas","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":1,"output":3.2,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-5"},"zai-org/glm-4.7-maas":{"id":"zai-org/glm-4.7-maas","name":"GLM-4.7","description":"GLM model for coding, tool use, and complex reasoning","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-06","last_updated":"2026-01-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.6,"output":2.2,"cache_read":0.06}},"google/gemma-4-26b-a4b-it-maas":{"id":"google/gemma-4-26b-a4b-it-maas","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"status":"alpha","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"openai/gpt-oss-20b-maas":{"id":"openai/gpt-oss-20b-maas","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.07,"output":0.25,"cache_read":0.007}},"openai/gpt-oss-120b-maas":{"id":"openai/gpt-oss-120b-maas","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.09,"output":0.36},"canonical_model_id":"openai/gpt-oss-120b"}}},"subconscious":{"id":"subconscious","env":["SUBCONSCIOUS_API_KEY"],"npm":"@ai-sdk/anthropic","api":"https://api.subconscious.dev/v1","name":"Subconscious","doc":"https://docs.subconscious.dev","models":{"subconscious/glm-5.3-marathon":{"id":"subconscious/glm-5.3-marathon","name":"GLM-5.3 Marathon","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":5000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"subconscious/deepseek-v4.1-flash-marathon":{"id":"subconscious/deepseek-v4.1-flash-marathon","name":"DeepSeek V4.1 Flash Marathon","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":5000000,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4.1-flash"}}},"inference":{"id":"inference","env":["INFERENCE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.net/v1","name":"Inference","doc":"https://inference.net/models","models":{"meta/llama-3.1-8b-instruct":{"id":"meta/llama-3.1-8b-instruct","name":"Llama 3.1 8B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4096},"cost":{"input":0.025,"output":0.025}},"meta/llama-3.2-11b-vision-instruct":{"id":"meta/llama-3.2-11b-vision-instruct","name":"Llama 3.2 11B Vision Instruct","description":"Open Llama multimodal model for image understanding and text reasoning","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4096},"cost":{"input":0.055,"output":0.055}},"meta/llama-3.2-3b-instruct":{"id":"meta/llama-3.2-3b-instruct","name":"Llama 3.2 3B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4096},"cost":{"input":0.02,"output":0.02}},"meta/llama-3.2-1b-instruct":{"id":"meta/llama-3.2-1b-instruct","name":"Llama 3.2 1B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4096},"cost":{"input":0.01,"output":0.01}},"qwen/qwen3-embedding-4b":{"id":"qwen/qwen3-embedding-4b","name":"Qwen 3 Embedding 4B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":2048},"cost":{"input":0.01,"output":0}},"qwen/qwen-2.5-7b-vision-instruct":{"id":"qwen/qwen-2.5-7b-vision-instruct","name":"Qwen 2.5 7B Vision Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":125000,"output":4096},"cost":{"input":0.2,"output":0.2}},"osmosis/osmosis-structure-0.6b":{"id":"osmosis/osmosis-structure-0.6b","name":"Osmosis Structure 0.6B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"osmosis","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4000,"output":2048},"cost":{"input":0.1,"output":0.5}},"google/gemma-3":{"id":"google/gemma-3","name":"Google Gemma 3","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":125000,"output":4096},"cost":{"input":0.15,"output":0.3}},"mistral/mistral-nemo-12b-instruct":{"id":"mistral/mistral-nemo-12b-instruct","name":"Mistral Nemo 12B Instruct","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4096},"cost":{"input":0.038,"output":0.1}}}},"minimax-cn-coding-plan":{"id":"minimax-cn-coding-plan","env":["MINIMAX_API_KEY"],"npm":"@ai-sdk/anthropic","api":"https://api.minimax.cn/anthropic/v1","name":"MiniMax Token Plan (minimax.cn)","doc":"https://platform.minimaxi.com/docs/token-plan/intro","models":{"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":524288},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.7-highspeed":{"id":"MiniMax-M2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M3.1-Flash":{"id":"MiniMax-M3.1-Flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"MiniMax-M2":{"id":"MiniMax-M2","name":"MiniMax-M2","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M3.1-Flash-Preview":{"id":"MiniMax-M3.1-Flash-Preview","name":"MiniMax-M3.1-Flash-Preview","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-27","last_updated":"2026-09-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":512000},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash-Preview"},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.1":{"id":"MiniMax-M2.1","name":"MiniMax-M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.5-highspeed":{"id":"MiniMax-M2.5-highspeed","name":"MiniMax-M2.5-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}}}},"llmgateway-providers":{"id":"llmgateway-providers","env":["LLMGATEWAY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.llmgateway.io/v1","name":"LLM Gateway","doc":"https://llmgateway.io/docs","models":{"anthropic/claude-haiku-4-5":{"id":"anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5 (Anthropic)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5 (Anthropic)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4-5-20251101":{"id":"anthropic/claude-opus-4-5-20251101","name":"Claude Opus 4.5 (Anthropic)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":31999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6 (Anthropic)","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-fable-5-1":{"id":"anthropic/claude-fable-5-1","name":"Claude Fable 5.1 (Anthropic)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-sonnet-4-5-20250929":{"id":"anthropic/claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5 (2025-09-29) (Anthropic)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":63999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"anthropic/claude-opus-4-8":{"id":"anthropic/claude-opus-4-8","name":"Claude Opus 4.8 (Anthropic)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-sonnet-4-5":{"id":"anthropic/claude-sonnet-4-5","name":"Claude Sonnet 4.5 (Anthropic)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":63999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6 (Anthropic)","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":63999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7 (Anthropic)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5 (Anthropic)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-haiku-4-5-20251001":{"id":"anthropic/claude-haiku-4-5-20251001","name":"Claude Haiku 4.5 (2025-10-01) (Anthropic)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"anthropic/claude-sonnet-5-5":{"id":"anthropic/claude-sonnet-5-5","name":"Claude Sonnet 5.5 (Anthropic)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-haiku-5-5":{"id":"anthropic/claude-haiku-5-5","name":"Claude Haiku 5.5 (Anthropic)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-5-5":{"id":"anthropic/claude-opus-5-5","name":"Claude Opus 5.5 (Anthropic)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5 (Anthropic)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"xai/grok-4-5":{"id":"xai/grok-4-5","name":"Grok 4.5 (xAI)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3}},"xai/grok-4-7":{"id":"xai/grok-4-7","name":"Grok 4.7 (xAI)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5}},"xai/grok-4-20-beta-0309-reasoning":{"id":"xai/grok-4-20-beta-0309-reasoning","name":"Grok 4.20 Beta Reasoning (0309) (xAI)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"xai/grok-4-6":{"id":"xai/grok-4-6","name":"Grok 4.6 (xAI)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"xai/grok-build-0-1":{"id":"xai/grok-build-0-1","name":"Grok Build 0.1 (xAI)","description":"Grok coding model for agentic engineering, edits, and codebase workflows","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-20","last_updated":"2026-05-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}}},"xai/grok-4":{"id":"xai/grok-4","name":"Grok 4 (xAI)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":3,"output":15,"cache_read":0.75}},"xai/grok-4-3":{"id":"xai/grok-4-3","name":"Grok 4.3 (xAI)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"xai/grok-4-20-beta-0309-non-reasoning":{"id":"xai/grok-4-20-beta-0309-non-reasoning","name":"Grok 4.20 Beta Non-Reasoning (0309) (xAI)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"moonshot/kimi-k2.6":{"id":"moonshot/kimi-k2.6","name":"Kimi K2.6 (Moonshot AI)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshot/kimi-k3":{"id":"moonshot/kimi-k3","name":"Kimi K3 (Moonshot AI)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshot/kimi-k2.7-code-highspeed":{"id":"moonshot/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed (Moonshot AI)","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshot/kimi-k2.7-code":{"id":"moonshot/kimi-k2.7-code","name":"Kimi K2.7 Code (Moonshot AI)","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshot/kimi-k2.5":{"id":"moonshot/kimi-k2.5","name":"Kimi K2.5 (Moonshot AI)","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro (DeepSeek)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":393216},"cost":{"input":0.435,"output":0.87,"cache_read":0.003625},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (DeepSeek)","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":393216},"cost":{"input":0.15,"output":0.6,"cache_read":0.003}},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (DeepSeek)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":393216},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"quartz/gemini-3.1-pro-preview":{"id":"quartz/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro (Preview) (Quartz)","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"azure-ai-foundry/grok-4-1-fast-non-reasoning":{"id":"azure-ai-foundry/grok-4-1-fast-non-reasoning","name":"Grok 4.1 Fast Non-Reasoning (Azure AI Foundry)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5}},"azure-ai-foundry/grok-4-3":{"id":"azure-ai-foundry/grok-4-3","name":"Grok 4.3 (Azure AI Foundry)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":20000,"output":8192},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"azure-ai-foundry/grok-4-1-fast-reasoning":{"id":"azure-ai-foundry/grok-4-1-fast-reasoning","name":"Grok 4.1 Fast Reasoning (Azure AI Foundry)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5}},"google-ai-studio/gemini-pro-latest":{"id":"google-ai-studio/gemini-pro-latest","name":"Gemini Pro Latest (Google AI Studio)","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-27","last_updated":"2026-02-27","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2}},"google-ai-studio/gemini-3-flash-preview":{"id":"google-ai-studio/gemini-3-flash-preview","name":"Gemini 3 Flash (Preview) (Google AI Studio)","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3-flash-preview"},"google-ai-studio/gemini-3.1-flash-lite":{"id":"google-ai-studio/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite (Google AI Studio)","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google-ai-studio/gemini-3.1-pro-preview":{"id":"google-ai-studio/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro (Preview) (Google AI Studio)","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google-ai-studio/gemini-3.8-flash":{"id":"google-ai-studio/gemini-3.8-flash","name":"Gemini 3.8 Flash (Google AI Studio)","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.8-flash"},"google-ai-studio/gemini-4-argon":{"id":"google-ai-studio/gemini-4-argon","name":"Gemini 4 Argon (Google AI Studio)","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-4-argon"},"google-ai-studio/gemini-3.5-flash-lite":{"id":"google-ai-studio/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite (Google AI Studio)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google-ai-studio/gemini-3.6-flash":{"id":"google-ai-studio/gemini-3.6-flash","name":"Gemini 3.6 Flash (Google AI Studio)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.6-flash"},"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3 (Meta)","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1 (Meta)","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-wonder-1.0":{"id":"meta/muse-wonder-1.0","name":"Muse Wonder 1.0 (Meta)","description":"Muse Wonder is a frontier multimodal reasoning model from Meta for personal and professional agents, long-horizon coding, and computer use. It combines Active Memory, which carries relevant context across conversations, with Active Focus, which keeps it working through blockers toward a goal.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2 (Meta)","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"fireworks/glm-5.3-flash":{"id":"fireworks/glm-5.3-flash","name":"GLM-5.3 Flash (Fireworks AI)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"fireworks/kimi-k3-fast":{"id":"fireworks/kimi-k3-fast","name":"Kimi K3 Fast (Fireworks AI)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1040384,"output":1040384},"cost":{"input":4.5,"output":22.5,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"fireworks/kimi-k3":{"id":"fireworks/kimi-k3","name":"Kimi K3 (Fireworks AI)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1040384,"output":1040384},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"fireworks/glm-5.3":{"id":"fireworks/glm-5.3","name":"GLM-5.3 (Fireworks AI)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"fireworks/deepseek-v4.1-flash":{"id":"fireworks/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Fireworks AI)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.22,"output":0.66,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"bytedance/seed-1-6-250615":{"id":"bytedance/seed-1-6-250615","name":"Seed 1.6 (250615) (ByteDance)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-06-25","last_updated":"2025-06-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"bytedance/seed-2-0-code-preview-260328":{"id":"bytedance/seed-2-0-code-preview-260328","name":"Seed 2.0 Code Preview (260328) (ByteDance)","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-28","last_updated":"2026-03-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.5,"output":3,"cache_read":0.1}},"bytedance/deepseek-v4-pro":{"id":"bytedance/deepseek-v4-pro","name":"DeepSeek V4 Pro (ByteDance)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro"},"bytedance/seed-1-6-flash-250715":{"id":"bytedance/seed-1-6-flash-250715","name":"Seed 1.6 Flash (250715) (ByteDance)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.07,"output":0.3,"cache_read":0.015}},"bytedance/glm-4.7":{"id":"bytedance/glm-4.7","name":"GLM-4.7 (ByteDance)","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"bytedance/seed-2-0-lite-260428":{"id":"bytedance/seed-2-0-lite-260428","name":"Seed 2.0 Lite (260428) (ByteDance)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"bytedance/seed-1-6-250915":{"id":"bytedance/seed-1-6-250915","name":"Seed 1.6 (250915) (ByteDance)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"bytedance/seed-2-0-mini-260428":{"id":"bytedance/seed-2-0-mini-260428","name":"Seed 2.0 Mini (260428) (ByteDance)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.1,"output":0.4,"cache_read":0.02}},"bytedance/deepseek-v4-flash":{"id":"bytedance/deepseek-v4-flash","name":"DeepSeek V4 Flash (ByteDance)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash"},"bytedance/seed-2-0-pro-260328":{"id":"bytedance/seed-2-0-pro-260328","name":"Seed 2.0 Pro (260328) (ByteDance)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-28","last_updated":"2026-03-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.5,"output":3,"cache_read":0.1}},"bytedance/deepseek-v3.2":{"id":"bytedance/deepseek-v3.2","name":"DeepSeek V3.2 (ByteDance)","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.28,"output":0.42,"cache_read":0.056},"canonical_model_id":"deepseek/deepseek-v3.2"},"bytedance/gpt-oss-120b":{"id":"bytedance/gpt-oss-120b","name":"GPT OSS 120B (ByteDance)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.1,"output":0.5,"cache_read":0.02},"canonical_model_id":"openai/gpt-oss-120b"},"bytedance/glm-5.2":{"id":"bytedance/glm-5.2","name":"GLM-5.2 (ByteDance)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"bytedance/seed-1-8-251228":{"id":"bytedance/seed-1-8-251228","name":"Seed 1.8 (251228) (ByteDance)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"reka-ai/deepseek-v4-pro":{"id":"reka-ai/deepseek-v4-pro","name":"DeepSeek V4 Pro (reka-ai)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":1.2,"output":3.3,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v4-pro"},"reka-ai/glm-5.3":{"id":"reka-ai/glm-5.3","name":"GLM-5.3 (reka-ai)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":1.17,"output":3.96,"cache_read":0.234},"canonical_model_id":"zhipuai/glm-5.3"},"reka-ai/reka-edge-2603":{"id":"reka-ai/reka-edge-2603","name":"Reka Edge 2603 (reka-ai)","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"reka","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-09","last_updated":"2026-10-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":16384},"cost":{"input":0.1,"output":0.1}},"vertex-openai/grok-4-7":{"id":"vertex-openai/grok-4-7","name":"Grok 4.7 (Vertex AI (OpenAI-compatible))","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":524288},"cost":{"input":2,"output":6,"cache_read":0.5}},"vertex-openai/qwen3-235b-a22b-instruct-2507":{"id":"vertex-openai/qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B A22B Instruct 2507 (Vertex AI (OpenAI-compatible))","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.22,"output":0.88},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"vertex-openai/qwen3-next-80b-a3b-instruct":{"id":"vertex-openai/qwen3-next-80b-a3b-instruct","name":"Qwen3 Next 80B A3B Instruct (Vertex AI (OpenAI-compatible))","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"vertex-openai/glm-4.7":{"id":"vertex-openai/glm-4.7","name":"GLM-4.7 (Vertex AI (OpenAI-compatible))","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":128000},"cost":{"input":0.6,"output":2.2},"canonical_model_id":"zhipuai/glm-4.7"},"vertex-openai/qwen3-next-80b-a3b-thinking":{"id":"vertex-openai/qwen3-next-80b-a3b-thinking","name":"Qwen3 Next 80B A3B Thinking (Vertex AI (OpenAI-compatible))","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"vertex-openai/grok-4-6":{"id":"vertex-openai/grok-4-6","name":"Grok 4.6 (Vertex AI (OpenAI-compatible))","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"vertex-openai/qwen3-coder-480b-a35b-instruct":{"id":"vertex-openai/qwen3-coder-480b-a35b-instruct","name":"Qwen3 Coder 480B A35B Instruct (Vertex AI (OpenAI-compatible))","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.22,"output":1.8,"cache_read":0.022},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"vertex-openai/grok-4-20-reasoning":{"id":"vertex-openai/grok-4-20-reasoning","name":"Grok 4.20 Reasoning (Vertex AI (OpenAI-compatible))","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"vertex-openai/glm-5":{"id":"vertex-openai/glm-5","name":"GLM-5 (Vertex AI (OpenAI-compatible))","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":32768},"cost":{"input":1,"output":3.2,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-5"},"vertex-openai/deepseek-v3.2":{"id":"vertex-openai/deepseek-v3.2","name":"DeepSeek V3.2 (Vertex AI (OpenAI-compatible))","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":65536},"cost":{"input":0.56,"output":1.68,"cache_read":0.056},"canonical_model_id":"deepseek/deepseek-v3.2"},"vertex-openai/grok-4-20-non-reasoning":{"id":"vertex-openai/grok-4-20-non-reasoning","name":"Grok 4.20 Non-Reasoning (Vertex AI (OpenAI-compatible))","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"vertex-openai/kimi-k2-thinking":{"id":"vertex-openai/kimi-k2-thinking","name":"Kimi K2 Thinking (Vertex AI (OpenAI-compatible))","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.6,"output":2.5,"cache_read":0.06},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"consensusprotocol/glm-5.3-flash":{"id":"consensusprotocol/glm-5.3-flash","name":"GLM-5.3 Flash (Consensus Protocol)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.25,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-5.3-flash"},"consensusprotocol/Qwen3.8-27B":{"id":"consensusprotocol/Qwen3.8-27B","name":"Qwen3.8 27B (Consensus Protocol)","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":32768},"cost":{"input":0.08,"output":0.35,"cache_read":0.05}},"consensusprotocol/gpt-oss-20b":{"id":"consensusprotocol/gpt-oss-20b","name":"GPT OSS 20B (Consensus Protocol)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.04,"output":0.19,"cache_read":0.01},"canonical_model_id":"openai/gpt-oss-20b"},"consensusprotocol/mimo-v2.6-flash":{"id":"consensusprotocol/mimo-v2.6-flash","name":"MiMo V2.6 Flash (Consensus Protocol)","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.12,"output":0.25,"cache_read":0.0025},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"consensusprotocol/deepseek-v4-flash":{"id":"consensusprotocol/deepseek-v4-flash","name":"DeepSeek V4 Flash (Consensus Protocol)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.1,"output":0.2,"cache_read":0.005},"canonical_model_id":"deepseek/deepseek-v4-flash"},"consensusprotocol/minimax-m3":{"id":"consensusprotocol/minimax-m3","name":"MiniMax M3 (Consensus Protocol)","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.28,"output":1.1,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M3"},"consensusprotocol/gemma-4-31b-it":{"id":"consensusprotocol/gemma-4-31b-it","name":"Gemma 4 31B IT (Consensus Protocol)","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.1,"output":0.25,"cache_read":0.01},"canonical_model_id":"google/gemma-4-31b-it"},"consensusprotocol/deepseek-v4.1-flash":{"id":"consensusprotocol/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Consensus Protocol)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.15,"output":0.55,"cache_read":0.005},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"baidu/deepseek-v4-pro":{"id":"baidu/deepseek-v4-pro","name":"DeepSeek V4 Pro (Baidu)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.32,"output":3.96,"cache_read":0.042},"canonical_model_id":"deepseek/deepseek-v4-pro"},"baidu/kimi-k2.6":{"id":"baidu/kimi-k2.6","name":"Kimi K2.6 (Baidu)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"baidu/glm-5.3":{"id":"baidu/glm-5.3","name":"GLM-5.3 (Baidu)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"baidu/glm-5.1":{"id":"baidu/glm-5.1","name":"GLM-5.1 (Baidu)","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"baidu/deepseek-v4-flash":{"id":"baidu/deepseek-v4-flash","name":"DeepSeek V4 Flash (Baidu)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash"},"baidu/glm-5.2":{"id":"baidu/glm-5.2","name":"GLM-5.2 (Baidu)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"baidu/deepseek-v4.1-flash":{"id":"baidu/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Baidu)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"aws-mantle/gpt-5.6-terra":{"id":"aws-mantle/gpt-5.6-terra","name":"GPT-5.6 Terra (AWS Mantle)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":921600,"output":128000},"cost":{"input":2.2,"output":13.2,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"openai/gpt-5.6-terra"},"aws-mantle/gpt-5.6-luna":{"id":"aws-mantle/gpt-5.6-luna","name":"GPT-5.6 Luna (AWS Mantle)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":921600,"output":128000},"cost":{"input":0.22,"output":1.32,"cache_read":0.022,"cache_write":0.275},"canonical_model_id":"openai/gpt-5.6-luna"},"aws-mantle/gpt-6-astra":{"id":"aws-mantle/gpt-6-astra","name":"GPT-6 Astra (AWS Mantle)","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"openai/gpt-6-astra"},"aws-mantle/gpt-5.6-sol":{"id":"aws-mantle/gpt-5.6-sol","name":"GPT-5.6 Sol (AWS Mantle)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":921600,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.44,"cache_write":5.5},"canonical_model_id":"openai/gpt-5.6-sol"},"novita/glm-4.6v":{"id":"novita/glm-4.6v","name":"GLM-4.6V (NovitaAI)","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16000},"cost":{"input":0.3,"output":0.9,"cache_read":0.055},"canonical_model_id":"zhipuai/glm-4.6v"},"novita/qwen3-235b-a22b-thinking-2507":{"id":"novita/qwen3-235b-a22b-thinking-2507","name":"Qwen3 235B A22B Thinking 2507 (NovitaAI)","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":3}},"novita/glm-5.3-flash":{"id":"novita/glm-5.3-flash","name":"GLM-5.3 Flash (NovitaAI)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"novita/qwen3-vl-30b-a3b-instruct":{"id":"novita/qwen3-vl-30b-a3b-instruct","name":"Qwen3 VL 30B A3B Instruct (NovitaAI)","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-10-05","last_updated":"2025-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.2,"output":0.7}},"novita/minimax-m2.7":{"id":"novita/minimax-m2.7","name":"MiniMax M2.7 (NovitaAI)","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"novita/qwen3-vl-235b-a22b-instruct":{"id":"novita/qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct (NovitaAI)","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":1.5},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"novita/llama-3-70b-instruct":{"id":"novita/llama-3-70b-instruct","name":"Llama 3 70B Instruct (NovitaAI)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-04-18","last_updated":"2024-04-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8000},"cost":{"input":0.51,"output":0.74}},"novita/qwen3.5-122b-a10b":{"id":"novita/qwen3.5-122b-a10b","name":"Qwen3.5 122B A10B (NovitaAI)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":3.2},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"novita/mimo-v2.5-pro":{"id":"novita/mimo-v2.5-pro","name":"MiMo V2.5 Pro (NovitaAI)","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.522,"output":1.044,"cache_read":0.0043,"tiers":[{"input":2,"output":6,"cache_read":0.4,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.4}},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"novita/llama-3.2-3b-instruct":{"id":"novita/llama-3.2-3b-instruct","name":"Llama 3.2 3B Instruct (NovitaAI)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-09-18","last_updated":"2024-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32000},"cost":{"input":0.03,"output":0.05}},"novita/qwen3.6-35b-a3b":{"id":"novita/qwen3.6-35b-a3b","name":"Qwen3.6 35B A3B (NovitaAI)","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":64000},"cost":{"input":0.248,"output":1.485},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"novita/llama-3.3-70b-instruct":{"id":"novita/llama-3.3-70b-instruct","name":"Llama 3.3 70B Instruct (NovitaAI)","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":120000},"cost":{"input":0.135,"output":0.4},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"novita/kimi-k2.6":{"id":"novita/kimi-k2.6","name":"Kimi K2.6 (NovitaAI)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.8,"output":3.4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"novita/kimi-k3":{"id":"novita/kimi-k3","name":"Kimi K3 (NovitaAI)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"novita/qwen3-235b-a22b-instruct-2507":{"id":"novita/qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B A22B Instruct 2507 (NovitaAI)","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.09,"output":0.58},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"novita/glm-4.5v":{"id":"novita/glm-4.5v","name":"GLM-4.5V (NovitaAI)","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16000},"cost":{"input":0.6,"output":1.8,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5v"},"novita/ernie-4.5-vl-424b-a47b":{"id":"novita/ernie-4.5-vl-424b-a47b","name":"ERNIE 4.5 VL 424B A47B (NovitaAI)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ernie","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-06-30","last_updated":"2025-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":123000,"output":16000},"cost":{"input":0.42,"output":1.25}},"novita/mimo-v2.5":{"id":"novita/mimo-v2.5","name":"MiMo V2.5 (NovitaAI)","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.168,"output":0.336,"cache_read":0.0034,"tiers":[{"input":0.8,"output":4,"cache_read":0.16,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.8,"output":4,"cache_read":0.16}},"canonical_model_id":"xiaomi/mimo-v2.5"},"novita/qwen3.8-flash":{"id":"novita/qwen3.8-flash","name":"Qwen3.8 Flash (NovitaAI)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-flash"},"novita/kimi-k2":{"id":"novita/kimi-k2","name":"Kimi K2 (NovitaAI)","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":0.57,"output":2.3}},"novita/qwen3-235b-a22b-fp8":{"id":"novita/qwen3-235b-a22b-fp8","name":"Qwen3 235B A22B FP8 (NovitaAI)","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":40960,"output":20000},"cost":{"input":0.2,"output":0.8}},"novita/qwen3-vl-235b-a22b-thinking":{"id":"novita/qwen3-vl-235b-a22b-thinking","name":"Qwen3 VL 235B A22B Thinking (NovitaAI)","description":"Qwen vision-language thinking model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.98,"output":3.95},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-thinking"},"novita/step-3.7-flash":{"id":"novita/step-3.7-flash","name":"Step 3.7 Flash (NovitaAI)","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"novita/qwen3.8-2.4t-a95b":{"id":"novita/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B (NovitaAI)","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"novita/hy3":{"id":"novita/hy3","name":"Hy3 (NovitaAI)","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":262144},"cost":{"input":0.14,"output":0.58,"cache_read":0.035},"canonical_model_id":"tencent/hy3"},"novita/ling-3.0-flash":{"id":"novita/ling-3.0-flash","name":"InclusionAI Ling 3.0 Flash (NovitaAI)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-02","last_updated":"2026-08-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012}},"novita/qwen3-next-80b-a3b-instruct":{"id":"novita/qwen3-next-80b-a3b-instruct","name":"Qwen3 Next 80B A3B Instruct (NovitaAI)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":1.5},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"novita/glm-5.3":{"id":"novita/glm-5.3","name":"GLM-5.3 (NovitaAI)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"novita/glm-4.7":{"id":"novita/glm-4.7","name":"GLM-4.7 (NovitaAI)","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128000},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"novita/glm-5.1":{"id":"novita/glm-5.1","name":"GLM-5.1 (NovitaAI)","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1.38,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"novita/minimax-m2.1":{"id":"novita/minimax-m2.1","name":"MiniMax M2.1 (NovitaAI)","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.1"},"novita/qwen3-coder-480b-a35b-instruct":{"id":"novita/qwen3-coder-480b-a35b-instruct","name":"Qwen3 Coder 480B A35B Instruct (NovitaAI)","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.38,"output":1.55},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"novita/gemma-4-26b-a4b-it":{"id":"novita/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT (NovitaAI)","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.13,"output":0.4},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"novita/deepseek-v4-flash":{"id":"novita/deepseek-v4-flash","name":"DeepSeek V4 Flash (NovitaAI)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":393216},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"novita/llama-4-scout-17b-instruct":{"id":"novita/llama-4-scout-17b-instruct","name":"Llama 4 Scout 17B Instruct (NovitaAI)","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.18,"output":0.59},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"novita/qwen3.8-27b":{"id":"novita/qwen3.8-27b","name":"Qwen3.8 27B (NovitaAI)","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.42,"output":3,"cache_read":0.085},"canonical_model_id":"alibaba/qwen3.8-27b"},"novita/glm-5":{"id":"novita/glm-5","name":"GLM-5 (NovitaAI)","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"novita/kimi-k2.7-code":{"id":"novita/kimi-k2.7-code","name":"Kimi K2.7 Code (NovitaAI)","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"novita/qwen3-coder-30b-a3b-instruct":{"id":"novita/qwen3-coder-30b-a3b-instruct","name":"Qwen3 Coder 30B A3B Instruct (NovitaAI)","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":160000,"output":32768},"cost":{"input":0.07,"output":0.27},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"novita/deepseek-v3.2":{"id":"novita/deepseek-v3.2","name":"DeepSeek V3.2 (NovitaAI)","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":65536},"cost":{"input":0.269,"output":0.4,"cache_read":0.1345},"canonical_model_id":"deepseek/deepseek-v3.2"},"novita/llama-4-maverick-17b-instruct":{"id":"novita/llama-4-maverick-17b-instruct","name":"Llama 4 Maverick 17B Instruct (NovitaAI)","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":8192},"cost":{"input":0.27,"output":0.85},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"novita/deepseek-v4.1-flash-1017":{"id":"novita/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (NovitaAI)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"novita/gemma-4-31b-it":{"id":"novita/gemma-4-31b-it","name":"Gemma 4 31B IT (NovitaAI)","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"novita/qwen3.7-max":{"id":"novita/qwen3.7-max","name":"Qwen3.7 Max (NovitaAI)","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.25,"output":3.75,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.7-max"},"novita/glm-4.6":{"id":"novita/glm-4.6","name":"GLM-4.6 (NovitaAI)","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.55,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"novita/glm-5.2":{"id":"novita/glm-5.2","name":"GLM-5.2 (NovitaAI)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"novita/deepseek-v4.1-flash":{"id":"novita/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (NovitaAI)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"novita/qwen3.8-max":{"id":"novita/qwen3.8-max","name":"Qwen3.8 Max (NovitaAI)","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max"},"novita/minimax-m2.5":{"id":"novita/minimax-m2.5","name":"MiniMax M2.5 (NovitaAI)","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"novita/qwen3-max":{"id":"novita/qwen3-max","name":"Qwen3 Max (NovitaAI)","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.845,"output":3.38},"canonical_model_id":"alibaba/qwen3-max"},"novita/qwen35-397b-a17b":{"id":"novita/qwen35-397b-a17b","name":"Qwen3.5 397B A17B (NovitaAI)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":64000},"cost":{"input":0.6,"output":3.6}},"novita/ling-3.1-flash":{"id":"novita/ling-3.1-flash","name":"Ling 3.1 Flash (NovitaAI)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"inclusionai/ling-3.1-flash"},"canopywave/deepseek-v4-pro":{"id":"canopywave/deepseek-v4-pro","name":"DeepSeek V4 Pro (CanopyWave)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":1.74,"output":3.48,"cache_read":0.01},"canonical_model_id":"deepseek/deepseek-v4-pro"},"canopywave/kimi-k2.6":{"id":"canopywave/kimi-k2.6","name":"Kimi K2.6 (CanopyWave)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"canopywave/kimi-k3":{"id":"canopywave/kimi-k3","name":"Kimi K3 (CanopyWave)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"canopywave/deepseek-v4-flash":{"id":"canopywave/deepseek-v4-flash","name":"DeepSeek V4 Flash (CanopyWave)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.14,"output":0.28,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4-flash"},"canopywave/glm-5.2":{"id":"canopywave/glm-5.2","name":"GLM-5.2 (CanopyWave)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32768},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"minimax/minimax-m2.1-lightning":{"id":"minimax/minimax-m2.1-lightning","name":"MiniMax M2.1 Lightning (MiniMax)","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":196608,"output":131072},"cost":{"input":0.12,"output":0.48}},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax M2.7 (MiniMax)","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax M2 (MiniMax)","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.2,"output":1,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax M2.1 (MiniMax)","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.27,"output":1.1},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-text-01":{"id":"minimax/minimax-text-01","name":"MiniMax Text 01 (MiniMax)","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-01-15","last_updated":"2025-01-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.2,"output":1.1}},"minimax/minimax-m2.5-highspeed":{"id":"minimax/minimax-m2.5-highspeed","name":"MiniMax M2.5 Highspeed (MiniMax)","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.6,"output":2.4,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5-highspeed"},"minimax/minimax-m3.1-flash":{"id":"minimax/minimax-m3.1-flash","name":"MiniMax M3.1 Flash (MiniMax)","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":524288},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax M2.7 Highspeed (MiniMax)","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.6,"output":2.4,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax M3 (MiniMax)","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax M2.5 (MiniMax)","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"runpod/kimi-k3":{"id":"runpod/kimi-k3","name":"Kimi K3 (Runpod)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"azure/gpt-5.5":{"id":"azure/gpt-5.5","name":"GPT-5.5 (Azure)","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"azure/gpt-4.1-mini":{"id":"azure/gpt-4.1-mini","name":"GPT-4.1 Mini (Azure)","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"azure/gpt-5.4-pro":{"id":"azure/gpt-5.4-pro","name":"GPT-5.4 Pro (Azure)","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.4-pro"},"azure/gpt-5.2":{"id":"azure/gpt-5.2","name":"GPT-5.2 (Azure)","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"azure/gpt-6-luna":{"id":"azure/gpt-6-luna","name":"GPT-6 Luna (Azure)","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"openai/gpt-6-luna"},"azure/gpt-5":{"id":"azure/gpt-5","name":"GPT-5 (Azure)","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"azure/gpt-5.6-terra":{"id":"azure/gpt-5.6-terra","name":"GPT-5.6 Terra (Azure)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"azure/gpt-4.1":{"id":"azure/gpt-4.1","name":"GPT-4.1 (Azure)","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"azure/gpt-5-mini":{"id":"azure/gpt-5-mini","name":"GPT-5 Mini (Azure)","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"azure/gpt-5.6-luna":{"id":"azure/gpt-5.6-luna","name":"GPT-5.6 Luna (Azure)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"azure/gpt-5.4":{"id":"azure/gpt-5.4","name":"GPT-5.4 (Azure)","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"azure/gpt-4o":{"id":"azure/gpt-4o","name":"GPT-4o (Azure)","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"azure/gpt-5.2-codex":{"id":"azure/gpt-5.2-codex","name":"GPT-5.2 Codex (Azure)","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"azure/gpt-5.1-codex-mini":{"id":"azure/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini (Azure)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"azure/gpt-5.1-codex":{"id":"azure/gpt-5.1-codex","name":"GPT-5.1 Codex (Azure)","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":1.25,"output":10},"canonical_model_id":"openai/gpt-5.1-codex"},"azure/gpt-6-astra":{"id":"azure/gpt-6-astra","name":"GPT-6 Astra (Azure)","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"openai/gpt-6-astra"},"azure/gpt-5.6-sol":{"id":"azure/gpt-5.6-sol","name":"GPT-5.6 Sol (Azure)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5},"canonical_model_id":"openai/gpt-5.6-sol"},"azure/gpt-6.1-sol":{"id":"azure/gpt-6.1-sol","name":"GPT-6.1 Sol (Azure)","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"openai/gpt-6.1-sol"},"azure/gpt-5.4-nano":{"id":"azure/gpt-5.4-nano","name":"GPT-5.4 Nano (Azure)","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"azure/gpt-oss-120b":{"id":"azure/gpt-oss-120b","name":"GPT OSS 120B (Azure)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"azure/gpt-5.1":{"id":"azure/gpt-5.1","name":"GPT-5.1 (Azure)","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"azure/gpt-6-sol":{"id":"azure/gpt-6-sol","name":"GPT-6 Sol (Azure)","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-6-sol"},"azure/gpt-5.4-mini":{"id":"azure/gpt-5.4-mini","name":"GPT-5.4 Mini (Azure)","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"azure/gpt-5-nano":{"id":"azure/gpt-5-nano","name":"GPT-5 Nano (Azure)","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"azure/gpt-5.3-codex":{"id":"azure/gpt-5.3-codex","name":"GPT-5.3 Codex (Azure)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"luminal/deepseek-v4.1-flash":{"id":"luminal/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (luminal)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.135,"output":0.54,"cache_read":0.0027},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"google-vertex/gemini-3-flash-preview":{"id":"google-vertex/gemini-3-flash-preview","name":"Gemini 3 Flash (Preview) (Google Vertex AI)","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3-flash-preview"},"google-vertex/gemini-3.1-flash-lite":{"id":"google-vertex/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite (Google Vertex AI)","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google-vertex/gemini-3.7-flash":{"id":"google-vertex/gemini-3.7-flash","name":"Gemini 3.7 Flash (Google Vertex AI)","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.7-flash"},"google-vertex/gemini-3.1-pro-preview":{"id":"google-vertex/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro (Preview) (Google Vertex AI)","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google-vertex/gemini-3.8-flash":{"id":"google-vertex/gemini-3.8-flash","name":"Gemini 3.8 Flash (Google Vertex AI)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333}},"google-vertex/gemini-3.5-flash":{"id":"google-vertex/gemini-3.5-flash","name":"Gemini 3.5 Flash (Google Vertex AI)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.5-flash"},"google-vertex/gemini-4-argon":{"id":"google-vertex/gemini-4-argon","name":"Gemini 4 Argon (Google Vertex AI)","description":"Gemini multimodal model for text, image, audio, video, and document tasks","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-30","last_updated":"2026-09-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":0.375}},"google-vertex/gemini-3.5-flash-lite":{"id":"google-vertex/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite (Google Vertex AI)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google-vertex/gemini-3.6-flash":{"id":"google-vertex/gemini-3.6-flash","name":"Gemini 3.6 Flash (Google Vertex AI)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.6-flash"},"azure-anthropic/claude-fable-5":{"id":"azure-anthropic/claude-fable-5","name":"Claude Fable 5 (Azure Anthropic)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"azure-anthropic/claude-opus-4-6":{"id":"azure-anthropic/claude-opus-4-6","name":"Claude Opus 4.6 (Azure Anthropic)","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"azure-anthropic/claude-fable-5-1":{"id":"azure-anthropic/claude-fable-5-1","name":"Claude Fable 5.1 (Azure Anthropic)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"azure-anthropic/claude-opus-4-8":{"id":"azure-anthropic/claude-opus-4-8","name":"Claude Opus 4.8 (Azure Anthropic)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"azure-anthropic/claude-opus-4-7":{"id":"azure-anthropic/claude-opus-4-7","name":"Claude Opus 4.7 (Azure Anthropic)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"azure-anthropic/claude-opus-5":{"id":"azure-anthropic/claude-opus-5","name":"Claude Opus 5 (Azure Anthropic)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"azure-anthropic/claude-sonnet-5-5":{"id":"azure-anthropic/claude-sonnet-5-5","name":"Claude Sonnet 5.5 (Azure Anthropic)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"azure-anthropic/claude-opus-5-5":{"id":"azure-anthropic/claude-opus-5-5","name":"Claude Opus 5.5 (Azure Anthropic)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"azure-anthropic/claude-sonnet-5":{"id":"azure-anthropic/claude-sonnet-5","name":"Claude Sonnet 5 (Azure Anthropic)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"scx-ai-gp/glm-5.3-flash":{"id":"scx-ai-gp/glm-5.3-flash","name":"GLM-5.3 Flash (SCX.ai)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.088,"output":0.25,"cache_read":0.025},"canonical_model_id":"zhipuai/glm-5.3-flash"},"scx-ai-gp/glm-5.3-fast":{"id":"scx-ai-gp/glm-5.3-fast","name":"GLM 5.3 Fast (SCX.ai)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":2.8,"output":8.8,"cache_read":0.56},"canonical_model_id":"zhipuai/glm-5.3"},"scx-ai-gp/kimi-k3":{"id":"scx-ai-gp/kimi-k3","name":"Kimi K3 (SCX.ai)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3.135,"output":15.675,"cache_read":0.319},"canonical_model_id":"moonshotai/kimi-k3"},"scx-ai-gp/glm-5.2-fast":{"id":"scx-ai-gp/glm-5.2-fast","name":"GLM-5.2 Turbo (SCX.ai)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2.156,"output":6.776,"cache_read":0.429},"canonical_model_id":"zhipuai/glm-5.2"},"scx-ai-gp/glm-5.3":{"id":"scx-ai-gp/glm-5.3","name":"GLM-5.3 (SCX.ai)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.924,"output":2.904,"cache_read":0.1716},"canonical_model_id":"zhipuai/glm-5.3"},"scx-ai-gp/kimi-k2.7-code":{"id":"scx-ai-gp/kimi-k2.7-code","name":"Kimi K2.7 Code (SCX.ai)","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.737,"output":3.08,"cache_read":0.143},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"scx-ai-gp/glm-5.2":{"id":"scx-ai-gp/glm-5.2","name":"GLM-5.2 (SCX.ai)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.605,"output":2.123,"cache_read":0.154},"canonical_model_id":"zhipuai/glm-5.2"},"scx-ai-gp/deepseek-v4.1-flash":{"id":"scx-ai-gp/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (SCX.ai)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.01},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"scx-ai-gp/qwen3.8-max":{"id":"scx-ai-gp/qwen3.8-max","name":"Qwen3.8 Max (SCX.ai)","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.54,"output":4.631,"cache_read":0.198,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"perplexity/sonar":{"id":"perplexity/sonar","name":"Sonar (Perplexity)","description":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","family":"sonar","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":130000,"output":4096},"cost":{"input":0.25,"output":2.5,"cache_write":0.25},"canonical_model_id":"perplexity/sonar"},"xiaomi/mimo-v2.6-flash":{"id":"xiaomi/mimo-v2.6-flash","name":"MiMo V2.6 Flash (Xiaomi)","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo V2.6 Pro (Xiaomi)","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"deepinfra/qwen3-vl-30b-a3b-instruct":{"id":"deepinfra/qwen3-vl-30b-a3b-instruct","name":"Qwen3 VL 30B A3B Instruct (DeepInfra)","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-10-05","last_updated":"2025-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.15,"output":0.6}},"deepinfra/qwen3.5-9b":{"id":"deepinfra/qwen3.5-9b","name":"Qwen3.5 9B (DeepInfra)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.1,"output":0.15},"canonical_model_id":"alibaba/qwen3.5-9b"},"deepinfra/inkling":{"id":"deepinfra/inkling","name":"Inkling (DeepInfra)","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":262144},"cost":{"input":0.95,"output":4.05,"cache_read":0.16},"canonical_model_id":"thinkingmachines/inkling"},"deepinfra/qwen3-vl-235b-a22b-instruct":{"id":"deepinfra/qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct (DeepInfra)","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":0.88,"cache_read":0.11},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"deepinfra/deepseek-v4-pro":{"id":"deepinfra/deepseek-v4-pro","name":"DeepSeek V4 Pro (DeepInfra)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":1.3,"output":2.6,"cache_read":0.1},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepinfra/granite-4.2-8b":{"id":"deepinfra/granite-4.2-8b","name":"Granite 4.2 8B (DeepInfra)","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"granite","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":117964},"cost":{"input":0.06,"output":0.25,"cache_read":0.015}},"deepinfra/muse-glimmer-30b":{"id":"deepinfra/muse-glimmer-30b","name":"Muse Glimmer 30B (DeepInfra)","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.3,"output":1.2,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"deepinfra/ling-3.0-flash-vl":{"id":"deepinfra/ling-3.0-flash-vl","name":"Ling 3.0 Flash VL (DeepInfra)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012}},"deepinfra/ling-3.0-flash":{"id":"deepinfra/ling-3.0-flash","name":"InclusionAI Ling 3.0 Flash (DeepInfra)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-02","last_updated":"2026-08-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012}},"deepinfra/mimo-v2.6-flash":{"id":"deepinfra/mimo-v2.6-flash","name":"MiMo V2.6 Flash (DeepInfra)","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"deepinfra/gemma-4-26b-a4b-it":{"id":"deepinfra/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT (DeepInfra)","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.07,"output":0.34},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"deepinfra/deepseek-v4-flash":{"id":"deepinfra/deepseek-v4-flash","name":"DeepSeek V4 Flash (DeepInfra)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.08,"output":0.18,"cache_read":0.016},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepinfra/deepseek-v3.2":{"id":"deepinfra/deepseek-v3.2","name":"DeepSeek V3.2 (DeepInfra)","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":160000,"output":65536},"cost":{"input":0.26,"output":0.38,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepinfra/mimo-v2.6-pro":{"id":"deepinfra/mimo-v2.6-pro","name":"MiMo V2.6 Pro (DeepInfra)","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"deepinfra/deepseek-v4.1-flash-1017":{"id":"deepinfra/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (DeepInfra)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.2,"output":0.6,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepinfra/gemma-4-31b-it":{"id":"deepinfra/gemma-4-31b-it","name":"Gemma 4 31B IT (DeepInfra)","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"deepinfra/nemotron-3-ultra-550b":{"id":"deepinfra/nemotron-3-ultra-550b","name":"Nemotron 3 Ultra 550B (DeepInfra)","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":2.2,"cache_read":0.1}},"deepinfra/deepseek-v4.1-flash":{"id":"deepinfra/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (DeepInfra)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.2,"output":0.6,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"runware/glm-5.3-flash":{"id":"runware/glm-5.3-flash","name":"GLM-5.3 Flash (Runware)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"runware/deepseek-v4-pro":{"id":"runware/deepseek-v4-pro","name":"DeepSeek V4 Pro (Runware)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.961,"output":1.922,"cache_read":0.079},"canonical_model_id":"deepseek/deepseek-v4-pro"},"runware/kimi-k2.6":{"id":"runware/kimi-k2.6","name":"Kimi K2.6 (Runware)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.6,"output":3.05,"cache_read":0.13},"canonical_model_id":"moonshotai/kimi-k2.6"},"runware/kimi-k3":{"id":"runware/kimi-k3","name":"Kimi K3 (Runware)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"runware/glm-5.3":{"id":"runware/glm-5.3","name":"GLM-5.3 (Runware)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5.3"},"runware/deepseek-v4-flash":{"id":"runware/deepseek-v4-flash","name":"DeepSeek V4 Flash (Runware)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.076,"output":0.153,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash"},"runware/gpt-oss-120b":{"id":"runware/gpt-oss-120b","name":"GPT OSS 120B (Runware)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.032,"output":0.14,"cache_read":0.032},"canonical_model_id":"openai/gpt-oss-120b"},"runware/gemma-4-31b-it":{"id":"runware/gemma-4-31b-it","name":"Gemma 4 31B IT (Runware)","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.102,"output":0.297,"cache_read":0.012},"canonical_model_id":"google/gemma-4-31b-it"},"runware/glm-5.2":{"id":"runware/glm-5.2","name":"GLM-5.2 (Runware)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":128000},"cost":{"input":0.8,"output":2.55,"cache_read":0.16},"canonical_model_id":"zhipuai/glm-5.2"},"runware/deepseek-v4.1-flash":{"id":"runware/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Runware)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.01},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"groq/gpt-oss-20b":{"id":"groq/gpt-oss-20b","name":"GPT OSS 20B (Groq)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32766},"cost":{"input":0.1,"output":0.5},"canonical_model_id":"openai/gpt-oss-20b"},"groq/gpt-oss-120b":{"id":"groq/gpt-oss-120b","name":"GPT OSS 120B (Groq)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32766},"cost":{"input":0.15,"output":0.75},"canonical_model_id":"openai/gpt-oss-120b"},"ranoai/deepseek-v4-flash":{"id":"ranoai/deepseek-v4-flash","name":"DeepSeek V4 Flash (RanoAI)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"gonka24/minimax-m2.7":{"id":"gonka24/minimax-m2.7","name":"MiniMax M2.7 (Gonka24)","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.08,"output":0.32,"cache_read":0.017},"canonical_model_id":"minimax/MiniMax-M2.7"},"embercloud/glm-4.5-air":{"id":"embercloud/glm-4.5-air","name":"GLM-4.5 Air (EmberCloud)","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":96000},"cost":{"input":0.13,"output":0.85,"cache_read":0.025},"canonical_model_id":"zhipuai/glm-4.5-air"},"embercloud/glm-4.7":{"id":"embercloud/glm-4.7","name":"GLM-4.7 (EmberCloud)","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131000},"cost":{"input":0.38,"output":1.98,"cache_read":0.19},"canonical_model_id":"zhipuai/glm-4.7"},"embercloud/glm-5.1":{"id":"embercloud/glm-5.1","name":"GLM-5.1 (EmberCloud)","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131000},"cost":{"input":0.931,"output":2.93,"cache_read":0.173},"canonical_model_id":"zhipuai/glm-5.1"},"embercloud/glm-4.7-flash":{"id":"embercloud/glm-4.7-flash","name":"GLM-4.7 Flash (EmberCloud)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131000},"cost":{"input":0.06,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flash"},"embercloud/glm-5":{"id":"embercloud/glm-5","name":"GLM-5 (EmberCloud)","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131000},"cost":{"input":0.72,"output":2.3,"cache_read":0.144},"canonical_model_id":"zhipuai/glm-5"},"embercloud/qwen3-coder-next":{"id":"embercloud/qwen3-coder-next","name":"Qwen3 Coder Next (EmberCloud)","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.108,"output":0.675,"cache_read":0.06},"canonical_model_id":"alibaba/qwen3-coder-next"},"embercloud/glm-5.2":{"id":"embercloud/glm-5.2","name":"GLM-5.2 (EmberCloud)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131000},"cost":{"input":1.26,"output":3.96,"cache_read":0.234},"canonical_model_id":"zhipuai/glm-5.2"},"embercloud/glm-4.5":{"id":"embercloud/glm-4.5","name":"GLM-4.5 (EmberCloud)","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":96000},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5"},"embercloud/kimi-k2.5":{"id":"embercloud/kimi-k2.5","name":"Kimi K2.5 (EmberCloud)","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.405,"output":1.98,"cache_read":0.225},"canonical_model_id":"moonshotai/kimi-k2.5"},"zai/glm-4.6v":{"id":"zai/glm-4.6v","name":"GLM-4.6V (Z AI)","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16000},"cost":{"input":0.3,"output":0.9,"cache_read":0.05},"canonical_model_id":"zhipuai/glm-4.6v"},"zai/glm-5.3-flash":{"id":"zai/glm-5.3-flash","name":"GLM-5.3 Flash (Z AI)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai/glm-4.5v":{"id":"zai/glm-4.5v","name":"GLM-4.5V (Z AI)","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16000},"cost":{"input":0.6,"output":1.8,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5v"},"zai/glm-4.6v-flashx":{"id":"zai/glm-4.6v-flashx","name":"GLM-4.6V FlashX (Z AI)","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000},"cost":{"input":0.04,"output":0.4,"cache_read":0.004}},"zai/glm-4.5-airx":{"id":"zai/glm-4.5-airx","name":"GLM-4.5 AirX (Z AI)","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":1.1,"output":4.5,"cache_read":0.22}},"zai/glm-4.5-air":{"id":"zai/glm-4.5-air","name":"GLM-4.5 Air (Z AI)","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":98304},"cost":{"input":0.2,"output":1.1,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-4.5-air"},"zai/glm-5.3":{"id":"zai/glm-5.3","name":"GLM-5.3 (Z AI)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"zai/glm-4.7":{"id":"zai/glm-4.7","name":"GLM-4.7 (Z AI)","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"zai/glm-5.1":{"id":"zai/glm-5.1","name":"GLM-5.1 (Z AI)","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai/glm-4-32b-0414-128k":{"id":"zai/glm-4-32b-0414-128k","name":"GLM-4 32B (0414-128k) (Z AI)","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.1,"output":0.1}},"zai/glm-5":{"id":"zai/glm-5","name":"GLM-5 (Z AI)","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":131100},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"zai/glm-4.6":{"id":"zai/glm-4.6","name":"GLM-4.6 (Z AI)","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"zai/glm-5.2":{"id":"zai/glm-5.2","name":"GLM-5.2 (Z AI)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai/glm-4.5":{"id":"zai/glm-4.5","name":"GLM-4.5 (Z AI)","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5"},"zai/glm-4.7-flashx":{"id":"zai/glm-4.7-flashx","name":"GLM-4.7 FlashX (Z AI)","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.07,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"zai/glm-4.5-x":{"id":"zai/glm-4.5-x","name":"GLM-4.5 X (Z AI)","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":2.2,"output":8.9,"cache_read":0.45}},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 Preview (Tencent Cloud)","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"tencent/minimax-m2.7":{"id":"tencent/minimax-m2.7","name":"MiniMax M2.7 (Tencent Cloud)","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"tencent/deepseek-v4-pro":{"id":"tencent/deepseek-v4-pro","name":"DeepSeek V4 Pro (Tencent Cloud)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.435,"output":0.87,"cache_read":0.00363},"canonical_model_id":"deepseek/deepseek-v4-pro"},"tencent/mimo-v2.5-pro":{"id":"tencent/mimo-v2.5-pro","name":"MiMo V2.5 Pro (Tencent Cloud)","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036,"tiers":[{"input":2,"output":6,"cache_read":0.4,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.4}},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"tencent/kimi-k2.6":{"id":"tencent/kimi-k2.6","name":"Kimi K2.6 (Tencent Cloud)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.858,"output":3.566,"cache_read":0.145},"canonical_model_id":"moonshotai/kimi-k2.6"},"tencent/kimi-k3":{"id":"tencent/kimi-k3","name":"Kimi K3 (Tencent Cloud)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"tencent/kimi-k2.7-code-highspeed":{"id":"tencent/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed (Tencent Cloud)","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3 (Tencent Cloud)","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.132,"output":0.528,"cache_read":0.033},"canonical_model_id":"tencent/hy3"},"tencent/hy-mt2-plus":{"id":"tencent/hy-mt2-plus","name":"Hy-MT2 Plus (Tencent Cloud)","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":4096},"cost":{"input":0.074,"output":0.295}},"tencent/glm-5.1":{"id":"tencent/glm-5.1","name":"GLM-5.1 (Tencent Cloud)","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"tencent/deepseek-v4-flash":{"id":"tencent/deepseek-v4-flash","name":"DeepSeek V4 Flash (Tencent Cloud)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"tencent/glm-5":{"id":"tencent/glm-5","name":"GLM-5 (Tencent Cloud)","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"tencent/kimi-k2.7-code":{"id":"tencent/kimi-k2.7-code","name":"Kimi K2.7 Code (Tencent Cloud)","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"tencent/glm-5v-turbo":{"id":"tencent/glm-5v-turbo","name":"GLM-5V Turbo (Tencent Cloud)","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"tencent/glm-5-turbo":{"id":"tencent/glm-5-turbo","name":"GLM-5 Turbo (Tencent Cloud)","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"tencent/minimax-m3":{"id":"tencent/minimax-m3","name":"MiniMax M3 (Tencent Cloud)","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"tencent/glm-5.2":{"id":"tencent/glm-5.2","name":"GLM-5.2 (Tencent Cloud)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"meta-contributor/muse-spark-1.3-contributor":{"id":"meta-contributor/muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor (Meta Contributor)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta-contributor/muse-spark-1.2-contributor":{"id":"meta-contributor/muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor (Meta Contributor)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"sakana/fugu-max":{"id":"sakana/fugu-max","name":"Fugu Max (Sakana AI)","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":2,"output":6,"cache_read":0.25}},"sakana/fugu-ultra-v2.0":{"id":"sakana/fugu-ultra-v2.0","name":"Fugu Ultra v2.0 (Sakana AI)","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":5,"output":30,"cache_read":0.5}},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra (Sakana AI)","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"sakana/fugu-ultra"},"aws-bedrock/claude-opus-4-1-20250805":{"id":"aws-bedrock/claude-opus-4-1-20250805","name":"Claude Opus 4.1 (AWS Bedrock)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":31999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"aws-bedrock/claude-haiku-4-5":{"id":"aws-bedrock/claude-haiku-4-5","name":"Claude Haiku 4.5 (AWS Bedrock)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"aws-bedrock/claude-fable-5":{"id":"aws-bedrock/claude-fable-5","name":"Claude Fable 5 (AWS Bedrock)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"aws-bedrock/grok-4-7":{"id":"aws-bedrock/grok-4-7","name":"Grok 4.7 (AWS Bedrock)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":524288},"cost":{"input":2,"output":6,"cache_read":0.5}},"aws-bedrock/kimi-k3":{"id":"aws-bedrock/kimi-k3","name":"Kimi K3 (AWS Bedrock)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"moonshotai/kimi-k3"},"aws-bedrock/claude-opus-4-5-20251101":{"id":"aws-bedrock/claude-opus-4-5-20251101","name":"Claude Opus 4.5 (AWS Bedrock)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":31999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"aws-bedrock/claude-opus-4-6":{"id":"aws-bedrock/claude-opus-4-6","name":"Claude Opus 4.6 (AWS Bedrock)","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"aws-bedrock/claude-fable-5-1":{"id":"aws-bedrock/claude-fable-5-1","name":"Claude Fable 5.1 (AWS Bedrock)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"aws-bedrock/claude-sonnet-4-5-20250929":{"id":"aws-bedrock/claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5 (2025-09-29) (AWS Bedrock)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":63999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"aws-bedrock/claude-opus-4-8":{"id":"aws-bedrock/claude-opus-4-8","name":"Claude Opus 4.8 (AWS Bedrock)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"aws-bedrock/grok-4-6":{"id":"aws-bedrock/grok-4-6","name":"Grok 4.6 (AWS Bedrock)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"aws-bedrock/claude-sonnet-4-5":{"id":"aws-bedrock/claude-sonnet-4-5","name":"Claude Sonnet 4.5 (AWS Bedrock)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":63999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"aws-bedrock/claude-sonnet-4-6":{"id":"aws-bedrock/claude-sonnet-4-6","name":"Claude Sonnet 4.6 (AWS Bedrock)","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":63999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"aws-bedrock/llama-4-scout-17b-instruct":{"id":"aws-bedrock/llama-4-scout-17b-instruct","name":"Llama 4 Scout 17B Instruct (AWS Bedrock)","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":2048},"cost":{"input":0.17,"output":0.66},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"aws-bedrock/claude-opus-4-7":{"id":"aws-bedrock/claude-opus-4-7","name":"Claude Opus 4.7 (AWS Bedrock)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"aws-bedrock/claude-opus-5":{"id":"aws-bedrock/claude-opus-5","name":"Claude Opus 5 (AWS Bedrock)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"aws-bedrock/llama-4-maverick-17b-instruct":{"id":"aws-bedrock/llama-4-maverick-17b-instruct","name":"Llama 4 Maverick 17B Instruct (AWS Bedrock)","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":2048},"cost":{"input":0.24,"output":0.97},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"aws-bedrock/claude-haiku-4-5-20251001":{"id":"aws-bedrock/claude-haiku-4-5-20251001","name":"Claude Haiku 4.5 (2025-10-01) (AWS Bedrock)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"aws-bedrock/claude-sonnet-5-5":{"id":"aws-bedrock/claude-sonnet-5-5","name":"Claude Sonnet 5.5 (AWS Bedrock)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"aws-bedrock/claude-haiku-5-5":{"id":"aws-bedrock/claude-haiku-5-5","name":"Claude Haiku 5.5 (AWS Bedrock)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"anthropic/claude-haiku-5-5"},"aws-bedrock/llama-3.1-70b-instruct":{"id":"aws-bedrock/llama-3.1-70b-instruct","name":"Llama 3.1 70B Instruct (AWS Bedrock)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":2048},"cost":{"input":0.72,"output":0.72}},"aws-bedrock/claude-opus-5-5":{"id":"aws-bedrock/claude-opus-5-5","name":"Claude Opus 5.5 (AWS Bedrock)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"aws-bedrock/grok-4-3":{"id":"aws-bedrock/grok-4-3","name":"Grok 4.3 (AWS Bedrock)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"aws-bedrock/claude-sonnet-5":{"id":"aws-bedrock/claude-sonnet-5","name":"Claude Sonnet 5 (AWS Bedrock)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"inference.net/glm-5.3-flash":{"id":"inference.net/glm-5.3-flash","name":"GLM-5.3 Flash (Inference.net)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"cost":{"input":0.09,"output":0.28,"cache_read":0.02},"canonical_model_id":"zhipuai/glm-5.3-flash"},"inference.net/llama-3.2-11b-instruct":{"id":"inference.net/llama-3.2-11b-instruct","name":"Llama 3.2 11B Instruct (Inference.net)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.07,"output":0.33}},"inference.net/glm-5.3":{"id":"inference.net/glm-5.3","name":"GLM-5.3 (Inference.net)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.9,"output":3,"cache_read":0.15},"canonical_model_id":"zhipuai/glm-5.3"},"inference.net/deepseek-v4.1-flash-1017":{"id":"inference.net/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (Inference.net)","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.3,"output":1,"cache_read":0.007}},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4 Mini (OpenAI)","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o Mini (OpenAI)","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4":{"id":"openai/gpt-4","name":"GPT-4 (OpenAI)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":30,"output":60},"canonical_model_id":"openai/gpt-4"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5 (OpenAI)","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-live-transcribe":{"id":"openai/gpt-live-transcribe","type":"transcription","name":"GPT Live Transcribe (OpenAI)","description":"Low-latency streaming speech-to-text model with keyword and language hints","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-07-28","last_updated":"2026-07-28","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-live-transcribe"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 Nano (OpenAI)","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 Mini (OpenAI)","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro (OpenAI)","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2 (OpenAI)","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 Pro (OpenAI)","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna (OpenAI)","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5 (OpenAI)","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra (OpenAI)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3 Mini (OpenAI)","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5 Turbo (OpenAI)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1 (OpenAI)","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-4o-transcribe":{"id":"openai/gpt-4o-transcribe","name":"GPT-4o Transcribe (OpenAI)","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-20","last_updated":"2025-03-20","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":2000},"cost":{"input":2.5,"output":10}},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini (OpenAI)","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna (OpenAI)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4 (OpenAI)","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o (OpenAI)","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/o3":{"id":"openai/o3","name":"o3 (OpenAI)","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/o1":{"id":"openai/o1","name":"o1 (OpenAI)","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2 Pro (OpenAI)","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra (OpenAI)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-03","last_updated":"2026-09-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5}},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo (OpenAI)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol (OpenAI)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol (OpenAI)","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 Nano (OpenAI)","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1 (OpenAI)","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol (OpenAI)","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-4o-mini-transcribe":{"id":"openai/gpt-4o-mini-transcribe","name":"GPT-4o Mini Transcribe (OpenAI)","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-20","last_updated":"2025-03-20","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":2000},"cost":{"input":1.25,"output":5}},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 Mini (OpenAI)","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro (OpenAI)","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano (OpenAI)","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-transcribe":{"id":"openai/gpt-transcribe","type":"transcription","name":"GPT Transcribe (OpenAI)","description":"High-accuracy speech-to-text model for audio files and committed Realtime turns","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-07-28","last_updated":"2026-07-28","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-transcribe"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex (OpenAI)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"vertex-anthropic/claude-haiku-4-5":{"id":"vertex-anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5 (Vertex AI (Anthropic))","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"vertex-anthropic/claude-opus-4-5-20251101":{"id":"vertex-anthropic/claude-opus-4-5-20251101","name":"Claude Opus 4.5 (Vertex AI (Anthropic))","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":31999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"vertex-anthropic/claude-opus-4-6":{"id":"vertex-anthropic/claude-opus-4-6","name":"Claude Opus 4.6 (Vertex AI (Anthropic))","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"vertex-anthropic/claude-sonnet-4-5":{"id":"vertex-anthropic/claude-sonnet-4-5","name":"Claude Sonnet 4.5 (Vertex AI (Anthropic))","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":63999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"vertex-anthropic/claude-sonnet-4-6":{"id":"vertex-anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6 (Vertex AI (Anthropic))","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":63999},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"vertex-anthropic/claude-opus-4-7":{"id":"vertex-anthropic/claude-opus-4-7","name":"Claude Opus 4.7 (Vertex AI (Anthropic))","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"vertex-anthropic/claude-sonnet-5":{"id":"vertex-anthropic/claude-sonnet-5","name":"Claude Sonnet 5 (Vertex AI (Anthropic))","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"scx-ai/minimax-m2.7":{"id":"scx-ai/minimax-m2.7","name":"MiniMax M2.7 (SCX.ai (Turbo))","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":196608},"cost":{"input":0.48,"output":1.79,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M2.7"},"scx-ai/qwen3-32b":{"id":"scx-ai/qwen3-32b","name":"Qwen3 32B (SCX.ai (Turbo))","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0.36,"output":0.87},"canonical_model_id":"alibaba/qwen3-32b"},"scx-ai/gpt-oss-120b":{"id":"scx-ai/gpt-oss-120b","name":"GPT OSS 120B (SCX.ai (Turbo))","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.17,"output":0.55},"canonical_model_id":"openai/gpt-oss-120b"},"cerebras/llama-3.3-70b-instruct":{"id":"cerebras/llama-3.3-70b-instruct","name":"Llama 3.3 70B Instruct (Cerebras)","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.85,"output":1.2},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"cerebras/qwen3-235b-a22b-instruct-2507":{"id":"cerebras/qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B A22B Instruct 2507 (Cerebras)","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":8192},"cost":{"input":0.6,"output":1.2},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"cerebras/glm-4.7":{"id":"cerebras/glm-4.7","name":"GLM-4.7 (Cerebras)","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":2.25,"output":2.75},"canonical_model_id":"zhipuai/glm-4.7"},"cerebras/gpt-oss-120b":{"id":"cerebras/gpt-oss-120b","name":"GPT OSS 120B (Cerebras)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.35,"output":0.75},"canonical_model_id":"openai/gpt-oss-120b"},"cerebras/gemma-4-31b-it":{"id":"cerebras/gemma-4-31b-it","name":"Gemma 4 31B IT (Cerebras)","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.99,"output":1.49},"canonical_model_id":"google/gemma-4-31b-it"},"alibaba/qwen3.7-flash":{"id":"alibaba/qwen3.7-flash","name":"Qwen3.7 Flash (Alibaba Cloud)","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"input":991808,"output":65536},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.0375},"canonical_model_id":"alibaba/qwen3.7-flash"},"alibaba/qwen-plus":{"id":"alibaba/qwen-plus","name":"Qwen Plus (Alibaba Cloud)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32000},"cost":{"input":0.4,"output":1.2,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen-plus"},"alibaba/qwen-flash":{"id":"alibaba/qwen-flash","name":"Qwen Flash (Alibaba Cloud)","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":0.05,"output":0.4,"cache_read":0.01,"cache_write":0.0625},"canonical_model_id":"alibaba/qwen-flash"},"alibaba/qwen3.5-122b-a10b":{"id":"alibaba/qwen3.5-122b-a10b","name":"Qwen3.5 122B A10B (Alibaba Cloud)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":3.2},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"alibaba/deepseek-v4-pro":{"id":"alibaba/deepseek-v4-pro","name":"DeepSeek V4 Pro (Alibaba Cloud)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":2.4,"output":4.8,"cache_read":0.2},"canonical_model_id":"deepseek/deepseek-v4-pro"},"alibaba/qwen3.6-35b-a3b":{"id":"alibaba/qwen3.6-35b-a3b","name":"Qwen3.6 35B A3B (Alibaba Cloud)","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.375,"output":2.25},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"alibaba/qwen-max":{"id":"alibaba/qwen-max","name":"Qwen Max (Alibaba Cloud)","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":1.6,"output":6.4},"canonical_model_id":"alibaba/qwen-max"},"alibaba/kimi-k3":{"id":"alibaba/kimi-k3","name":"Kimi K3 (Alibaba Cloud)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"alibaba/qwen3.6-plus":{"id":"alibaba/qwen3.6-plus","name":"Qwen3.6 Plus (Alibaba Cloud)","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625,"tiers":[{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5}},"canonical_model_id":"alibaba/qwen3.6-plus"},"alibaba/qwen3.6-flash":{"id":"alibaba/qwen3.6-flash","name":"Qwen3.6 Flash (Alibaba Cloud)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.05,"cache_write":0.3125},"canonical_model_id":"alibaba/qwen3.6-flash"},"alibaba/qwen3-coder-flash":{"id":"alibaba/qwen3-coder-flash","name":"Qwen3 Coder Flash (Alibaba Cloud)","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":997952,"output":65536},"cost":{"input":0.3,"output":1.5,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"alibaba/qwen3-coder-flash"},"alibaba/qwen3.8-flash":{"id":"alibaba/qwen3.8-flash","name":"Qwen3.8 Flash (Alibaba Cloud)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"alibaba/qwen3.7-plus":{"id":"alibaba/qwen3.7-plus","name":"Qwen3.7 Plus (Alibaba Cloud)","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":65536},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen3.7-plus"},"alibaba/glm-5.3":{"id":"alibaba/glm-5.3","name":"GLM-5.3 (Alibaba Cloud)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.28},"canonical_model_id":"zhipuai/glm-5.3"},"alibaba/qwen3.5-27b":{"id":"alibaba/qwen3.5-27b","name":"Qwen3.5 27B (Alibaba Cloud)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":2.4},"canonical_model_id":"alibaba/qwen3.5-27b"},"alibaba/qwen3.5-35b-a3b":{"id":"alibaba/qwen3.5-35b-a3b","name":"Qwen3.5 35B A3B (Alibaba Cloud)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"alibaba/deepseek-v4-flash":{"id":"alibaba/deepseek-v4-flash","name":"DeepSeek V4 Flash (Alibaba Cloud)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.2,"output":0.4,"cache_read":0.04},"canonical_model_id":"deepseek/deepseek-v4-flash"},"alibaba/glm-5":{"id":"alibaba/glm-5","name":"GLM-5 (Alibaba Cloud)","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"cost":{"input":0.573,"output":2.58},"canonical_model_id":"zhipuai/glm-5"},"alibaba/qwen-plus-latest":{"id":"alibaba/qwen-plus-latest","name":"Qwen Plus Latest (Alibaba Cloud)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2024-09-09","last_updated":"2024-09-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":0.4,"output":1.2,"cache_read":0.08,"cache_write":0.5}},"alibaba/deepseek-v4.1-flash-1017":{"id":"alibaba/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (Alibaba Cloud)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"alibaba/qwen3.7-max":{"id":"alibaba/qwen3.7-max","name":"Qwen3.7 Max (Alibaba Cloud)","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":65536},"cost":{"input":2.5,"output":7.5,"cache_read":0.5,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.7-max"},"alibaba/glm-5.2":{"id":"alibaba/glm-5.2","name":"GLM-5.2 (Alibaba Cloud)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.28},"canonical_model_id":"zhipuai/glm-5.2"},"alibaba/deepseek-v4.1-flash":{"id":"alibaba/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Alibaba Cloud)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"alibaba/kimi-k2.5":{"id":"alibaba/kimi-k2.5","name":"Kimi K2.5 (Alibaba Cloud)","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.574,"output":3.011},"canonical_model_id":"moonshotai/kimi-k2.5"},"alibaba/qwen3-vl-plus":{"id":"alibaba/qwen3-vl-plus","name":"Qwen3 VL Plus (Alibaba Cloud)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":1.6,"cache_read":0.04,"cache_write":0.25},"canonical_model_id":"alibaba/qwen3-vl-plus"},"alibaba/qwen3.8-max":{"id":"alibaba/qwen3.8-max","name":"Qwen3.8 Max (Alibaba Cloud)","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"alibaba/qwen35-397b-a17b":{"id":"alibaba/qwen35-397b-a17b","name":"Qwen3.5 397B A17B (Alibaba Cloud)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6}},"mistral/ministral-8b-2512":{"id":"mistral/ministral-8b-2512","name":"Ministral 8B (Mistral AI)","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.15,"output":0.15}},"mistral/glm-5.3":{"id":"mistral/glm-5.3","name":"GLM-5.3 (Mistral AI)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.3"},"mistral/devstral-2512":{"id":"mistral/devstral-2512","name":"Devstral 2 (Mistral AI)","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/devstral-2512"},"mistral/ministral-3b-2512":{"id":"mistral/ministral-3b-2512","name":"Ministral 3B (Mistral AI)","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":0.1,"output":0.1}},"mistral/mistral-medium-3-5":{"id":"mistral/mistral-medium-3-5","name":"Mistral Medium 3.5 (Mistral AI)","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5}},"mistral/mistral-large-2512":{"id":"mistral/mistral-large-2512","name":"Mistral Large 3 (Mistral AI)","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/mistral-large-2512"},"mistral/mistral-small-2603":{"id":"mistral/mistral-small-2603","name":"Mistral Small 4 (Mistral AI)","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"mistral/mistral-small-2603"},"mistral/ministral-14b-2512":{"id":"mistral/ministral-14b-2512","name":"Ministral 14B (Mistral AI)","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.2,"output":0.2}},"mistral/mistral-large-4":{"id":"mistral/mistral-large-4","name":"Mistral Large 4 (Mistral AI)","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":262144},"cost":{"input":1.36,"output":4.18,"cache_read":0.14},"canonical_model_id":"mistral/mistral-large-4"},"mistral/mistral-small-2506":{"id":"mistral/mistral-small-2506","name":"Mistral Small 3.2 (Mistral AI)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/mistral-small-2506"},"mistral/codestral-2508":{"id":"mistral/codestral-2508","name":"Codestral (Mistral AI)","description":"Mistral coding model for code completion, generation, and developer workflows","family":"codestral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.3,"output":0.9}},"mistral/mistral-large-latest":{"id":"mistral/mistral-large-latest","name":"Mistral Large Latest (Mistral AI)","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":262144},"cost":{"input":4,"output":12},"canonical_model_id":"mistral/mistral-large-latest"},"together-ai/glm-5.3-flash":{"id":"together-ai/glm-5.3-flash","name":"GLM-5.3 Flash (Together AI)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943717},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"together-ai/deepseek-v4-pro":{"id":"together-ai/deepseek-v4-pro","name":"DeepSeek V4 Pro (Together AI)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":163840},"cost":{"input":1.32,"output":3.96,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v4-pro"},"together-ai/kimi-k3":{"id":"together-ai/kimi-k3","name":"Kimi K3 (Together AI)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":1000000},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"together-ai/muse-glimmer-30b":{"id":"together-ai/muse-glimmer-30b","name":"Muse Glimmer 30B (Together AI)","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.35,"output":1.5,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"together-ai/qwen3.8-2.4t-a95b":{"id":"together-ai/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B (Together AI)","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1010000,"output":909000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"together-ai/glm-5.3":{"id":"together-ai/glm-5.3","name":"GLM-5.3 (Together AI)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943717},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"together-ai/glm-4.7":{"id":"together-ai/glm-4.7","name":"GLM-4.7 (Together AI)","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":128000},"cost":{"input":0.45,"output":2},"canonical_model_id":"zhipuai/glm-4.7"},"together-ai/deepseek-v4-flash":{"id":"together-ai/deepseek-v4-flash","name":"DeepSeek V4 Flash (Together AI)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.14,"output":0.28,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4-flash"},"together-ai/minimax-m3":{"id":"together-ai/minimax-m3","name":"MiniMax M3 (Together AI)","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"together-ai/gpt-oss-120b":{"id":"together-ai/gpt-oss-120b","name":"GPT OSS 120B (Together AI)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"together-ai/deepseek-v4.1-flash-1017":{"id":"together-ai/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (Together AI)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"together-ai/deepseek-v4.1-flash":{"id":"together-ai/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Together AI)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"entrim/glm-5.3-flash":{"id":"entrim/glm-5.3-flash","name":"GLM-5.3 Flash (entrim)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.35,"cache_read":0.0175},"canonical_model_id":"zhipuai/glm-5.3-flash"},"entrim/glm-5.3":{"id":"entrim/glm-5.3","name":"GLM-5.3 (entrim)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.5,"output":2,"cache_read":0.15},"canonical_model_id":"zhipuai/glm-5.3"},"entrim/deepseek-v4.1-flash":{"id":"entrim/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (entrim)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.2,"output":0.6,"cache_read":0.005},"canonical_model_id":"deepseek/deepseek-v4.1-flash"}}},"dinference":{"id":"dinference","env":["DINFERENCE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.dinference.com/v1","name":"DInference","doc":"https://dinference.com","models":{"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.45,"output":1.65},"canonical_model_id":"zhipuai/glm-4.7"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":1.25,"output":3.89},"canonical_model_id":"zhipuai/glm-5.1"},"glm-5":{"id":"glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.75,"output":2.4},"canonical_model_id":"zhipuai/glm-5"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-08","last_updated":"2025-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.0675,"output":0.27}},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.25,"output":3.89},"canonical_model_id":"zhipuai/glm-5.2"},"minimax-m2.5":{"id":"minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"cost":{"input":0.22,"output":0.88},"canonical_model_id":"minimax/MiniMax-M2.5"}}},"lowrouter":{"id":"lowrouter","env":["LOWROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.lowrouter.ai/v1","name":"LowRouter","doc":"https://lowrouter.ai/docs/models/available","models":{"auto/mistralai/mistral-medium-3.5":{"id":"auto/mistralai/mistral-medium-3.5","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.5688,"output":7.8442,"cache_read":0.1569},"canonical_model_id":"mistral/mistral-medium-2604"},"auto/mistralai/ministral-8b-2512":{"id":"auto/mistralai/ministral-8b-2512","name":"Ministral 3 8B","description":"Compact open vision-language model for edge deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.1569,"output":0.1569,"cache_read":0.0157},"canonical_model_id":"mistral/ministral-3-8b-instruct-2512"},"auto/mistralai/ministral-3b-2512":{"id":"auto/mistralai/ministral-3b-2512","name":"Ministral 3 3B","description":"Compact open vision-language model for edge deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.1046,"output":0.1046,"cache_read":0.0105},"canonical_model_id":"mistral/ministral-3-3b-instruct-2512"},"auto/mistralai/mistral-large-2512":{"id":"auto/mistralai/mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5229,"output":1.5688,"cache_read":0.0523},"canonical_model_id":"mistral/mistral-large-2512"},"auto/mistralai/mistral-small-2603":{"id":"auto/mistralai/mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.1569,"output":0.6275,"cache_read":0.0157},"canonical_model_id":"mistral/mistral-small-2603"},"auto/mistralai/ministral-14b-2512":{"id":"auto/mistralai/ministral-14b-2512","name":"Ministral 3 14B","description":"Open vision-language model for efficient local deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.2092,"output":0.2092,"cache_read":0.0209},"canonical_model_id":"mistral/ministral-3-14b-instruct-2512"},"auto/mistralai/mistral-large-4":{"id":"auto/mistralai/mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":262144},"status":"beta","cost":{"input":0.7112,"output":2.1859,"cache_read":0.0732},"canonical_model_id":"mistral/mistral-large-4"},"auto/mistralai/codestral-2508":{"id":"auto/mistralai/codestral-2508","name":"Codestral 25.08","description":"Mistral code model for low-latency completion, fill-in-the-middle, and code generation","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.3138,"output":0.9413,"cache_read":0.0314},"canonical_model_id":"mistral/codestral-2508"},"auto/openai/gpt-4o-mini":{"id":"auto/openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"auto/openai/gpt-5.5":{"id":"auto/openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"auto/openai/gpt-4.1-mini":{"id":"auto/openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"auto/openai/gpt-5.2":{"id":"auto/openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"auto/openai/gpt-6-luna":{"id":"auto/openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"auto/openai/gpt-5":{"id":"auto/openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"auto/openai/gpt-5.6-terra":{"id":"auto/openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"auto/openai/gpt-4.1":{"id":"auto/openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"auto/openai/gpt-5-mini":{"id":"auto/openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"auto/openai/gpt-5.6-luna":{"id":"auto/openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"auto/openai/gpt-oss-safeguard-20b":{"id":"auto/openai/gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.08,"output":0.24},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"auto/openai/gpt-5.4":{"id":"auto/openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"auto/openai/gpt-4o":{"id":"auto/openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"auto/openai/gpt-oss-20b":{"id":"auto/openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.113,"output":0.4745},"canonical_model_id":"openai/gpt-oss-20b"},"auto/openai/o3":{"id":"auto/openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"auto/openai/gpt-6-astra":{"id":"auto/openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"auto/openai/gpt-oss-safeguard-120b":{"id":"auto/openai/gpt-oss-safeguard-120b","name":"GPT OSS Safeguard 120B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.18,"output":0.73},"canonical_model_id":"openai/gpt-oss-safeguard-120b"},"auto/openai/gpt-5.6-sol":{"id":"auto/openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"auto/openai/gpt-6.1-sol":{"id":"auto/openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"auto/openai/gpt-oss-120b":{"id":"auto/openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.1695,"output":0.6779},"canonical_model_id":"openai/gpt-oss-120b"},"auto/openai/gpt-6-sol":{"id":"auto/openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"auto/openai/gpt-5.4-mini":{"id":"auto/openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"auto/openai/gpt-5-nano":{"id":"auto/openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"}}},"stackit":{"id":"stackit","env":["STACKIT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.openai-compat.model-serving.eu01.onstackit.cloud/v1","name":"STACKIT","doc":"https://docs.stackit.cloud/products/data-and-ai/ai-model-serving/basics/available-shared-models","models":{"intfloat/e5-mistral-7b-instruct":{"id":"intfloat/e5-mistral-7b-instruct","name":"E5 Mistral 7B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2023-12-11","last_updated":"2023-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":0.02,"output":0.02}},"cortecs/Llama-3.3-70B-Instruct-FP8-Dynamic":{"id":"cortecs/Llama-3.3-70B-Instruct-FP8-Dynamic","name":"Llama 3.3 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.53,"output":0.76}},"Qwen/Qwen3.6-27B":{"id":"Qwen/Qwen3.6-27B","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.53,"output":0.76},"canonical_model_id":"alibaba/qwen3.6-27b"},"Qwen/Qwen3-VL-Embedding-8B":{"id":"Qwen/Qwen3-VL-Embedding-8B","name":"Qwen3-VL Embedding 8B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.09,"output":0.09}},"Qwen/Qwen3-VL-235B-A22B-Instruct-FP8":{"id":"Qwen/Qwen3-VL-235B-A22B-Instruct-FP8","name":"Qwen3-VL 235B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2024-11-01","last_updated":"2024-11-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":218000,"output":16384},"cost":{"input":1.76,"output":2.05}},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-05-17","last_updated":"2025-05-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":37000,"output":4096},"cost":{"input":0.53,"output":0.76}},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.18,"output":0.29},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":8192},"cost":{"input":0.53,"output":0.76},"canonical_model_id":"openai/gpt-oss-120b"}}},"engy":{"id":"engy","env":["ENGY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.engy.ai/v1","name":"engy","doc":"https://engy.ai/pricing","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":327680,"input":294912,"output":32768},"cost":{"input":0.135,"output":0.45,"cache_read":0.027},"canonical_model_id":"zhipuai/glm-5.3-flash"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":920576,"output":128000},"cost":{"input":0.045,"output":0.09,"cache_read":0.009},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":208192,"input":200000,"output":8192},"cost":{"input":0.045,"output":0.3,"cache_read":0.015},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1113088,"input":1047552,"output":65536},"cost":{"input":1.95,"output":9.75,"cache_read":0.195},"canonical_model_id":"moonshotai/kimi-k3"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":327680,"input":294912,"output":32768},"cost":{"input":0.98,"output":3.08,"cache_read":0.18},"canonical_model_id":"zhipuai/glm-5.3"},"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1001536,"input":936000,"output":65536},"cost":{"input":0.045,"output":0.32,"cache_read":0.015},"canonical_model_id":"alibaba/qwen3.8-27b"},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":983040,"output":65536},"cost":{"input":0.04,"output":0.08,"cache_read":0.008},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":229376,"output":32768},"cost":{"input":0.68,"output":1.5,"cache_read":0.18},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":983040,"output":65536},"cost":{"input":0.04,"output":0.08,"cache_read":0.008},"canonical_model_id":"deepseek/deepseek-v4.1-flash"}}},"upstage":{"id":"upstage","env":["UPSTAGE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.upstage.ai/v1/solar","name":"Upstage","doc":"https://developers.upstage.ai/docs/apis/chat","models":{"solar-pro4":{"id":"solar-pro4","name":"Solar Pro 4","description":"Upstage's flagship model, specialized for agentic use","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06}},"solar-pro3":{"id":"solar-pro3","name":"solar-pro3","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.25,"output":0.25}},"solar-mini":{"id":"solar-mini","name":"solar-mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"solar-mini","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2024-06-12","last_updated":"2025-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":4096},"cost":{"input":0.15,"output":0.15}},"solar-mini4":{"id":"solar-mini4","name":"Solar Mini 4","description":"Upstage's compact, cost-efficient model for agentic use","family":"solar-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.1,"output":0.4,"cache_read":0.01}},"solar-pro2":{"id":"solar-pro2","name":"solar-pro2","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":8192},"cost":{"input":0.25,"output":0.25}}}},"kimi-code-plan-cn":{"id":"kimi-code-plan-cn","env":["KIMI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.kimi.com/coding/v1","name":"Kimi For Coding (kimi.com)","doc":"https://www.kimi.com/code/docs/en/kimi-code/models.html","models":{"k3":{"id":"k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k3"},"kimi-for-coding":{"id":"kimi-for-coding","name":"kimi-for-coding","description":"Kimi coding model with more efficient thinking and up to 1M context, available through Kimi Code","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.8-preview"},"k3-256k":{"id":"k3-256k","name":"Kimi K3-256K","description":"256K-context version of Kimi K3, reducing token consumption for shorter coding sessions","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k3"},"kimi-for-coding-highspeed":{"id":"kimi-for-coding-highspeed","name":"Kimi For Coding HighSpeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"}}},"modelscope":{"id":"modelscope","env":["MODELSCOPE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api-inference.modelscope.cn/v1","name":"ModelScope","doc":"https://modelscope.cn/docs/model-service/API-Inference/intro","models":{"Qwen/Qwen3-Coder-30B-A3B-Instruct":{"id":"Qwen/Qwen3-Coder-30B-A3B-Instruct","name":"Qwen3 Coder 30B A3B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0}},"Qwen/Qwen3-235B-A22B-Thinking-2507":{"id":"Qwen/Qwen3-235B-A22B-Thinking-2507","name":"Qwen3-235B-A22B-Thinking-2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0,"output":0}},"Qwen/Qwen3-30B-A3B-Thinking-2507":{"id":"Qwen/Qwen3-30B-A3B-Thinking-2507","name":"Qwen3 30B A3B Thinking 2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0}},"Qwen/Qwen3-30B-A3B-Instruct-2507":{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507","name":"Qwen3 30B A3B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0,"output":0}},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0,"output":0}},"ZhipuAI/GLM-4.6":{"id":"ZhipuAI/GLM-4.6","name":"GLM-4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":98304},"cost":{"input":0,"output":0}},"ZhipuAI/GLM-4.5":{"id":"ZhipuAI/GLM-4.5","name":"GLM-4.5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0,"output":0}}}},"cohere":{"id":"cohere","env":["COHERE_API_KEY"],"npm":"@ai-sdk/cohere","name":"Cohere","doc":"https://docs.cohere.com/docs/models","models":{"command-a-reasoning-08-2025":{"id":"command-a-reasoning-08-2025","name":"Command A Reasoning","description":"Cohere reasoning model for multilingual enterprise agents, tools, and complex workflows","family":"command-a","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a-reasoning-08-2025"},"command-a-03-2025":{"id":"command-a-03-2025","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a-03-2025"},"command-a-plus-05-2026":{"id":"command-a-plus-05-2026","name":"Command A Plus","description":"Cohere's stronger command model for multilingual agents and enterprise workflows","family":"command-a","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04-01","release_date":"2026-05-20","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":64000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a-plus-05-2026"},"command-r7b-arabic-02-2025":{"id":"command-r7b-arabic-02-2025","name":"Command R7B Arabic","description":"Open Command R model optimized for Arabic enterprise chat, RAG, and cultural knowledge","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-02-27","last_updated":"2025-02-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.0375,"output":0.15},"canonical_model_id":"cohere/command-r7b-arabic-02-2025"},"command-a-vision-07-2025":{"id":"command-a-vision-07-2025","name":"Command A Vision","description":"Cohere vision model for multilingual document analysis, OCR, and image understanding","family":"command-a","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a-vision-07-2025"},"c4ai-aya-vision-32b":{"id":"c4ai-aya-vision-32b","name":"Aya Vision 32B","description":"Open multilingual vision model for OCR, visual reasoning, and image question answering","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-04","last_updated":"2025-05-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4000},"canonical_model_id":"cohere/c4ai-aya-vision-32b"},"tiny-aya-global":{"id":"tiny-aya-global","name":"Tiny Aya Global","description":"Compact multilingual model balanced across 70 languages and regions","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"output":8000},"canonical_model_id":"cohere/tiny-aya-global"},"command-a-translate-08-2025":{"id":"command-a-translate-08-2025","name":"Command A Translate","description":"Translation model for multilingual conversion, localization, and cross-language workflows","family":"command-a","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"output":8000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a-translate-08-2025"},"north-mini-code-1-0":{"id":"north-mini-code-1-0","name":"North Mini Code","description":"Cohere coding model for practical software engineering and agentic edits","family":"north","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-09-23","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://api.cohere.ai/compatibility/v1"},"cost":{"input":0,"output":0},"canonical_model_id":"cohere/north-mini-code-1-0"},"command-r-08-2024":{"id":"command-r-08-2024","name":"Command R","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"cohere/command-r-08-2024"},"tiny-aya-fire":{"id":"tiny-aya-fire","name":"Tiny Aya Fire","description":"Compact multilingual model specialized for South Asian languages","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"output":8000},"canonical_model_id":"cohere/tiny-aya-fire"},"tiny-aya-earth":{"id":"tiny-aya-earth","name":"Tiny Aya Earth","description":"Compact multilingual model specialized for West Asian and African languages","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"output":8000},"canonical_model_id":"cohere/tiny-aya-earth"},"tiny-aya-water":{"id":"tiny-aya-water","name":"Tiny Aya Water","description":"Compact multilingual model specialized for European and Asia-Pacific languages","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"output":8000},"canonical_model_id":"cohere/tiny-aya-water"},"command-r7b-12-2024":{"id":"command-r7b-12-2024","name":"Command R7B","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-12-13","last_updated":"2024-12-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.0375,"output":0.15},"canonical_model_id":"cohere/command-r7b-12-2024"},"north-small-translate-09-2026":{"id":"north-small-translate-09-2026","name":"North Small Translate","description":"Open-weight translation model for more than 50 languages and locale variants","family":"north","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-09-09","last_updated":"2026-09-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16000},"cost":{"input":0,"output":0},"canonical_model_id":"cohere/north-small-translate-1-0"},"c4ai-aya-expanse-32b":{"id":"c4ai-aya-expanse-32b","name":"Aya Expanse 32B","description":"Open multilingual model optimized for generation across 23 languages","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-10-24","last_updated":"2024-10-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"cohere/c4ai-aya-expanse-32b"},"command-r-plus-08-2024":{"id":"command-r-plus-08-2024","name":"Command R+","description":"Cohere's RAG workhorse for long-context enterprise search and tool use","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-r-plus-08-2024"}}},"umans-ai-coding-plan":{"id":"umans-ai-coding-plan","env":["UMANS_AI_CODING_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.code.umans.ai/v1","name":"Umans AI Coding Plan","doc":"https://app.umans.ai/offers/code/docs","models":{"umans-glm-5.3":{"id":"umans-glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131071},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"umans-flash":{"id":"umans-flash","name":"Umans Flash","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"umans-deepseek-v4.1-flash":{"id":"umans-deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393215},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"umans-kimi-k3":{"id":"umans-kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k3"},"umans-coder":{"id":"umans-coder","name":"Umans Coder","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131071},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"umans-qwen3.6-35b-a3b":{"id":"umans-qwen3.6-35b-a3b","name":"Qwen3.6 35B A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"umans-glm-5.3-flash":{"id":"umans-glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131071},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"umans-deepseek-v4-flash-0731":{"id":"umans-deepseek-v4-flash-0731","name":"DeepSeek V4 Flash","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393215},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"}}},"perplexity":{"id":"perplexity","env":["PERPLEXITY_API_KEY"],"npm":"@ai-sdk/perplexity","name":"Perplexity","doc":"https://docs.perplexity.ai","models":{"sonar-deep-research":{"id":"sonar-deep-research","name":"Perplexity Sonar Deep Research","description":"Sonar search model for current answers, retrieval, and citation-backed chat","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-02-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":2,"output":8,"reasoning":3}},"sonar-pro":{"id":"sonar-pro","name":"Sonar Pro","description":"Deeper Sonar search model with broader retrieval and stronger synthesis","family":"sonar-pro","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":3,"output":15}},"sonar":{"id":"sonar","name":"Sonar","description":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","family":"sonar","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":1,"output":1}},"sonar-reasoning-pro":{"id":"sonar-reasoning-pro","name":"Sonar Reasoning Pro","description":"Web-grounded Sonar for multi-step research questions that need cited reasoning","family":"sonar-reasoning","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":2,"output":8}}}},"infomaniak":{"id":"infomaniak","env":["INFOMANIAK_API_KEY","INFOMANIAK_PRODUCT_ID"],"npm":"@ai-sdk/openai-compatible","api":"https://api.infomaniak.com/2/ai/${INFOMANIAK_PRODUCT_ID}/openai/v1","name":"Infomaniak","doc":"https://www.infomaniak.com/en/hosting/ai-services/open-source-models","models":{"mini_lm_l12_v2":{"id":"mini_lm_l12_v2","name":"All-MiniLM-L12-v2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2021-08-30","last_updated":"2026-08-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128,"input":128,"output":384},"cost":{"input":0,"output":0}},"bge_multilingual_gemma2":{"id":"bge_multilingual_gemma2","name":"BGE Multilingual Gemma2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-07-25","last_updated":"2026-08-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"input":8000,"output":3584},"cost":{"input":0.08,"output":0}},"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8":{"id":"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8","name":"Nemotron 3 Nano 30B A3B FP8","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2026-08-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":262144},"status":"beta","cost":{"input":0.06,"output":0.25},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"swiss-ai/Apertus-v1.5-70B":{"id":"swiss-ai/Apertus-v1.5-70B","name":"Apertus v1.5 70B","description":"Open, ethically-sourced Swiss AI model for multilingual, multimodal chat and instruction following","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-07-24","last_updated":"2026-08-01","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":100000,"input":100000,"output":8192},"status":"beta","cost":{"input":0.87,"output":3.1}},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-08-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"status":"beta","cost":{"input":0.74,"output":3.72},"canonical_model_id":"moonshotai/kimi-k2.6"},"mistralai/Mistral-Small-4-119B-2603":{"id":"mistralai/Mistral-Small-4-119B-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-08-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.25,"output":0.93},"canonical_model_id":"mistral/mistral-small-2603"},"mistralai/Ministral-3-14B-Instruct-2512":{"id":"mistralai/Ministral-3-14B-Instruct-2512","name":"Ministral 3 14B Instruct","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2026-08-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":100000,"input":100000,"output":25600},"status":"beta","cost":{"input":0.37,"output":0.5}},"Qwen/Qwen3.5-122B-A10B-FP8":{"id":"Qwen/Qwen3.5-122B-A10B-FP8","name":"Qwen3.5 122B-A10B FP8","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-08-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65536},"cost":{"input":0.5,"output":3.97},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"Qwen/Qwen3.5-397B-A17B-FP8":{"id":"Qwen/Qwen3.5-397B-A17B-FP8","name":"Qwen3.5 397B-A17B FP8","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-08-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65536},"status":"beta","cost":{"input":0.99,"output":4.46},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-08-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":100000,"input":100000,"output":32768},"cost":{"input":0.25,"output":0.5},"canonical_model_id":"google/gemma-4-31b-it"}}},"ainetcafe":{"id":"ainetcafe","env":["AINETCAFE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://microquickjs.com/v1","name":"ainetcafe","doc":"https://ainetcafe.com/k3/guides/","models":{"Kimi-K3":{"id":"Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":2.1,"output":10.5,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"}}},"tokengo":{"id":"tokengo","env":["TOKENGO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.tokengo.com/v1","name":"TokenGo","doc":"https://www.tokengo.com/docs","models":{"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.435,"output":0.87},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v3.1":{"id":"deepseek/deepseek-v3.1","name":"DeepSeek-V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.19,"output":0.71,"cache_read":0.06},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.098,"output":0.196,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":64000},"cost":{"input":0.2174,"output":0.326,"cache_read":0.06},"canonical_model_id":"deepseek/deepseek-v3.2"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.075,"output":0.025,"cache_read":0.015},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.89,"output":3.2647,"cache_read":0.2226},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":2.65,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"}}},"zhipuai-coding-plan":{"id":"zhipuai-coding-plan","env":["ZHIPU_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://open.bigmodel.cn/api/coding/paas/v4","name":"Zhipu AI Coding Plan","doc":"https://docs.bigmodel.cn/cn/coding-plan/overview","models":{"glm-4.6v":{"id":"glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6v"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5.3-highspeed":{"id":"glm-5.3-highspeed","name":"GLM-5.3 Highspeed","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"}}},"hyper":{"id":"hyper","env":["HYPER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://hyper.charm.land/v1","name":"Charm Hyper","doc":"https://hyper.charm.land","models":{"qwen3.7-flash":{"id":"qwen3.7-flash","name":"Qwen3.7 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":64000},"cost":{"input":0.2,"output":0.8,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.7-flash"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.16332,"output":0.5444,"cache_read":0.031575},"canonical_model_id":"zhipuai/glm-5.3-flash"},"inkling":{"id":"inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":1.0888,"output":4.40964,"cache_read":0.185096},"canonical_model_id":"thinkingmachines/inkling"},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262100,"output":6553},"cost":{"input":0.484,"output":1.852,"cache_read":0.242},"canonical_model_id":"minimax/MiniMax-M2.7"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.44,"output":1.32,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":16000},"cost":{"input":3.2664,"output":16.332,"cache_read":0.32664},"canonical_model_id":"moonshotai/kimi-k3"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen3.8-2.4t-a95b":{"id":"qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":1.2,"output":4.8,"cache_read":0.24},"canonical_model_id":"alibaba/qwen3.7-plus"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":262144},"cost":{"input":1.437216,"output":4.311648,"cache_read":0.047907},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.52432,"output":4.79072,"cache_read":0.283088},"canonical_model_id":"zhipuai/glm-5.3"},"gemma-4-26b-a4b-it":{"id":"gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":25600},"cost":{"input":0.098,"output":0.334,"cache_read":0.049},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"prism":{"id":"prism","name":"Prism","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":true,"release_date":"1970-01-01","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":16000},"cost":{"input":3.2664,"output":16.332,"cache_read":0.32664}},"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.5,"output":3,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3.8-27b"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16000},"cost":{"input":1.03436,"output":4.3552,"cache_read":0.206872},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":512000},"cost":{"input":0.32664,"output":1.30656,"cache_read":0.064239},"canonical_model_id":"minimax/MiniMax-M3"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":13107},"cost":{"input":0.178,"output":0.68,"cache_read":0.089},"canonical_model_id":"openai/gpt-oss-120b"},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0.33,"output":1.31,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2.5,"output":7.5,"cache_read":0.5},"canonical_model_id":"alibaba/qwen3.7-max"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":1.52432,"output":4.79072,"cache_read":0.152432},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0.33,"output":1.31,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max Preview","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-19","last_updated":"2026-07-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"kimi-k2-thinking":{"id":"kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":26214},"cost":{"input":0.6,"output":2.5,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k2-thinking"}}},"neosmith":{"id":"neosmith","env":["NEOSMITH_API_KEY"],"npm":"@ai-sdk/openai","api":"https://router.neosmith.ai/v1","name":"NeoSmith","doc":"https://neosmith.ai/docs","models":{"neosmith.neolite":{"id":"neosmith.neolite","name":"NeoSmith NeoLite","description":"Sealed single-model budget tier. 512K context, text and images, tool use, and no escalation of any kind.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-20","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":512000,"output":64000},"cost":{"input":0.6,"output":2.4,"cache_read":0.08,"cache_write":0}},"neosmith.intelligent-basic":{"id":"neosmith.intelligent-basic","name":"NeoSmith Basic","description":"Cost-capped tier. Intelligent routing with a Claude Sonnet ceiling — Opus is never invoked.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-26","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":1.17,"output":4.37,"cache_read":0.22,"cache_write":0}},"neosmith.intelligent-maestro":{"id":"neosmith.intelligent-maestro","name":"NeoSmith Maestro","description":"Highest-accuracy coding tier. Hard, self-contained problems run NeoSmith's premium multi-model solver; everything else gets the strongest intelligence tier.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.4,"output":12,"cache_read":0.35,"cache_write":0}},"neosmith.intelligent-pro":{"id":"neosmith.intelligent-pro","name":"NeoSmith Pro","description":"Default production tier. Intelligent NeoSmith routing with a Claude Opus ceiling on escalation.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-26","last_updated":"2026-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":1.81,"output":8.39,"cache_read":0.3,"cache_write":0}}}},"xiaomi":{"id":"xiaomi","env":["XIAOMI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.xiaomimimo.com/v1","name":"Xiaomi","doc":"https://platform.xiaomimimo.com/#/docs","models":{"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-06-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.435,"output":0.87,"cache_read":0.0036}},"mimo-v2.6-pro-ultraspeed":{"id":"mimo-v2.6-pro-ultraspeed","name":"MiMo-V2.6-Pro-UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.35,"output":8.7,"cache_read":0.036}},"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-06-24","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.0028}},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028}},"mimo-v2.5-pro-ultraspeed":{"id":"mimo-v2.5-pro-ultraspeed","name":"MiMo-V2.5-Pro-UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-06-08","last_updated":"2026-06-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":1.305,"output":2.61,"cache_read":0.0108}},"mimo-v2-flash":{"id":"mimo-v2-flash","name":"MiMo-V2-Flash","description":"Legacy model retained for compatibility with older integrations","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12-01","release_date":"2025-12-16","last_updated":"2026-06-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.0028}},"mimo-v2-omni":{"id":"mimo-v2-omni","name":"MiMo-V2-Omni","description":"Legacy model retained for compatibility with older integrations","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-06-24","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.0028}},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"MiMo Pro model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036}},"mimo-v2-pro":{"id":"mimo-v2-pro","name":"MiMo-V2-Pro","description":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-06-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.435,"output":0.87,"cache_read":0.0036}}}},"submodel":{"id":"submodel","env":["SUBMODEL_INSTAGEN_ACCESS_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://llm.submodel.ai/v1","name":"submodel","doc":"https://submodel.gitbook.io","models":{"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-08-23","last_updated":"2025-08-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":75000,"output":163840},"cost":{"input":0.2,"output":0.8}},"deepseek-ai/DeepSeek-R1-0528":{"id":"deepseek-ai/DeepSeek-R1-0528","name":"DeepSeek R1 0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-08-23","last_updated":"2025-08-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":75000,"output":163840},"cost":{"input":0.5,"output":2.15}},"deepseek-ai/DeepSeek-V3-0324":{"id":"deepseek-ai/DeepSeek-V3-0324","name":"DeepSeek V3 0324","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-08-23","last_updated":"2025-08-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":75000,"output":163840},"cost":{"input":0.2,"output":0.8}},"Qwen/Qwen3-235B-A22B-Thinking-2507":{"id":"Qwen/Qwen3-235B-A22B-Thinking-2507","name":"Qwen3 235B A22B Thinking 2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-08-23","last_updated":"2025-08-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.2,"output":0.6}},"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8":{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8","name":"Qwen3 Coder 480B A35B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-08-23","last_updated":"2025-08-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.2,"output":0.8}},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-08-23","last_updated":"2025-08-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.2,"output":0.3}},"zai-org/GLM-4.5-FP8":{"id":"zai-org/GLM-4.5-FP8","name":"GLM 4.5 FP8","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.2,"output":0.8}},"zai-org/GLM-4.5-Air":{"id":"zai-org/GLM-4.5-Air","name":"GLM 4.5 Air","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-air","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.1,"output":0.5}},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-08-23","last_updated":"2025-08-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.1,"output":0.5}}}},"wandb":{"id":"wandb","env":["WANDB_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.inference.wandb.ai/v1","name":"CoreWeave","doc":"https://docs.wandb.ai/inference","models":{"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B":{"id":"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B","name":"Nemotron 3.5 Lightning","description":"Nemotron 3.5 Lightning is an MoE model built for fast, reliable agentic tasks across use cases such as financial services, cybersecurity, telecom, and retail.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.07,"output":0.2,"cache_read":0.04},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B":{"id":"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B","name":"Nemotron 3 Ultra","description":"Nemotron 3 Ultra is a powerful MoE model designed for long-running agents across coding, deep research, and enterprise automation.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":2.15,"cache_read":0.1},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B":{"id":"nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B","name":"Nemotron 3.5 Super VL","description":"Nemotron 3.5 Super VL is a multimodal MoE model built for visual understanding, reasoning, and agentic workflows.","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.15,"output":0.5,"cache_read":0.02},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek V3.1","description":"A large hybrid model that supports both thinking and non-thinking modes via prompt templates.","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":161000,"output":161000},"cost":{"input":0.55,"output":1.65,"cache_read":0.55},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"DeepSeek V4-Flash-0731 is an MoE model great for coding, reasoning, and agentic workloads.","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.13,"output":0.28,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-V4.1-Flash":{"id":"deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash is a multimodal MoE model for coding, reasoning, and agentic workloads with long contexts.","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.15,"output":0.65,"cache_read":0.002},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4-Pro-0813 is a 1.6T-parameter MoE model excelling at advanced reasoning, coding, and complex agentic workloads.","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.31,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"meta-llama/Llama-3.3-70B-Instruct":{"id":"meta-llama/Llama-3.3-70B-Instruct","name":"Llama 3.3 70B","description":"Multilingual model excelling in conversational tasks, detailed instruction-following, and coding.","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.71,"output":0.71,"cache_read":0.71},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/Llama-3.1-8B-Instruct":{"id":"meta-llama/Llama-3.1-8B-Instruct","name":"Llama 3.1 8B","description":"Efficient conversational model optimized for responsive multilingual chatbot interactions.","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.22,"output":0.22,"cache_read":0.22},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Kimi K2.6 is a multimodal Mixture-of-Experts language model featuring 32 billion activated parameters and a total of 1 trillion parameters.","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.65,"output":3.41,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Kimi K2.7 Code is a 1T-parameter MoE model with 32B active parameters purpose-built for long-horizon agentic coding and software engineering.","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.71,"output":3.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax M3","description":"MiniMax M3 is a multimodal MoE model with 23B active parameters optimized for coding and agentic workflows.","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.23,"output":0.96,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M3"},"reflection/Beam-1.0-501B-A23B":{"id":"reflection/Beam-1.0-501B-A23B","name":"Beam 1.0 501B A23B","description":"Beam is a 501B-parameter MoE model with 23B active parameters built for coding, reasoning, and agentic workflows.","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.3,"output":1,"cache_read":0.04},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"Qwen/Qwen3.6-35B-A3B":{"id":"Qwen/Qwen3.6-35B-A3B","name":"Qwen3.6 35B A3B","description":"Qwen3.6-35B-A3B is an MoE multimodal model with 262K context optimized for agentic coding workflows.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.25,"output":1.25,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8 27B","description":"Qwen3.8-27B is a dense multimodal model suited for coding, research, vision, and long-running agent tasks.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.4,"output":3,"cache_read":0.15},"canonical_model_id":"alibaba/qwen3.8-27b"},"ibm-granite/granite-4.2-8b":{"id":"ibm-granite/granite-4.2-8b","name":"Granite 4.2 8B","description":"Granite 4.2 8B is an instruct model capable of enhanced tool calling, instruction following, and chat capabilities.","family":"granite","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-25","last_updated":"2026-08-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.1,"output":0.15,"cache_read":0.05},"canonical_model_id":"ibm/granite-4.2-8b"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM 5.3","description":"GLM-5.3 is a 744B-parameter MoE model built for advanced reasoning, software engineering, and long-horizon agentic tasks.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.2,"output":4,"cache_read":0.19},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM 5.2","description":"GLM-5.2 is a Mixture-of-Experts language model featuring 40 billion activated parameters and a total of 744 billion parameters.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.62,"output":2.42,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM 5.3 Flash","description":"GLM-5.3-Flash is a natively multimodal model with 320B total parameters and 18B active parameters.","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.15,"output":0.5,"cache_read":0.02},"canonical_model_id":"zhipuai/glm-5.3-flash"},"google/gemma-4-26B-A4B-it":{"id":"google/gemma-4-26B-A4B-it","name":"Gemma 4 26B A4B","description":"Gemma 4 26B A4B is a multimodal MoE model with LoRA support and function calling for agentic workflows.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.3,"cache_read":0.05},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B","description":"Gemma 4 31B Dense is designed for advanced reasoning, agentic workflows, and longer context and is natively trained on 140+ languages.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.34,"cache_read":0.1},"canonical_model_id":"google/gemma-4-31b-it"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"gpt-oss-20b","description":"Lower latency Mixture-of-Experts model trained on OpenAI's Harmony response format with reasoning capabilities.","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.03,"output":0.13,"cache_read":0.03},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"gpt-oss-120b","description":"Efficient Mixture-of-Experts model designed for high-reasoning, agentic and general-purpose use cases.","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.03,"output":0.17,"cache_read":0.03},"canonical_model_id":"openai/gpt-oss-120b"}}},"vivgrid":{"id":"vivgrid","env":["VIVGRID_API_KEY"],"npm":"@ai-sdk/openai","api":"https://api.vivgrid.com/v1","name":"Vivgrid","doc":"https://docs.vivgrid.com/models","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.04},"canonical_model_id":"zhipuai/glm-5.3-flash"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.435,"output":0.87,"cache_read":0.003625},"canonical_model_id":"deepseek/deepseek-v4-pro"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":10,"output":50,"cache_read":1.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT 5.6 Terra","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"openai/gpt-5.6-terra"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":10,"output":50,"cache_read":0.5,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.25,"output":2,"cache_read":0.03}},"viv-fast":{"id":"viv-fast","name":"Viv Fast","description":"Fast coding model","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-09","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":256000},"cost":{"input":0.13,"output":0.4,"cache_read":0.05},"canonical_model_id":"vivgrid/viv-fast"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.35,"output":3,"reasoning":3,"cache_read":0.05},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT 5.6 Luna","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":1,"output":6,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"openai/gpt-5.6-luna"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}}},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.2,"output":4.2,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.75,"output":3.75,"cache_read":0.15},"canonical_model_id":"google/gemini-3.8-flash"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":2.5,"output":15,"cache_read":0.25}},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-01-14","last_updated":"2026-01-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.3,"reasoning":0.3,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT 5.6 Sol","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.28,"output":0.42}},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.2,"output":1.25,"cache_read":0.02}},"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"gpt-5.1-codex-max":{"id":"gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075}},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.2,"output":4.2,"cache_read":0.3},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.31,"output":1.23,"cache_read":0.01},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"gemini-3.1-flash-lite-preview":{"id":"gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":1}},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}}}},"edenai":{"id":"edenai","env":["EDENAI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.edenai.run/v3","name":"Eden AI","doc":"https://docs.edenai.co","models":{"deepseek-v4-flash-vision-exp":{"id":"deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-latest":{"id":"anthropic/claude-opus-latest","name":"Claude Opus Latest (Claude Opus 5.5)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4-5-20251101":{"id":"anthropic/claude-opus-4-5-20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-haiku-latest":{"id":"anthropic/claude-haiku-latest","name":"Claude Haiku Latest (Claude Haiku 5.5)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-fable-5-1":{"id":"anthropic/claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-opus-4-8":{"id":"anthropic/claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-latest":{"id":"anthropic/claude-fable-latest","name":"Claude Fable Latest (Claude Fable 5.1)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-sonnet-5-5":{"id":"anthropic/claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-haiku-5-5":{"id":"anthropic/claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-5-5":{"id":"anthropic/claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-opus-4-5":{"id":"anthropic/claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-latest":{"id":"anthropic/claude-sonnet-latest","name":"Claude Sonnet Latest (Claude Sonnet 5.5)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"flexai/Muse-Glimmer-30B":{"id":"flexai/Muse-Glimmer-30B","name":"Muse Glimmer 30B (FlexAI)","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.045},"canonical_model_id":"meta/muse-glimmer-30b"},"flexai/DeepSeek-V4-Flash-0731":{"id":"flexai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731 (FlexAI)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.119,"output":0.238,"cache_read":0.0178},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"flexai/DeepSeek-V4.1-Flash":{"id":"flexai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash (FlexAI)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.1,"output":0.66,"cache_read":0.015},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"flexai/gpt-oss-20b":{"id":"flexai/gpt-oss-20b","name":"GPT OSS 20B (FlexAI)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.02,"output":0.1,"cache_read":0.003},"canonical_model_id":"openai/gpt-oss-20b"},"flexai/DeepSeek-V4-Flash-0731@eu":{"id":"flexai/DeepSeek-V4-Flash-0731@eu","name":"DeepSeek V4 Flash 0731 (FlexAI, EU)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.119,"output":0.238,"cache_read":0.0178},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"flexai/gpt-oss-120b":{"id":"flexai/gpt-oss-120b","name":"GPT OSS 120B (FlexAI)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.05,"output":0.25,"cache_read":0.0075},"canonical_model_id":"openai/gpt-oss-120b"},"flexai/Step-3.7-Flash":{"id":"flexai/Step-3.7-Flash","name":"Step 3.7 Flash (FlexAI)","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.03},"canonical_model_id":"stepfun/step-3.7-flash"},"xai/grok-4.3":{"id":"xai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"xai/grok-4-fast-non-reasoning":{"id":"xai/grok-4-fast-non-reasoning","name":"X-Ai/Grok-4-Fast-Non-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05,"tiers":[{"input":0.4,"output":1,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"xai/grok-4-fast-non-reasoning"},"xai/grok-4.20-0309-reasoning":{"id":"xai/grok-4.20-0309-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"xai/grok-4.7":{"id":"xai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"xai/grok-build-0.1":{"id":"xai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}},"canonical_model_id":"xai/grok-build-0.1"},"xai/grok-4.5":{"id":"xai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"xai/grok-latest":{"id":"xai/grok-latest","name":"Grok Latest (Grok 4.7)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"xai/grok-4.6":{"id":"xai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"xai/grok-4.20-0309-non-reasoning":{"id":"xai/grok-4.20-0309-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"amazon/mistral.pixtral-large-2502-v1:0@us":{"id":"amazon/mistral.pixtral-large-2502-v1:0@us","name":"Pixtral Large (25.02) (Amazon Bedrock, US)","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-08","last_updated":"2025-04-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/pixtral-large-2502"},"amazon/openai.gpt-oss-safeguard-20b@us":{"id":"amazon/openai.gpt-oss-safeguard-20b@us","name":"GPT OSS Safeguard 20B (Amazon Bedrock, US)","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.07,"output":0.2},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"amazon/openai.gpt-oss-safeguard-20b":{"id":"amazon/openai.gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B (Amazon Bedrock)","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.07,"output":0.2},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"amazon/mistral.voxtral-mini-3b-2507":{"id":"amazon/mistral.voxtral-mini-3b-2507","name":"Voxtral Mini 3B 2507 (Amazon Bedrock)","description":"Open audio-language model for speech transcription, audio understanding, and voice-driven tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.04,"output":0.04},"canonical_model_id":"mistral/voxtral-mini-3b-2507"},"amazon/mistral.voxtral-small-24b-2507":{"id":"amazon/mistral.voxtral-small-24b-2507","name":"Voxtral Small 24B 2507 (Amazon Bedrock)","description":"Open audio-language model for speech transcription, audio understanding, and voice-driven tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/voxtral-small-24b-2507"},"amazon/amazon.nova-lite-v1:0@us":{"id":"amazon/amazon.nova-lite-v1:0@us","name":"Nova Lite (US)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.06,"output":0.24,"cache_read":0.015},"canonical_model_id":"amazon/nova-lite"},"amazon/mistral.pixtral-large-2502-v1:0":{"id":"amazon/mistral.pixtral-large-2502-v1:0","name":"Pixtral Large (25.02) (Amazon Bedrock)","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-08","last_updated":"2025-04-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/pixtral-large-2502"},"amazon/zai.glm-4.7-flash":{"id":"amazon/zai.glm-4.7-flash","name":"GLM-4.7-Flash (Amazon Bedrock)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131072},"cost":{"input":0.07,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"amazon/amazon.nova-micro-v1:0":{"id":"amazon/amazon.nova-micro-v1:0","name":"Nova Micro","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.035,"output":0.14,"cache_read":0.00875},"canonical_model_id":"amazon/nova-micro"},"amazon/google.gemma-3-4b-it@us":{"id":"amazon/google.gemma-3-4b-it@us","name":"Gemma 3 4B IT (Amazon Bedrock, US)","description":"Open multimodal Gemma instruction model for efficient text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.04,"output":0.08},"canonical_model_id":"google/gemma-3-4b-it"},"amazon/amazon.nova-pro-v1:0@us":{"id":"amazon/amazon.nova-pro-v1:0@us","name":"Nova Pro (US)","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.8,"output":3.2,"cache_read":0.2},"canonical_model_id":"amazon/nova-pro"},"amazon/google.gemma-3-12b-it":{"id":"amazon/google.gemma-3-12b-it","name":"Gemma 3 12B IT (Amazon Bedrock)","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.09,"output":0.29},"canonical_model_id":"google/gemma-3-12b-it"},"amazon/mistral.voxtral-small-24b-2507@us":{"id":"amazon/mistral.voxtral-small-24b-2507@us","name":"Voxtral Small 24B 2507 (Amazon Bedrock, US)","description":"Open audio-language model for speech transcription, audio understanding, and voice-driven tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/voxtral-small-24b-2507"},"amazon/amazon.nova-lite-v1:0":{"id":"amazon/amazon.nova-lite-v1:0","name":"Nova Lite","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.06,"output":0.24,"cache_read":0.015},"canonical_model_id":"amazon/nova-lite"},"amazon/google.gemma-3-4b-it":{"id":"amazon/google.gemma-3-4b-it","name":"Gemma 3 4B IT (Amazon Bedrock)","description":"Open multimodal Gemma instruction model for efficient text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.04,"output":0.08},"canonical_model_id":"google/gemma-3-4b-it"},"amazon/qwen.qwen3-coder-next":{"id":"amazon/qwen.qwen3-coder-next","name":"Qwen3 Coder Next (Amazon Bedrock)","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.5,"output":1.2},"canonical_model_id":"alibaba/qwen3-coder-next"},"amazon/google.gemma-3-12b-it@us":{"id":"amazon/google.gemma-3-12b-it@us","name":"Gemma 3 12B IT (Amazon Bedrock, US)","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.09,"output":0.29},"canonical_model_id":"google/gemma-3-12b-it"},"amazon/amazon.nova-micro-v1:0@us":{"id":"amazon/amazon.nova-micro-v1:0@us","name":"Nova Micro (US)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.035,"output":0.14,"cache_read":0.00875},"canonical_model_id":"amazon/nova-micro"},"amazon/google.gemma-3-27b-it@us":{"id":"amazon/google.gemma-3-27b-it@us","name":"Gemma 3 27B IT (Amazon Bedrock, US)","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.23,"output":0.38},"canonical_model_id":"google/gemma-3-27b-it"},"amazon/moonshotai.kimi-k2.5":{"id":"amazon/moonshotai.kimi-k2.5","name":"Kimi K2.5 (Amazon Bedrock)","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3},"canonical_model_id":"moonshotai/kimi-k2.5"},"amazon/mistral.voxtral-mini-3b-2507@us":{"id":"amazon/mistral.voxtral-mini-3b-2507@us","name":"Voxtral Mini 3B 2507 (Amazon Bedrock, US)","description":"Open audio-language model for speech transcription, audio understanding, and voice-driven tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.04,"output":0.04},"canonical_model_id":"mistral/voxtral-mini-3b-2507"},"amazon/moonshot.kimi-k2-thinking":{"id":"amazon/moonshot.kimi-k2-thinking","name":"Kimi K2 Thinking (Amazon Bedrock)","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":262144},"cost":{"input":0.6,"output":2.5},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"amazon/zai.glm-4.7-flash@us":{"id":"amazon/zai.glm-4.7-flash@us","name":"GLM-4.7-Flash (Amazon Bedrock, US)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131072},"cost":{"input":0.07,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"amazon/google.gemma-3-27b-it":{"id":"amazon/google.gemma-3-27b-it","name":"Gemma 3 27B IT (Amazon Bedrock)","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.23,"output":0.38},"canonical_model_id":"google/gemma-3-27b-it"},"amazon/amazon.nova-pro-v1:0":{"id":"amazon/amazon.nova-pro-v1:0","name":"Nova Pro","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.8,"output":3.2,"cache_read":0.2},"canonical_model_id":"amazon/nova-pro"},"moonshot/kimi-k2.6":{"id":"moonshot/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshot/kimi-k3":{"id":"moonshot/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshot/kimi-k2.7-code-highspeed":{"id":"moonshot/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshot/kimi-k2.7-code":{"id":"moonshot/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"databricks/databricks-inkling":{"id":"databricks/databricks-inkling","name":"Inkling (Databricks)","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":1048576},"cost":{"input":1,"output":4.05,"cache_read":0.1,"cache_write":1.00002},"canonical_model_id":"thinkingmachines/inkling"},"databricks/databricks-deepseek-v4-pro-0813":{"id":"databricks/databricks-deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813 (Databricks)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.959999,"cache_read":0.132,"cache_write":1.31999},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"databricks/databricks-qwen3-next-80b-a3b-instruct@eu":{"id":"databricks/databricks-qwen3-next-80b-a3b-instruct@eu","name":"Qwen3-Next 80B-A3B Instruct (Databricks, EU)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15001,"output":1.20001,"cache_read":0.015001,"cache_write":0.15001},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"databricks/databricks-qwen3-next-80b-a3b-instruct":{"id":"databricks/databricks-qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct (Databricks)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15001,"output":1.20001,"cache_read":0.015001,"cache_write":0.15001},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"databricks/databricks-gpt-oss-120b":{"id":"databricks/databricks-gpt-oss-120b","name":"GPT OSS 120B (Databricks)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15001,"output":0.59997,"cache_read":0.015001,"cache_write":0.15001},"canonical_model_id":"openai/gpt-oss-120b"},"databricks/databricks-gpt-oss-20b":{"id":"databricks/databricks-gpt-oss-20b","name":"GPT OSS 20B (Databricks)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.07,"output":0.30002,"cache_read":0.007,"cache_write":0.07},"canonical_model_id":"openai/gpt-oss-20b"},"databricks/databricks-gpt-oss-120b@eu":{"id":"databricks/databricks-gpt-oss-120b@eu","name":"GPT OSS 120B (Databricks, EU)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15001,"output":0.59997,"cache_read":0.015001,"cache_write":0.15001},"canonical_model_id":"openai/gpt-oss-120b"},"databricks/databricks-deepseek-v4-flash-0731":{"id":"databricks/databricks-deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731 (Databricks)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.014,"cache_write":0.14},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"databricks/databricks-gpt-oss-20b@eu":{"id":"databricks/databricks-gpt-oss-20b@eu","name":"GPT OSS 20B (Databricks, EU)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.07,"output":0.30002,"cache_read":0.007,"cache_write":0.07},"canonical_model_id":"openai/gpt-oss-20b"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-chat":{"id":"deepseek/deepseek-chat","name":"DeepSeek Chat","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-chat"},"ovhcloud/gpt-oss-20b":{"id":"ovhcloud/gpt-oss-20b","name":"GPT OSS 20B (OVHcloud)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.05,"output":0.18},"canonical_model_id":"openai/gpt-oss-20b"},"ovhcloud/gpt-oss-120b":{"id":"ovhcloud/gpt-oss-120b","name":"GPT OSS 120B (OVHcloud)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.09,"output":0.47},"canonical_model_id":"openai/gpt-oss-120b"},"together_ai/deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"together_ai/deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731 (Together AI)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"together_ai/deepseek-ai/DeepSeek-V4.1-Flash":{"id":"together_ai/deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash (Together AI)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"together_ai/deepseek-ai/DeepSeek-V4.1-Flash-1017":{"id":"together_ai/deepseek-ai/DeepSeek-V4.1-Flash-1017","name":"DeepSeek V4.1 Flash 1017 (Together AI)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"together_ai/deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"together_ai/deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813 (Together AI)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"together_ai/thinkingmachines/Inkling":{"id":"together_ai/thinkingmachines/Inkling","name":"Inkling (Together AI)","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":1048576},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"together_ai/reflection/Beam-1.0-501B-A23B":{"id":"together_ai/reflection/Beam-1.0-501B-A23B","name":"Beam 1.0 501B A23B (Together AI)","description":"Sparse mixture-of-experts reasoning model with 23B active parameters for coding and agentic workflows","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.4,"output":1.2,"cache_read":0.04},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"together_ai/meta-models/Muse-Glimmer-30B":{"id":"together_ai/meta-models/Muse-Glimmer-30B","name":"Muse Glimmer 30B (Together AI)","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.35,"output":1.5,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"together_ai/openai/gpt-oss-120b":{"id":"together_ai/openai/gpt-oss-120b","name":"GPT OSS 120B (Together AI)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"qwen/deepseek-v4-flash-0731":{"id":"qwen/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731 (Alibaba)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.22,"output":0.66,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"qwen/qwen3.8-omni-flash":{"id":"qwen/qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"qwen/qwen-max":{"id":"qwen/qwen-max","name":"Qwen Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":1.6,"output":6.4,"cache_read":0.32},"canonical_model_id":"alibaba/qwen-max"},"qwen/qwen3-coder-flash":{"id":"qwen/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":1.5,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"alibaba/qwen3-coder-flash"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3.8-2.4t-a95b":{"id":"qwen/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"qwen/deepseek-v4-pro-0813":{"id":"qwen/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813 (Alibaba)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.066},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":0.5,"output":3,"cache_read":0.1,"cache_write":0.625},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/deepseek-v4.1-flash-1017":{"id":"qwen/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (Alibaba)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"qwen/deepseek-v4.1-flash":{"id":"qwen/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Alibaba)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen/qwen-mt-plus":{"id":"qwen/qwen-mt-plus","name":"Qwen MT Plus","description":"Flagship Qwen translation model with terminology control and support for 92 languages","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-24","last_updated":"2025-07-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4096,"output":8192},"cost":{"input":2.46,"output":7.37},"canonical_model_id":"alibaba/qwen-mt-plus"},"qwen/qwen3.8-max":{"id":"qwen/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen/qwen3.8-max-0902":{"id":"qwen/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"greenpt/deepseek-v4-flash-0731":{"id":"greenpt/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731 (Greenpt)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15715,"output":0.392875,"cache_read":0.0449},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"greenpt/llama-3.3-70b-instruct":{"id":"greenpt/llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct (Greenpt)","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":1.234749,"output":1.234749},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"greenpt/qwen3-235b-a22b-instruct-2507":{"id":"greenpt/qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B-A22B Instruct 2507 (Greenpt)","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":250000,"output":16384},"cost":{"input":1.01025,"output":3.030749},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"greenpt/gpt-oss-120b":{"id":"greenpt/gpt-oss-120b","name":"GPT OSS 120B (Greenpt)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.2245,"output":0.78575},"canonical_model_id":"openai/gpt-oss-120b"},"greenpt/deepseek-v4.1-flash":{"id":"greenpt/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Greenpt)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.24695,"output":1.234749,"cache_read":0.012347},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"nebius/nvidia/Nemotron-3-Ultra-550b-a55b":{"id":"nebius/nvidia/Nemotron-3-Ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B (Nebius)","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"cost":{"input":1,"output":3,"cache_read":1},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nebius/nvidia/nemotron-3-super-120b-a12b":{"id":"nebius/nvidia/nemotron-3-super-120b-a12b","name":"Nemotron 3 Super 120B A12B (Nebius)","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":0.9,"cache_read":0.3},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nebius/deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"nebius/deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731 (Nebius)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.14},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"nebius/deepseek-ai/DeepSeek-V4.1-Flash":{"id":"nebius/deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash (Nebius)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.3},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"nebius/deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"nebius/deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813 (Nebius)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":979000,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":1.32},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"nebius/meta-models/muse-spark-open":{"id":"nebius/meta-models/muse-spark-open","name":"Muse Spark Open (Nebius)","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":1.25},"canonical_model_id":"meta/muse-spark-open"},"nebius/Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"nebius/Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B-A22B Instruct 2507 (Nebius)","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.2,"output":0.6,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"nebius/google/gemma-3-27b-it":{"id":"nebius/google/gemma-3-27b-it","name":"Gemma 3 27B IT (Nebius)","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":110000,"output":131072},"cost":{"input":0.1,"output":0.3,"cache_read":0.1},"canonical_model_id":"google/gemma-3-27b-it"},"nebius/openai/gpt-oss-120b":{"id":"nebius/openai/gpt-oss-120b","name":"GPT OSS 120B (Nebius)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.15},"canonical_model_id":"openai/gpt-oss-120b"},"ionos/meta-llama/Llama-3.3-70B-Instruct":{"id":"ionos/meta-llama/Llama-3.3-70B-Instruct","name":"Llama-3.3-70B-Instruct (IONOS)","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.729625,"output":0.729625},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"ionos/Qwen/Qwen3-Coder-Next":{"id":"ionos/Qwen/Qwen3-Coder-Next","name":"Qwen3 Coder Next (IONOS)","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":65536},"cost":{"input":0.168375,"output":0.898},"canonical_model_id":"alibaba/qwen3-coder-next"},"ionos/openai/gpt-oss-120b":{"id":"ionos/openai/gpt-oss-120b","name":"GPT OSS 120B (IONOS)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.168375,"output":0.729625},"canonical_model_id":"openai/gpt-oss-120b"},"fireworks_ai/gpt-oss-120b":{"id":"fireworks_ai/gpt-oss-120b","name":"GPT OSS 120B (Fireworks AI)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.014},"canonical_model_id":"openai/gpt-oss-120b"},"fireworks_ai/accounts/fireworks/models/inkling":{"id":"fireworks_ai/accounts/fireworks/models/inkling","name":"Inkling (Fireworks AI)","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"fireworks_ai/accounts/fireworks/models/ember-1":{"id":"fireworks_ai/accounts/fireworks/models/ember-1","name":"Ember-1 (Fireworks AI)","description":"Specialized model from Fireworks built on Kimi K3, producing shorter reasoning traces with approximately 40% fewer tokens while maintaining comparable quality","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"fireworks/ember-1"},"fireworks_ai/accounts/fireworks/models/muse-spark-open":{"id":"fireworks_ai/accounts/fireworks/models/muse-spark-open","name":"Muse Spark Open (Fireworks AI)","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"fireworks_ai/accounts/fireworks/models/gpt-oss-120b":{"id":"fireworks_ai/accounts/fireworks/models/gpt-oss-120b","name":"GPT OSS 120B (Fireworks AI)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"openai/gpt-oss-120b"},"cloudflare/@cf/meta/llama-guard-3-8b":{"id":"cloudflare/@cf/meta/llama-guard-3-8b","name":"Llama-Guard-3-8B (Cloudflare)","description":"Llama 3.1-based safety classifier for moderating prompts and model responses","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.484,"output":0.03},"canonical_model_id":"meta/llama-guard-3-8b"},"cloudflare/@cf/meta/llama-3.2-1b-instruct":{"id":"cloudflare/@cf/meta/llama-3.2-1b-instruct","name":"Llama 3.2 1B Instruct (Cloudflare)","description":"Compact instruction-tuned Llama model for multilingual dialogue and on-device assistants","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":60000,"output":32000},"cost":{"input":0.027,"output":0.201},"canonical_model_id":"meta/llama-3.2-1b-instruct"},"cloudflare/@cf/deepseek-ai/deepseek-v4-flash-0731":{"id":"cloudflare/@cf/deepseek-ai/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731 (Cloudflare)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"cloudflare/@cf/deepseek-ai/deepseek-v4-pro-0813":{"id":"cloudflare/@cf/deepseek-ai/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813 (Cloudflare)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"cloudflare/@cf/deepseek-ai/deepseek-v4.1-flash-1017":{"id":"cloudflare/@cf/deepseek-ai/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (Cloudflare)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"cloudflare/@cf/qwen/qwen2.5-coder-32b-instruct":{"id":"cloudflare/@cf/qwen/qwen2.5-coder-32b-instruct","name":"Qwen2.5-Coder-32B-Instruct (Cloudflare)","description":"Open coding-focused Qwen model for code generation, repair, and repository reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0.66,"output":1},"canonical_model_id":"alibaba/qwen2.5-coder-32b-instruct"},"cloudflare/@cf/mistralai/mistral-small-3.1-24b-instruct":{"id":"cloudflare/@cf/mistralai/mistral-small-3.1-24b-instruct","name":"Mistral Small 3.1 24B (Cloudflare)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":102400},"cost":{"input":0.351,"output":0.555},"canonical_model_id":"mistral/mistral-small-3.1-24b-instruct"},"cloudflare/@cf/zai-org/glm-4.7-flash":{"id":"cloudflare/@cf/zai-org/glm-4.7-flash","name":"GLM-4.7-Flash (Cloudflare)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.0605,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"cloudflare/@cf/openai/gpt-oss-20b":{"id":"cloudflare/@cf/openai/gpt-oss-20b","name":"GPT OSS 20B (Cloudflare)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.2,"output":0.3},"canonical_model_id":"openai/gpt-oss-20b"},"cloudflare/@cf/openai/gpt-oss-120b":{"id":"cloudflare/@cf/openai/gpt-oss-120b","name":"GPT OSS 120B (Cloudflare)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.35,"output":0.75},"canonical_model_id":"openai/gpt-oss-120b"},"cloudflare/@cf/aisingapore/gemma-sea-lion-v4-27b-it":{"id":"cloudflare/@cf/aisingapore/gemma-sea-lion-v4-27b-it","name":"Gemma-SEA-LION-v4-27B-IT (Cloudflare)","description":"Gemma 3 27B tuned by AI Singapore for Southeast Asian languages and instruction following","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.351,"output":0.555},"canonical_model_id":"aisingapore/gemma-sea-lion-v4-27b-it"},"greenference/gpt-oss-20b":{"id":"greenference/gpt-oss-20b","name":"GPT OSS 20B (Greenference)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.009,"output":0.045,"cache_read":0.0045},"canonical_model_id":"openai/gpt-oss-20b"},"greenference/deepseek-v4.1-flash":{"id":"greenference/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Greenference)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.22,"output":0.88,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"tensorx/deepseek/deepseek-v4-flash-0731":{"id":"tensorx/deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731 (TensorX)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.25,"output":0.3,"cache_read":0.0625},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"tensorx/deepseek/deepseek-r1-0528":{"id":"tensorx/deepseek/deepseek-r1-0528","name":"DeepSeek R1 0528 (TensorX)","description":"May 2025 DeepSeek R1 update with improved mathematics, coding, and reasoning","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":32768},"cost":{"input":0.66,"output":2.6,"cache_read":0.165},"canonical_model_id":"deepseek/deepseek-r1-0528"},"tensorx/deepseek/deepseek-v4-pro-0813":{"id":"tensorx/deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813 (TensorX)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":2,"output":4,"cache_read":0.5},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"tensorx/deepseek/deepseek-v4.1-flash-1017":{"id":"tensorx/deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (TensorX)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.5,"output":1.5,"cache_read":0.125},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"tensorx/deepseek/deepseek-v4.1-flash":{"id":"tensorx/deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (TensorX)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.5,"output":1.5,"cache_read":0.125},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"tensorx/moonshotai/kimi-k2.5":{"id":"tensorx/moonshotai/kimi-k2.5","name":"Kimi K2.5 (TensorX)","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":2.8,"cache_read":0.125},"canonical_model_id":"moonshotai/kimi-k2.5"},"minimax/MiniMax-M3":{"id":"minimax/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/MiniMax-M3.1-Flash":{"id":"minimax/MiniMax-M3.1-Flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":512000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"minimax/MiniMax-M2":{"id":"minimax/MiniMax-M2","name":"MiniMax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/MiniMax-M2.5":{"id":"minimax/MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"minimax/MiniMax-M2.7":{"id":"minimax/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/MiniMax-M2.1":{"id":"minimax/MiniMax-M2.1","name":"MiniMax-M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.1"},"azure/gpt-5.2-codex":{"id":"azure/gpt-5.2-codex","name":"GPT-5.2 Codex (Azure)","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"azure/gpt-5.1-codex-mini":{"id":"azure/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini (Azure)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"azure/gpt-5.1-codex":{"id":"azure/gpt-5.1-codex","name":"GPT-5.1 Codex (Azure)","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"azure/gpt-5.1-codex-max":{"id":"azure/gpt-5.1-codex-max","name":"GPT-5.1 Codex Max (Azure)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex-max"},"cohere/command-a-03-2025":{"id":"cohere/command-a-03-2025","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":288000,"output":8000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a-03-2025"},"cohere/command-r-08-2024":{"id":"cohere/command-r-08-2024","name":"Command R","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"cohere/command-r-08-2024"},"cohere/command-r7b-12-2024":{"id":"cohere/command-r7b-12-2024","name":"Command R7B","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-12-13","last_updated":"2024-12-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":132000,"output":4000},"cost":{"input":0.0375,"output":0.15},"canonical_model_id":"cohere/command-r7b-12-2024"},"cohere/c4ai-aya-expanse-32b":{"id":"cohere/c4ai-aya-expanse-32b","name":"Aya Expanse 32B","description":"Open multilingual model optimized for generation across 23 languages","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-10-24","last_updated":"2024-10-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"cohere/c4ai-aya-expanse-32b"},"cohere/command-r-plus-08-2024":{"id":"cohere/command-r-plus-08-2024","name":"Command R+","description":"Cohere's RAG workhorse for long-context enterprise search and tool use","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-r-plus-08-2024"},"infomaniak/mistralai/Ministral-3-14B-Instruct-2512":{"id":"infomaniak/mistralai/Ministral-3-14B-Instruct-2512","name":"Ministral 3 14B (Infomaniak)","description":"Open vision-language model for efficient local deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":100000,"output":262144},"cost":{"input":0.33675,"output":0.449},"canonical_model_id":"mistral/ministral-3-14b-instruct-2512"},"perplexityai/sonar-deep-research":{"id":"perplexityai/sonar-deep-research","name":"Sonar Deep Research","description":"Sonar search model for autonomous research and citation-backed long-form reports","family":"sonar","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-02-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":2,"output":8,"reasoning":3},"canonical_model_id":"perplexity/sonar-deep-research"},"perplexityai/sonar-pro":{"id":"perplexityai/sonar-pro","name":"Sonar Pro","description":"Deeper Sonar search model with broader retrieval and stronger synthesis","family":"sonar-pro","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":3,"output":15},"canonical_model_id":"perplexity/sonar-pro"},"perplexityai/sonar":{"id":"perplexityai/sonar","name":"Sonar","description":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","family":"sonar","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":127072,"output":4096},"cost":{"input":1,"output":1},"canonical_model_id":"perplexity/sonar"},"perplexityai/sonar-reasoning-pro":{"id":"perplexityai/sonar-reasoning-pro","name":"Sonar Reasoning Pro","description":"Web-grounded Sonar for multi-step research questions that need cited reasoning","family":"sonar-reasoning","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":2,"output":8},"canonical_model_id":"perplexity/sonar-reasoning-pro"},"deepinfra/nemotron-3-ultra-550b-a55b":{"id":"deepinfra/nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B (Deep Infra)","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"cost":{"input":0.5,"output":2.2,"cache_read":0.1},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"deepinfra/nemotron-3-super-120b-a12b":{"id":"deepinfra/nemotron-3-super-120b-a12b","name":"Nemotron 3 Super 120B A12B (Deep Infra)","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.085,"output":0.4},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"deepinfra/Gryphe/MythoMax-L2-13b":{"id":"deepinfra/Gryphe/MythoMax-L2-13b","name":"MythoMax L2 13B (Deep Infra)","description":"Llama 2 merge combining MythoLogic and Huginn for creative writing and roleplay","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2023-08-10","last_updated":"2023-08-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":3200},"cost":{"input":0.4,"output":0.4},"canonical_model_id":"gryphe/mythomax-l2-13b"},"deepinfra/microsoft/WizardLM-2-8x22B":{"id":"deepinfra/microsoft/WizardLM-2-8x22B","name":"WizardLM 2 8x22B (Deep Infra)","description":"Microsoft WizardLM instruction-tuned Mixtral MoE for multilingual dialogue and complex tasks","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-04-15","last_updated":"2024-04-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":8000},"cost":{"input":0.48,"output":0.48},"canonical_model_id":"microsoft/wizardlm-2-8x22b"},"deepinfra/nvidia/Nemotron-3-Nano-30B-A3B":{"id":"deepinfra/nvidia/Nemotron-3-Nano-30B-A3B","name":"Nemotron 3 Nano 30B A3B (Deep Infra)","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.05,"output":0.2,"cache_read":0.025},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"deepinfra/nvidia/Llama-3.1-Nemotron-70B-Instruct":{"id":"deepinfra/nvidia/Llama-3.1-Nemotron-70B-Instruct","name":"Llama 3.1 Nemotron 70B Instruct (Deep Infra)","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.6,"output":0.6},"canonical_model_id":"nvidia/llama-3.1-nemotron-70b-instruct"},"deepinfra/deepseek-ai/DeepSeek-V3":{"id":"deepinfra/deepseek-ai/DeepSeek-V3","name":"DeepSeek-V3 (Deep Infra)","description":"Open DeepSeek MoE chat model for coding, math, and general reasoning","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":8192},"cost":{"input":0.32,"output":0.89},"canonical_model_id":"deepseek/deepseek-v3"},"deepinfra/deepseek-ai/DeepSeek-R1-0528":{"id":"deepinfra/deepseek-ai/DeepSeek-R1-0528","name":"DeepSeek R1 0528 (Deep Infra)","description":"May 2025 DeepSeek R1 update with improved mathematics, coding, and reasoning","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.5,"output":2.15,"cache_read":0.35},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepinfra/deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepinfra/deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731 (Deep Infra)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.06,"output":0.18,"cache_read":0.015},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepinfra/deepseek-ai/DeepSeek-V4.1-Flash":{"id":"deepinfra/deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash (Deep Infra)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.2,"output":0.6,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepinfra/deepseek-ai/DeepSeek-V4.1-Flash-1017":{"id":"deepinfra/deepseek-ai/DeepSeek-V4.1-Flash-1017","name":"DeepSeek V4.1 Flash 1017 (Deep Infra)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.2,"output":0.6,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepinfra/deepseek-ai/DeepSeek-V3-0324":{"id":"deepinfra/deepseek-ai/DeepSeek-V3-0324","name":"DeepSeek V3 0324 (Deep Infra)","description":"March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.24,"output":0.9,"cache_read":0.135},"canonical_model_id":"deepseek/deepseek-v3-0324"},"deepinfra/deepseek-ai/DeepSeek-R1":{"id":"deepinfra/deepseek-ai/DeepSeek-R1","name":"DeepSeek-R1 (Deep Infra)","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.7,"output":2.4},"canonical_model_id":"deepseek/deepseek-r1"},"deepinfra/deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepinfra/deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813 (Deep Infra)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.3,"output":2.6,"cache_read":0.1},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepinfra/meta-llama/Llama-3.3-70B-Instruct":{"id":"deepinfra/meta-llama/Llama-3.3-70B-Instruct","name":"Llama-3.3-70B-Instruct (Deep Infra)","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.1,"output":0.32},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"deepinfra/meta-llama/Llama-Guard-3-8B":{"id":"deepinfra/meta-llama/Llama-Guard-3-8B","name":"Llama-Guard-3-8B (Deep Infra)","description":"Llama 3.1-based safety classifier for moderating prompts and model responses","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.055,"output":0.055},"canonical_model_id":"meta/llama-guard-3-8b"},"deepinfra/meta-llama/Llama-3.2-11B-Vision-Instruct":{"id":"deepinfra/meta-llama/Llama-3.2-11B-Vision-Instruct","name":"Llama-3.2-11B-Vision-Instruct (Deep Infra)","description":"Open multimodal Llama model for image understanding, captioning, and visual QA","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.345,"output":0.345},"canonical_model_id":"meta/llama-3.2-11b-vision-instruct"},"deepinfra/moonshotai/Kimi-K2-Instruct":{"id":"deepinfra/moonshotai/Kimi-K2-Instruct","name":"Kimi K2 Instruct (Deep Infra)","description":"Non-thinking Kimi K2 MoE instruction model for coding and tool-using agents","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":2},"canonical_model_id":"moonshotai/kimi-k2-instruct"},"deepinfra/moonshotai/Kimi-K2.5":{"id":"deepinfra/moonshotai/Kimi-K2.5","name":"Kimi K2.5 (Deep Infra)","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.45,"output":2.25,"cache_read":0.07},"canonical_model_id":"moonshotai/kimi-k2.5"},"deepinfra/stepfun-ai/Step-3.5-Flash":{"id":"deepinfra/stepfun-ai/Step-3.5-Flash","name":"Step 3.5 Flash (Deep Infra)","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.09,"output":0.3,"cache_read":0.02},"canonical_model_id":"stepfun/step-3.5-flash"},"deepinfra/stepfun-ai/Step-5-Preview":{"id":"deepinfra/stepfun-ai/Step-5-Preview","name":"Step 5 Preview (Deep Infra)","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1000000},"cost":{"input":1,"output":2.7,"cache_read":0.1},"canonical_model_id":"stepfun/step-5-preview"},"deepinfra/stepfun-ai/Step-3.7-Flash":{"id":"deepinfra/stepfun-ai/Step-3.7-Flash","name":"Step 3.7 Flash (Deep Infra)","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"deepinfra/thinkingmachines/Inkling":{"id":"deepinfra/thinkingmachines/Inkling","name":"Inkling (Deep Infra)","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":1048576},"cost":{"input":0.95,"output":4.05,"cache_read":0.16},"canonical_model_id":"thinkingmachines/inkling"},"deepinfra/thinkingmachines/Inkling-Small":{"id":"deepinfra/thinkingmachines/Inkling-Small","name":"Inkling Small (Deep Infra)","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":1048576},"cost":{"input":0.45,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"deepinfra/meta-models/Muse-Glimmer-30B":{"id":"deepinfra/meta-models/Muse-Glimmer-30B","name":"Muse Glimmer 30B (Deep Infra)","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"deepinfra/mistralai/Mistral-Small-24B-Instruct-2501":{"id":"deepinfra/mistralai/Mistral-Small-24B-Instruct-2501","name":"Mistral Small 3 (Deep Infra)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-01-30","last_updated":"2025-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.05,"output":0.08},"canonical_model_id":"mistral/mistral-small-24b-instruct-2501"},"deepinfra/Qwen/Qwen3-VL-235B-A22B-Instruct":{"id":"deepinfra/Qwen/Qwen3-VL-235B-A22B-Instruct","name":"Qwen3 VL 235B A22B Instruct (Deep Infra)","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":0.88,"cache_read":0.11},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"deepinfra/Qwen/Qwen3-Coder-480B-A35B-Instruct":{"id":"deepinfra/Qwen/Qwen3-Coder-480B-A35B-Instruct","name":"Qwen3-Coder 480B-A35B Instruct (Deep Infra)","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":1.6},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"deepinfra/Qwen/Qwen3-Next-80B-A3B-Instruct":{"id":"deepinfra/Qwen/Qwen3-Next-80B-A3B-Instruct","name":"Qwen3-Next 80B-A3B Instruct (Deep Infra)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.09,"output":1.1},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"deepinfra/Qwen/Qwen3-VL-30B-A3B-Instruct":{"id":"deepinfra/Qwen/Qwen3-VL-30B-A3B-Instruct","name":"Qwen3 VL 30B A3B Instruct (Deep Infra)","description":"Open vision-language MoE model for image and video understanding, documents, and visual agents","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"deepinfra/Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"deepinfra/Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B-A22B Instruct 2507 (Deep Infra)","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.09,"output":0.55},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"deepinfra/Qwen/Qwen3-Max":{"id":"deepinfra/Qwen/Qwen3-Max","name":"Qwen3 Max (Deep Infra)","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"cost":{"input":1.2,"output":6,"cache_read":0.24},"canonical_model_id":"alibaba/qwen3-max"},"deepinfra/ibm-granite/granite-4.2-8b":{"id":"deepinfra/ibm-granite/granite-4.2-8b","name":"Granite 4.2 8B (Deep Infra)","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"granite","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-25","last_updated":"2026-08-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.06,"output":0.25,"cache_read":0.015},"canonical_model_id":"ibm/granite-4.2-8b"},"deepinfra/zai-org/GLM-4.7-Flash":{"id":"deepinfra/zai-org/GLM-4.7-Flash","name":"GLM-4.7-Flash (Deep Infra)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.06,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flash"},"deepinfra/NousResearch/Hermes-3-Llama-3.1-405B":{"id":"deepinfra/NousResearch/Hermes-3-Llama-3.1-405B","name":"Hermes 3 405B Instruct (Deep Infra)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-16","last_updated":"2024-08-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":1,"output":1},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-405b"},"deepinfra/NousResearch/Hermes-3-Llama-3.1-70B":{"id":"deepinfra/NousResearch/Hermes-3-Llama-3.1-70B","name":"Hermes 3 70B Instruct (Deep Infra)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-18","last_updated":"2024-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.7,"output":0.7},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-70b"},"deepinfra/google/gemma-3-12b-it":{"id":"deepinfra/google/gemma-3-12b-it","name":"Gemma 3 12B IT (Deep Infra)","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.15},"canonical_model_id":"google/gemma-3-12b-it"},"deepinfra/google/gemma-3-27b-it":{"id":"deepinfra/google/gemma-3-27b-it","name":"Gemma 3 27B IT (Deep Infra)","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.08,"output":0.16},"canonical_model_id":"google/gemma-3-27b-it"},"deepinfra/google/gemma-3-4b-it":{"id":"deepinfra/google/gemma-3-4b-it","name":"Gemma 3 4B IT (Deep Infra)","description":"Open multimodal Gemma instruction model for efficient text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.1},"canonical_model_id":"google/gemma-3-4b-it"},"deepinfra/tencent/Hy4-preview":{"id":"deepinfra/tencent/Hy4-preview","name":"Hy4 preview (Deep Infra)","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"deepinfra/tencent/Hy3":{"id":"deepinfra/tencent/Hy3","name":"Hy3 (Deep Infra)","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"cost":{"input":0.13,"output":0.53,"cache_read":0.033},"canonical_model_id":"tencent/hy3"},"deepinfra/openai/gpt-oss-20b":{"id":"deepinfra/openai/gpt-oss-20b","name":"GPT OSS 20B (Deep Infra)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.03,"output":0.14},"canonical_model_id":"openai/gpt-oss-20b"},"deepinfra/openai/gpt-oss-120b":{"id":"deepinfra/openai/gpt-oss-120b","name":"GPT OSS 120B (Deep Infra)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.037,"output":0.17},"canonical_model_id":"openai/gpt-oss-120b"},"deepinfra/ByteDance/Seed-2.0-code":{"id":"deepinfra/ByteDance/Seed-2.0-code","name":"Seed 2.0 Code (Deep Infra)","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":131072},"cost":{"input":0.5,"output":3,"cache_read":0.1},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"deepinfra/ByteDance/Seed-2.0-mini":{"id":"deepinfra/ByteDance/Seed-2.0-mini","name":"Seed 2.0 Mini (Deep Infra)","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.1,"output":0.4,"cache_read":0.02},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"groq/openai/gpt-oss-safeguard-20b":{"id":"groq/openai/gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B (Groq)","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"groq/openai/gpt-oss-20b":{"id":"groq/openai/gpt-oss-20b","name":"GPT OSS 20B (Groq)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-oss-20b"},"groq/openai/gpt-oss-120b":{"id":"groq/openai/gpt-oss-120b","name":"GPT OSS 120B (Groq)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-oss-120b"},"vertex/gemini-4-argon@us":{"id":"vertex/gemini-4-argon@us","name":"Gemini 4 Argon (Vertex AI, US)","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":4,"output":20,"reasoning":20,"cache_read":0.2,"cache_write":0.375,"input_audio":4},"canonical_model_id":"google/gemini-4-argon"},"vertex/gemini-3.5-flash-lite@eu":{"id":"vertex/gemini-3.5-flash-lite@eu","name":"Gemini 3.5 Flash Lite (Vertex AI, EU)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333,"input_audio":0.3},"canonical_model_id":"google/gemini-3.5-flash-lite"},"vertex/gemini-pro-latest":{"id":"vertex/gemini-pro-latest","name":"Gemini Pro Latest (Gemini 4 Argon, Vertex AI)","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":4,"output":20,"reasoning":20,"cache_read":0.2,"cache_write":0.375,"input_audio":4},"canonical_model_id":"google/gemini-4-argon"},"vertex/gemini-3.1-flash-image":{"id":"vertex/gemini-3.1-flash-image","type":"image","name":"Nano Banana 2 (Vertex AI)","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["image","text"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image"},"vertex/gemini-3.1-flash-lite@us":{"id":"vertex/gemini-3.1-flash-lite@us","name":"Gemini 3.1 Flash Lite (Vertex AI, US)","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.083333,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"vertex/gemini-3.5-flash@us":{"id":"vertex/gemini-3.5-flash@us","name":"Gemini 3.5 Flash (Vertex AI, US)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"reasoning":9,"cache_read":0.15,"cache_write":0.083333,"input_audio":3},"canonical_model_id":"google/gemini-3.5-flash"},"vertex/gemini-3-flash-preview":{"id":"vertex/gemini-3-flash-preview","name":"Gemini 3 Flash Preview (Vertex AI)","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"reasoning":3,"cache_read":0.05,"cache_write":0.083333,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"vertex/gemini-3.5-flash@eu":{"id":"vertex/gemini-3.5-flash@eu","name":"Gemini 3.5 Flash (Vertex AI, EU)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"reasoning":9,"cache_read":0.15,"cache_write":0.083333,"input_audio":3},"canonical_model_id":"google/gemini-3.5-flash"},"vertex/gemini-3.8-flash@eu":{"id":"vertex/gemini-3.8-flash@eu","name":"Gemini 3.8 Flash (Vertex AI, EU)","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.8-flash"},"vertex/gemini-3.1-flash-lite":{"id":"vertex/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite (Vertex AI)","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.083333,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"vertex/gemini-4-argon@eu":{"id":"vertex/gemini-4-argon@eu","name":"Gemini 4 Argon (Vertex AI, EU)","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":4,"output":20,"reasoning":20,"cache_read":0.2,"cache_write":0.375,"input_audio":4},"canonical_model_id":"google/gemini-4-argon"},"vertex/gemini-3.7-flash":{"id":"vertex/gemini-3.7-flash","name":"Gemini 3.7 Flash (Vertex AI)","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.7-flash"},"vertex/gemini-3.1-pro-preview":{"id":"vertex/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview (Vertex AI)","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"input_audio":2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":0.25}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"vertex/gemini-3.7-flash@eu":{"id":"vertex/gemini-3.7-flash@eu","name":"Gemini 3.7 Flash (Vertex AI, EU)","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.7-flash"},"vertex/gemini-3.5-flash-lite@us":{"id":"vertex/gemini-3.5-flash-lite@us","name":"Gemini 3.5 Flash Lite (Vertex AI, US)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333,"input_audio":0.3},"canonical_model_id":"google/gemini-3.5-flash-lite"},"vertex/gemini-3.8-flash":{"id":"vertex/gemini-3.8-flash","name":"Gemini 3.8 Flash (Vertex AI)","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.8-flash"},"vertex/gemini-3.6-flash@eu":{"id":"vertex/gemini-3.6-flash@eu","name":"Gemini 3.6 Flash (Vertex AI, EU)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"vertex/gemini-3-pro-image":{"id":"vertex/gemini-3-pro-image","type":"image","name":"Nano Banana Pro (Vertex AI)","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"input_audio":2},"canonical_model_id":"google/gemini-3-pro-image"},"vertex/gemini-3.5-flash":{"id":"vertex/gemini-3.5-flash","name":"Gemini 3.5 Flash (Vertex AI)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"reasoning":9,"cache_read":0.15,"cache_write":0.083333,"input_audio":3},"canonical_model_id":"google/gemini-3.5-flash"},"vertex/gemini-3.1-flash-lite-image":{"id":"vertex/gemini-3.1-flash-lite-image","type":"image","name":"Nano Banana 2 Lite (Vertex AI)","description":"Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":4096},"cost":{"input":0.25,"output":1.5},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"vertex/gemini-nano-banana-2.1":{"id":"vertex/gemini-nano-banana-2.1","type":"image","name":"Nano Banana 2.1 (Vertex AI)","description":"Image model for prompt-driven generation, multi-turn editing, and visual design workflows","family":"gemini-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"google/gemini-nano-banana-2.1"},"vertex/gemini-4-argon":{"id":"vertex/gemini-4-argon","name":"Gemini 4 Argon (Vertex AI)","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":4,"output":20,"reasoning":20,"cache_read":0.2,"cache_write":0.375,"input_audio":4},"canonical_model_id":"google/gemini-4-argon"},"vertex/gemini-3.6-flash@us":{"id":"vertex/gemini-3.6-flash@us","name":"Gemini 3.6 Flash (Vertex AI, US)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"vertex/gemini-flash-latest":{"id":"vertex/gemini-flash-latest","name":"Gemini Flash Latest (Gemini 3.8 Flash, Vertex AI)","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.8-flash"},"vertex/gemini-3.8-flash@us":{"id":"vertex/gemini-3.8-flash@us","name":"Gemini 3.8 Flash (Vertex AI, US)","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.8-flash"},"vertex/gemini-3.5-flash-lite":{"id":"vertex/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite (Vertex AI)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333,"input_audio":0.3},"canonical_model_id":"google/gemini-3.5-flash-lite"},"vertex/gemini-3.6-flash":{"id":"vertex/gemini-3.6-flash","name":"Gemini 3.6 Flash (Vertex AI)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"vertex/gemini-2.5-flash-image":{"id":"vertex/gemini-2.5-flash-image","type":"image","name":"Nano Banana (Vertex AI)","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash-image"},"vertex/gemini-3.1-flash-lite@eu":{"id":"vertex/gemini-3.1-flash-lite@eu","name":"Gemini 3.1 Flash Lite (Vertex AI, EU)","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.083333,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"vertex/gemini-3.7-flash@us":{"id":"vertex/gemini-3.7-flash@us","name":"Gemini 3.7 Flash (Vertex AI, US)","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.7-flash"},"zai/glm-4.6v":{"id":"zai/glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.05},"canonical_model_id":"zhipuai/glm-4.6v"},"zai/glm-5.3-flash":{"id":"zai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai/glm-5.3":{"id":"zai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"zai/glm-4.7":{"id":"zai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"zai/glm-5.1":{"id":"zai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai/glm-5":{"id":"zai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"zai/glm-5v-turbo":{"id":"zai/glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"zai/glm-5-turbo":{"id":"zai/glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"zai/glm-4.6":{"id":"zai/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"zai/glm-5.2":{"id":"zai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"google/gemini-pro-latest":{"id":"google/gemini-pro-latest","name":"Gemini Pro Latest (Gemini 4 Argon)","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":4,"output":20,"reasoning":20,"cache_read":0.2,"cache_write":0.375,"input_audio":4},"canonical_model_id":"google/gemini-4-argon"},"google/gemini-3.1-flash-image":{"id":"google/gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["image","text"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image"},"google/gemini-3-pro-image-preview":{"id":"google/gemini-3-pro-image-preview","type":"image","name":"Nano Banana Pro Preview","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"input_audio":2},"canonical_model_id":"google/gemini-3-pro-image-preview"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"reasoning":3,"cache_read":0.05,"cache_write":0.083333,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-3.1-flash-image-preview":{"id":"google/gemini-3.1-flash-image-preview","type":"image","name":"Nano Banana 2 Preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["image","text"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":65536},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image-preview"},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"input_audio":2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.083333,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"input_audio":2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3-pro-image":{"id":"google/gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"input_audio":2},"canonical_model_id":"google/gemini-3-pro-image"},"google/deep-research-preview-04-2026":{"id":"google/deep-research-preview-04-2026","name":"Gemini Deep Research Preview","description":"Agentic model for autonomous multi-step research, synthesis, and cited reports","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12},"canonical_model_id":"google/deep-research-preview-04-2026"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"reasoning":9,"cache_read":0.15,"cache_write":0.083333,"input_audio":3},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-3.1-flash-lite-image":{"id":"google/gemini-3.1-flash-lite-image","type":"image","name":"Nano Banana 2 Lite","description":"Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":4096},"cost":{"input":0.25,"output":1.5},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"google/gemini-nano-banana-2.1":{"id":"google/gemini-nano-banana-2.1","type":"image","name":"Nano Banana 2.1","description":"Image model for prompt-driven generation, multi-turn editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["image","text"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"google/gemini-nano-banana-2.1"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":4,"output":20,"reasoning":20,"cache_read":0.2,"cache_write":0.375,"input_audio":4},"canonical_model_id":"google/gemini-4-argon"},"google/gemini-flash-latest":{"id":"google/gemini-flash-latest","name":"Gemini Flash Latest (Gemini 3.8 Flash)","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.8-flash"},"google/deep-research-max-preview-04-2026":{"id":"google/deep-research-max-preview-04-2026","name":"Deep Research Max Preview","description":"Maximum-comprehensiveness agentic researcher for multi-step investigation, synthesis, and cited reports","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12},"canonical_model_id":"google/deep-research-max-preview-04-2026"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333,"input_audio":0.3},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-robotics-er-2-preview":{"id":"google/gemini-robotics-er-2-preview","name":"Gemini Robotics-ER 2 Preview","description":"Embodied reasoning model for robotics: spatial understanding, video progress, and multi-robot orchestration","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":1,"output":5,"reasoning":5,"cache_read":0.1,"input_audio":1},"canonical_model_id":"google/gemini-robotics-er-2-preview"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.1-flash-lite-preview":{"id":"google/gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.083333,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"google/gemini-2.5-flash-image":{"id":"google/gemini-2.5-flash-image","type":"image","name":"Nano Banana","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["audio","image","text","video"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash-image"},"scaleway/deepseek-v4-flash-0731":{"id":"scaleway/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731 (Scaleway)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":384000},"cost":{"input":0.449,"output":0.898,"cache_read":0.0898},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"scaleway/llama-3.3-70b-instruct":{"id":"scaleway/llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct (Scaleway)","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":1.01025,"output":1.01025},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"scaleway/gemma-3-27b-it":{"id":"scaleway/gemma-3-27b-it","name":"Gemma 3 27B IT (Scaleway)","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":40000,"output":131072},"cost":{"input":0.287125,"output":0.57425},"canonical_model_id":"google/gemma-3-27b-it"},"scaleway/qwen3-235b-a22b-instruct-2507":{"id":"scaleway/qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B-A22B Instruct 2507 (Scaleway)","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":250000,"output":16384},"cost":{"input":0.841875,"output":2.525624},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"scaleway/qwen3-coder-30b-a3b-instruct":{"id":"scaleway/qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct (Scaleway)","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":65536},"cost":{"input":0.2245,"output":0.898},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"scaleway/gpt-oss-120b":{"id":"scaleway/gpt-oss-120b","name":"GPT OSS 120B (Scaleway)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.168375,"output":0.6735},"canonical_model_id":"openai/gpt-oss-120b"},"scaleway/deepseek-v4.1-flash":{"id":"scaleway/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Scaleway)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":384000},"cost":{"input":0.6735,"output":2.020499,"cache_read":0.1347},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 preview (Tencent)","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3 (Tencent)","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0.132,"output":0.528,"cache_read":0.033},"canonical_model_id":"tencent/hy3"},"tencent/deepseek-v4-pro-0813":{"id":"tencent/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813 (Tencent)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"tencent/deepseek-v4.1-flash":{"id":"tencent/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Tencent)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4":{"id":"openai/gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8191,"output":8192},"cost":{"input":30,"output":60},"canonical_model_id":"openai/gpt-4"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/o3-pro":{"id":"openai/o3-pro","name":"o3-pro","description":"High-effort o3 tier for difficult technical reasoning and careful answers","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","pdf","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":20,"output":80},"canonical_model_id":"openai/o3-pro"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5-turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-latest":{"id":"openai/gpt-latest","name":"GPT Latest (GPT-6 Astra)","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5-search-api":{"id":"openai/gpt-5-search-api","type":"chat","name":"GPT-5 Search API","description":"GPT-5 model trained for web search in the Chat Completions API","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-10-14","last_updated":"2025-10-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":100000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5-search-api"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-4o-2024-08-06":{"id":"openai/gpt-4o-2024-08-06","name":"GPT-4o (2024-08-06)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-08-06"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-mini-latest":{"id":"openai/gpt-mini-latest","name":"GPT Mini Latest (GPT-5.6 Luna)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2 Pro","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/gpt-pro-latest":{"id":"openai/gpt-pro-latest","name":"GPT Pro Latest (GPT-5.6 Sol)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai/o1-pro":{"id":"openai/o1-pro","name":"o1-pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2025-03-19","last_updated":"2025-03-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":150,"output":600},"canonical_model_id":"openai/o1-pro"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-4o-2024-11-20":{"id":"openai/gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"cerebras/gpt-oss-120b":{"id":"cerebras/gpt-oss-120b","name":"GPT OSS 120B (Cerebras)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.35,"output":0.75,"cache_read":0.35},"canonical_model_id":"openai/gpt-oss-120b"},"mistral/mistral-medium-2604":{"id":"mistral/mistral-medium-2604","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"mistral/mistral-medium-2604"},"mistral/mistral-medium-3":{"id":"mistral/mistral-medium-3","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":104857},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/mistral-medium-3"},"mistral/magistral-medium-latest":{"id":"mistral/magistral-medium-latest","name":"Magistral Medium (latest)","description":"Mistral reasoning model for transparent analysis, math, and complex decisions","family":"magistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-03-17","last_updated":"2025-03-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":16384},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"mistral/magistral-medium-latest"},"mistral/codestral-latest":{"id":"mistral/codestral-latest","name":"Codestral (latest)","description":"Mistral code model for completions, refactors, and developer IDE workflows","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-05-29","last_updated":"2025-01-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":4096},"cost":{"input":0.3,"output":0.9,"cache_read":0.03},"canonical_model_id":"mistral/codestral-latest"},"mistral/mistral-small-latest":{"id":"mistral/mistral-small-latest","name":"Mistral Small (latest)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"mistral/mistral-small-latest"},"mistral/ministral-8b-2512":{"id":"mistral/ministral-8b-2512","name":"Ministral 3 8B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.15,"output":0.15,"cache_read":0.015},"canonical_model_id":"mistral/ministral-8b-2512"},"mistral/devstral-2512":{"id":"mistral/devstral-2512","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/devstral-2512"},"mistral/ministral-3b-2512":{"id":"mistral/ministral-3b-2512","name":"Ministral 3 3B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":104857},"cost":{"input":0.1,"output":0.1,"cache_read":0.01},"canonical_model_id":"mistral/ministral-3b-2512"},"mistral/mistral-medium-3-5":{"id":"mistral/mistral-medium-3-5","name":"Mistral Medium 3.5","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":209715},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"mistral/mistral-medium-3-5"},"mistral/mistral-large-2512":{"id":"mistral/mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"mistral/mistral-large-2512"},"mistral/voxtral-small-latest":{"id":"mistral/voxtral-small-latest","name":"Voxtral Small (latest)","description":"Instruct model with native audio input for speech understanding and tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32000},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"mistral/voxtral-small-latest"},"mistral/mistral-small-2603":{"id":"mistral/mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"mistral/mistral-small-2603"},"mistral/mistral-medium-2505":{"id":"mistral/mistral-medium-2505","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/mistral-medium-2505"},"mistral/devstral-medium-latest":{"id":"mistral/devstral-medium-latest","name":"Devstral 2 (latest)","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/devstral-medium-latest"},"mistral/ministral-14b-2512":{"id":"mistral/ministral-14b-2512","name":"Ministral 3 14B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.2,"output":0.2,"cache_read":0.02},"canonical_model_id":"mistral/ministral-14b-2512"},"mistral/mistral-large-2407":{"id":"mistral/mistral-large-2407","name":"Mistral Large 2407","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-03-31","release_date":"2024-07-24","last_updated":"2024-07-24","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mistral-large-2407"},"mistral/codestral-2508":{"id":"mistral/codestral-2508","name":"Codestral 25.08","description":"Mistral code model for low-latency completion, fill-in-the-middle, and code generation","family":"codestral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":8192},"cost":{"input":0.3,"output":0.9,"cache_read":0.03},"canonical_model_id":"mistral/codestral-2508"},"mistral/mistral-large-latest":{"id":"mistral/mistral-large-latest","name":"Mistral Large (latest)","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mistral-large-latest"},"mistral/mistral-medium-latest":{"id":"mistral/mistral-medium-latest","name":"Mistral Medium (latest)","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"mistral/mistral-medium-latest"}}},"above":{"id":"above","env":["ABOVE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.above.dev/v1","name":"above.dev","doc":"https://above.dev/docs","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.165,"output":0.55,"cache_read":0.0319},"canonical_model_id":"zhipuai/glm-5.3-flash"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.726,"output":2.178,"reasoning":2.178,"cache_read":0.0242},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"glm-5.2-fast":{"id":"glm-5.2-fast","name":"GLM 5.2 Fast","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2.31,"output":7.26,"cache_read":0.231},"canonical_model_id":"zhipuai/glm-5.2"},"mimo-v2.6-pro-ultraspeed":{"id":"mimo-v2.6-pro-ultraspeed","name":"MiMo V2.6 Pro UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":5.0769,"output":10.1538,"cache_read":0.0423},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo V2.6 Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1692,"output":0.3385,"cache_read":0.0034},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0.165,"output":0.66,"reasoning":0.66,"cache_read":0.0033},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo V2.6 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.5077,"output":1.0154,"cache_read":0.0042},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"glm-5.2":{"id":"glm-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.54,"output":4.84,"cache_read":0.154},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.165,"output":0.66,"reasoning":0.66,"cache_read":0.0033},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen 3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":2.2,"output":6.6,"cache_read":0.275},"canonical_model_id":"alibaba/qwen3.8-max"}}},"deepinfra":{"id":"deepinfra","env":["DEEPINFRA_API_KEY"],"npm":"@ai-sdk/deepinfra","name":"Deep Infra","doc":"https://deepinfra.com/models","models":{"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50},"canonical_model_id":"anthropic/claude-fable-5"},"nvidia/Llama-3.3-Nemotron-Super-49B-v1.5":{"id":"nvidia/Llama-3.3-Nemotron-Super-49B-v1.5","name":"Llama 3.3 Nemotron Super 49B v1.5","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":0.4,"output":0.4},"canonical_model_id":"nvidia/llama-3.3-nemotron-super-49b-v1.5"},"nvidia/Nemotron-3-Nano-30B-A3B":{"id":"nvidia/Nemotron-3-Nano-30B-A3B","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.05,"output":0.2,"cache_read":0.025},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning":{"id":"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning","name":"Nemotron 3 Nano Omni 30B A3B Reasoning","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.2,"output":0.8},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"deepseek-ai/DeepSeek-V4-Flash-Vision-Exp":{"id":"deepseek-ai/DeepSeek-V4-Flash-Vision-Exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek-V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":8192},"cost":{"input":0.25,"output":0.95,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek-ai/DeepSeek-V3":{"id":"deepseek-ai/DeepSeek-V3","name":"DeepSeek-V3","description":"Open DeepSeek MoE chat model for coding, math, and general reasoning","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":8192},"cost":{"input":0.32,"output":0.89},"canonical_model_id":"deepseek/deepseek-v3"},"deepseek-ai/DeepSeek-R1-0528":{"id":"deepseek-ai/DeepSeek-R1-0528","name":"DeepSeek-R1-0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":64000},"cost":{"input":0.5,"output":2.15,"cache_read":0.35}},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.06,"output":0.18,"cache_read":0.015},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-V4.1-Flash":{"id":"deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.2,"output":0.6,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-ai/DeepSeek-V4.1-Flash-1017":{"id":"deepseek-ai/DeepSeek-V4.1-Flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.2,"output":0.6,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-ai/DeepSeek-V3-0324":{"id":"deepseek-ai/DeepSeek-V3-0324","name":"DeepSeek V3 0324","description":"March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"status":"deprecated","cost":{"input":0.24,"output":0.9,"cache_read":0.135},"canonical_model_id":"deepseek/deepseek-v3-0324"},"deepseek-ai/DeepSeek-V3.2":{"id":"deepseek-ai/DeepSeek-V3.2","name":"DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":64000},"cost":{"input":0.26,"output":0.38,"cache_read":0.13}},"deepseek-ai/DeepSeek-V4-Flash":{"id":"deepseek-ai/DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":16384},"cost":{"input":0.09,"output":0.18,"cache_read":0.018},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":16384},"cost":{"input":1.3,"output":2.6,"cache_read":0.1},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.3,"output":2.6,"cache_read":0.1},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"meta-llama/Llama-4-Scout-17B-16E-Instruct":{"id":"meta-llama/Llama-4-Scout-17B-16E-Instruct","name":"Llama 4 Scout 17B","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":327680,"output":16384},"cost":{"input":0.1,"output":0.3}},"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8":{"id":"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8","name":"Llama 4 Maverick 17B FP8","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":16384},"status":"deprecated","cost":{"input":0.2,"output":0.8}},"meta-llama/Llama-3.3-70B-Instruct-Turbo":{"id":"meta-llama/Llama-3.3-70B-Instruct-Turbo","name":"Llama 3.3 70B Turbo","description":"Compact Llama instruction model for fast chat and local deployment","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.1,"output":0.32}},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.75,"output":3.5,"cache_read":0.15}},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.68,"output":3.4,"cache_read":0.136},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":2.85,"output":14.25,"cache_read":0.285},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.5":{"id":"moonshotai/Kimi-K2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.45,"output":2.25,"cache_read":0.07}},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":512000},"cost":{"input":0.28,"output":1.1,"cache_read":0.056},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-06","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"status":"deprecated","cost":{"input":0.15,"output":1.15,"cache_read":0.03}},"MiniMaxAI/MiniMax-M2.7":{"id":"MiniMaxAI/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"status":"deprecated","cost":{"input":0.25,"output":1,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M2.7"},"XiaomiMiMo/MiMo-V2.5-Pro":{"id":"XiaomiMiMo/MiMo-V2.5-Pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":16384},"status":"deprecated","cost":{"input":1,"output":3,"cache_read":0.2},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"XiaomiMiMo/MiMo-V2.5":{"id":"XiaomiMiMo/MiMo-V2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.5"},"XiaomiMiMo/MiMo-V2.6-Flash":{"id":"XiaomiMiMo/MiMo-V2.6-Flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"XiaomiMiMo/MiMo-V2.6-Pro":{"id":"XiaomiMiMo/MiMo-V2.6-Pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.43,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"stepfun-ai/Step-5-Preview":{"id":"stepfun-ai/Step-5-Preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1000000},"cost":{"input":1,"output":2.7,"cache_read":0.1},"canonical_model_id":"stepfun/step-5-preview"},"stepfun-ai/Step-3.7-Flash":{"id":"stepfun-ai/Step-3.7-Flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"status":"deprecated","cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"thinkingmachines/Inkling":{"id":"thinkingmachines/Inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":1048576},"cost":{"input":0.95,"output":4.05,"cache_read":0.16},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/Inkling-Small":{"id":"thinkingmachines/Inkling-Small","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":1048576},"cost":{"input":0.45,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"mistralai/Mistral-Small-24B-Instruct-2501":{"id":"mistralai/Mistral-Small-24B-Instruct-2501","name":"Mistral Small 3","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-01-30","last_updated":"2025-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.05,"output":0.08},"canonical_model_id":"mistral/mistral-small-24b-instruct-2501"},"Qwen/Qwen3-VL-235B-A22B-Instruct":{"id":"Qwen/Qwen3-VL-235B-A22B-Instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":0.88,"cache_read":0.11},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"Qwen/Qwen3.5-35B-A3B":{"id":"Qwen/Qwen3.5-35B-A3B","name":"Qwen 3.5 35B A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-01","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":81920},"cost":{"input":0.14,"output":1,"cache_read":0.05}},"Qwen/Qwen3.6-35B-A3B":{"id":"Qwen/Qwen3.6-35B-A3B","name":"Qwen3.6 35B A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":81920},"cost":{"input":0.1,"output":0.95,"cache_read":0.1}},"Qwen/Qwen3.8-Flash":{"id":"Qwen/Qwen3.8-Flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.113,"output":0.382,"cache_read":0.0141},"canonical_model_id":"alibaba/qwen3.8-flash"},"Qwen/Qwen3-Next-80B-A3B-Instruct":{"id":"Qwen/Qwen3-Next-80B-A3B-Instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.09,"output":1.1},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"Qwen/Qwen3-VL-30B-A3B-Instruct":{"id":"Qwen/Qwen3-VL-30B-A3B-Instruct","name":"Qwen3 VL 30B A3B Instruct","description":"Open vision-language MoE model for image and video understanding, documents, and visual agents","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo":{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo","name":"Qwen3 Coder 480B A35B Instruct Turbo","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":66536},"cost":{"input":0.3,"output":1,"cache_read":0.1}},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":2.5,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3.5-122B-A10B":{"id":"Qwen/Qwen3.5-122B-A10B","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.29,"output":2.4},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"Qwen/Qwen3.8-Max":{"id":"Qwen/Qwen3.8-Max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":131072},"cost":{"input":1.65,"output":4.951,"cache_read":0.206},"canonical_model_id":"alibaba/qwen3.8-max"},"Qwen/Qwen3.6-27B":{"id":"Qwen/Qwen3.6-27B","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.32,"output":3.2},"canonical_model_id":"alibaba/qwen3.6-27b"},"Qwen/Qwen3.7-Max":{"id":"Qwen/Qwen3.7-Max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.5,"tiers":[{"input":5,"output":15,"cache_read":1,"tier":{"type":"context","size":32000}},{"input":6.25,"output":18.5,"cache_read":1.25,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.7-max"},"Qwen/Qwen3.5-9B":{"id":"Qwen/Qwen3.5-9B","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.1,"output":0.15},"canonical_model_id":"alibaba/qwen3.5-9b"},"Qwen/Qwen3-32B":{"id":"Qwen/Qwen3-32B","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.08,"output":0.28},"canonical_model_id":"alibaba/qwen3-32b"},"Qwen/Qwen3-30B-A3B":{"id":"Qwen/Qwen3-30B-A3B","name":"Qwen3 30B A3B","description":"Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.5},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"Qwen/Qwen3.5-27B":{"id":"Qwen/Qwen3.5-27B","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.26,"output":2.6},"canonical_model_id":"alibaba/qwen3.5-27b"},"Qwen/Qwen3.5-397B-A17B":{"id":"Qwen/Qwen3.5-397B-A17B","name":"Qwen 3.5 397B A17B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-01","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":81920},"cost":{"input":0.45,"output":3,"cache_read":0.22}},"Qwen/Qwen3.8-2.4T-A95B":{"id":"Qwen/Qwen3.8-2.4T-A95B","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"status":"deprecated","cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B-A22B Instruct 2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.09,"output":0.55},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"Qwen/Qwen3-Max":{"id":"Qwen/Qwen3-Max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"status":"deprecated","cost":{"input":1.2,"output":6,"cache_read":0.24,"tiers":[{"input":2.4,"output":12,"cache_read":0.48,"tier":{"type":"context","size":32000}},{"input":3,"output":15,"cache_read":0.6,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3-max"},"zai-org/GLM-4.7":{"id":"zai-org/GLM-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"cost":{"input":0.4,"output":1.75,"cache_read":0.08}},"zai-org/GLM-4.6":{"id":"zai-org/GLM-4.6","name":"GLM-4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.5,"output":2,"cache_read":0.1}},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"status":"deprecated","cost":{"input":1.05,"output":3.5,"cache_read":0.205}},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.9,"output":4,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5":{"id":"zai-org/GLM-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"status":"deprecated","cost":{"input":0.6,"output":2.08,"cache_read":0.12}},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.75,"output":2.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-4.7-Flash":{"id":"zai-org/GLM-4.7-Flash","name":"GLM-4.7-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"status":"deprecated","cost":{"input":0.06,"output":0.4,"cache_read":0.01}},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.15},"canonical_model_id":"google/gemma-3-12b-it"},"google/gemma-4-E4B-it":{"id":"google/gemma-4-E4B-it","name":"Gemma 4 E4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.02,"output":0.1},"canonical_model_id":"google/gemma-4-E4B-it"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.08,"output":0.16},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemma-4-26B-A4B-it":{"id":"google/gemma-4-26B-A4B-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.07,"output":0.34},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemma-3-4b-it":{"id":"google/gemma-3-4b-it","name":"Gemma 3 4B IT","description":"Open multimodal Gemma instruction model for efficient text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.1},"canonical_model_id":"google/gemma-3-4b-it"},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"tencent/Hy4-preview":{"id":"tencent/Hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"tencent/Hy3":{"id":"tencent/Hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":128000},"status":"deprecated","cost":{"input":0.13,"output":0.53,"cache_read":0.033},"canonical_model_id":"tencent/hy3"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.03,"output":0.14}},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.037,"output":0.17}},"ByteDance/Seed-2.0-code":{"id":"ByteDance/Seed-2.0-code","name":"Seed 2.0 Code","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":131072},"cost":{"input":0.5,"output":3,"cache_read":0.1,"tiers":[{"input":1,"output":6,"cache_read":0.2,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"ByteDance/Seed-2.0-mini":{"id":"ByteDance/Seed-2.0-mini","name":"Seed 2.0 Mini","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.1,"output":0.4,"cache_read":0.02,"tiers":[{"input":0.2,"output":0.8,"cache_read":0.2,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"ByteDance/Seed-2.0-pro":{"id":"ByteDance/Seed-2.0-pro","name":"Seed 2.0 Pro","description":"Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.5,"output":3,"cache_read":0.1,"tiers":[{"input":1,"output":6,"cache_read":0.2,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-pro"}}},"alibaba-token-plan":{"id":"alibaba-token-plan","env":["ALIBABA_TOKEN_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://token-plan.ap-southeast-1.maas.aliyuncs.com/compatible-mode/v1","name":"Alibaba Token Plan","doc":"https://www.alibabacloud.com/help/en/model-studio/token-plan-overview","models":{"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":131072}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":131072}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen-image-2.0-pro":{"id":"qwen-image-2.0-pro","type":"image","name":"Qwen Image 2.0 Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"qwen3.8-max-preview":{"id":"qwen3.8-max-preview","name":"Qwen3.8 Max Preview","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-19","last_updated":"2026-07-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.8-flash"},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"input":196601,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"minimax/MiniMax-M2.5"},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.7-plus"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"happyhorse-1.1-r2v":{"id":"happyhorse-1.1-r2v","type":"video","name":"HappyHorse 1.1 Reference-to-Video","description":"Video model for reference-guided video generation","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-17","last_updated":"2026-07-17","modalities":{"input":["image","text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"wan2.7-image":{"id":"wan2.7-image","type":"image","name":"Wan2.7 Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"wan2.7-image-pro":{"id":"wan2.7-image-pro","type":"image","name":"Wan2.7 Image Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"glm-5":{"id":"glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":16384},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-03","last_updated":"2025-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"status":"deprecated","cost":{"input":0,"output":0}},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"happyhorse-1.1-t2v":{"id":"happyhorse-1.1-t2v","type":"video","name":"HappyHorse 1.1 Text-to-Video","description":"Video model for prompt-driven text-to-video generation","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-17","last_updated":"2026-07-17","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.7-max"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.5"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen-image-2.0":{"id":"qwen-image-2.0","type":"image","name":"Qwen Image 2.0","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"happyhorse-1.1-i2v":{"id":"happyhorse-1.1-i2v","type":"video","name":"HappyHorse 1.1 Image-to-Video","description":"Video model for image-to-video generation","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-17","last_updated":"2026-07-17","modalities":{"input":["image","text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}}}},"groq":{"id":"groq","env":["GROQ_API_KEY"],"npm":"@ai-sdk/groq","name":"Groq","doc":"https://console.groq.com/docs/models","models":{"whisper-large-v3":{"id":"whisper-large-v3","name":"Whisper","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2023-09-01","last_updated":"2025-09-05","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":0,"output":0}},"whisper-large-v3-turbo":{"id":"whisper-large-v3-turbo","name":"Whisper Large V3 Turbo","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-10-01","last_updated":"2024-10-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":0,"output":0}},"llama-3.1-8b-instant":{"id":"llama-3.1-8b-instant","name":"Llama 3.1 8B","description":"Compact Llama instruction model for fast chat and local deployment","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.08}},"llama-3.3-70b-versatile":{"id":"llama-3.3-70b-versatile","name":"Llama 3.3 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.59,"output":0.79}},"allam-2-7b":{"id":"allam-2-7b","name":"ALLaM-2-7b","description":"ALLaM-2-7b instruction tuned model by SDAIA","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-01-23","last_updated":"2025-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"sdaia/allam-2-7b"},"canopylabs/orpheus-arabic-saudi":{"id":"canopylabs/orpheus-arabic-saudi","name":"Canopy Labs Orpheus Arabic Saudi","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"canopylabs","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":4000,"output":50000},"status":"beta"},"canopylabs/orpheus-v1-english":{"id":"canopylabs/orpheus-v1-english","name":"Canopy Labs Orpheus V1 English","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"canopylabs","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-19","last_updated":"2025-12-19","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":4000,"output":50000},"status":"beta"},"meta-llama/llama-prompt-guard-2-86m":{"id":"meta-llama/llama-prompt-guard-2-86m","name":"Prompt Guard 2 86M","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-05-29","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":512},"status":"beta","cost":{"input":0.04,"output":0.04}},"meta-llama/llama-prompt-guard-2-22m":{"id":"meta-llama/llama-prompt-guard-2-22m","name":"Llama Prompt Guard 2 22M","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-05-29","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":512},"status":"beta","cost":{"input":0.03,"output":0.03}},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","default","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131042,"output":16384},"cost":{"input":0.8,"output":4},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","default"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.6,"output":3,"cache_read":0.3},"canonical_model_id":"alibaba/qwen3.6-27b"},"groq/compound-mini":{"id":"groq/compound-mini","name":"Compound Mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"groq","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-09-04","last_updated":"2025-09-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192}},"groq/compound":{"id":"groq/compound","name":"Compound","description":"General-purpose chat model for instruction following, writing, and analysis","family":"groq","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-09-04","last_updated":"2025-09-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192}},"openai/gpt-oss-safeguard-20b":{"id":"openai/gpt-oss-safeguard-20b","name":"Safety GPT OSS 20B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2026-06-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"status":"beta","cost":{"input":0.075,"output":0.3}},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.075,"output":0.3,"cache_read":0.0375}},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-10-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-oss-120b"}}},"siliconflow":{"id":"siliconflow","env":["SILICONFLOW_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.siliconflow.com/v1","name":"SiliconFlow","doc":"https://cloud.siliconflow.com/models","models":{"nex-agi/Nex-N2-Pro":{"id":"nex-agi/Nex-N2-Pro","name":"Nex-N2-Pro","description":"Open agentic MoE model (397B total, 17B active) for coding, tool use, and research workflows","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.5,"output":2.5,"cache_read":0.25},"canonical_model_id":"nex-agi/nex-n2-pro"},"inclusionAI/Ling-flash-2.0":{"id":"inclusionAI/Ling-flash-2.0","name":"inclusionAI/Ling-flash-2.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-18","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":131000},"cost":{"input":0.14,"output":0.57}},"meituan-longcat/LongCat-2.0":{"id":"meituan-longcat/LongCat-2.0","name":"LongCat-2.0","description":"Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1049000,"output":131072},"cost":{"input":0.75,"output":2.95,"cache_read":0.015},"canonical_model_id":"meituan/longcat-2.0"},"deepseek-ai/DeepSeek-V4-Flash-Vision-Exp":{"id":"deepseek-ai/DeepSeek-V4-Flash-Vision-Exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.44,"output":1.32,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"deepseek-ai/DeepSeek-V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-25","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":164000,"output":164000},"cost":{"input":0.27,"output":1}},"deepseek-ai/DeepSeek-V3.2-Exp":{"id":"deepseek-ai/DeepSeek-V3.2-Exp","name":"deepseek-ai/DeepSeek-V3.2-Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-10","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":164000,"output":164000},"cost":{"input":0.27,"output":0.41}},"deepseek-ai/DeepSeek-V3":{"id":"deepseek-ai/DeepSeek-V3","name":"deepseek-ai/DeepSeek-V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-12-26","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":164000,"output":164000},"cost":{"input":0.25,"output":1}},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.22,"output":0.66,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-V3.2":{"id":"deepseek-ai/DeepSeek-V3.2","name":"deepseek-ai/DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-03","last_updated":"2025-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":164000,"output":164000},"cost":{"input":0.27,"output":0.42,"cache_read":0.135}},"deepseek-ai/DeepSeek-V4-Flash":{"id":"deepseek-ai/DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.13,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.50162,"output":3.135,"cache_read":0.135},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/DeepSeek-V3.1-Terminus":{"id":"deepseek-ai/DeepSeek-V3.1-Terminus","name":"deepseek-ai/DeepSeek-V3.1-Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-29","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":164000,"output":164000},"cost":{"input":0.27,"output":1}},"deepseek-ai/DeepSeek-R1":{"id":"deepseek-ai/DeepSeek-R1","name":"deepseek-ai/DeepSeek-R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":164000,"output":164000},"cost":{"input":0.5,"output":2.18}},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"moonshotai/Kimi-K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-21","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.77,"output":3.4,"cache_read":0.14}},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.85916,"output":3.8,"cache_read":0.17993},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262000},"cost":{"input":2.7,"output":13.5,"cache_read":0.27},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.5":{"id":"moonshotai/Kimi-K2.5","name":"moonshotai/Kimi-K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.45,"output":2.25,"cache_read":0.07}},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMaxAI/MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-06-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":197000,"output":131000},"cost":{"input":0.3,"output":1.2,"cache_read":0.03}},"stepfun-ai/Step-3.5-Flash":{"id":"stepfun-ai/Step-3.5-Flash","name":"stepfun-ai/Step-3.5-Flash","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","family":"step","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.1,"output":0.3}},"ByteDance-Seed/Seed-OSS-36B-Instruct":{"id":"ByteDance-Seed/Seed-OSS-36B-Instruct","name":"ByteDance-Seed/Seed-OSS-36B-Instruct","description":"Tool-capable chat model for instruction following and agentic application workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-04","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.21,"output":0.57}},"Qwen/Qwen3-VL-8B-Instruct":{"id":"Qwen/Qwen3-VL-8B-Instruct","name":"Qwen/Qwen3-VL-8B-Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.18,"output":0.68}},"Qwen/Qwen3.5-35B-A3B":{"id":"Qwen/Qwen3.5-35B-A3B","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.24,"output":1.8},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"Qwen/Qwen3.6-35B-A3B":{"id":"Qwen/Qwen3.6-35B-A3B","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.2,"output":1.6},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen/Qwen3-VL-30B-A3B-Thinking":{"id":"Qwen/Qwen3-VL-30B-A3B-Thinking","name":"Qwen/Qwen3-VL-30B-A3B-Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-11","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.29,"output":1}},"Qwen/Qwen3-VL-32B-Thinking":{"id":"Qwen/Qwen3-VL-32B-Thinking","name":"Qwen/Qwen3-VL-32B-Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-21","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.2,"output":1.5}},"Qwen/Qwen2.5-72B-Instruct":{"id":"Qwen/Qwen2.5-72B-Instruct","name":"Qwen/Qwen2.5-72B-Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-09-18","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":33000,"output":4000},"cost":{"input":0.59,"output":0.59}},"Qwen/Qwen3-Coder-480B-A35B-Instruct":{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct","name":"Qwen/Qwen3-Coder-480B-A35B-Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-31","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.25,"output":1}},"Qwen/Qwen3-VL-30B-A3B-Instruct":{"id":"Qwen/Qwen3-VL-30B-A3B-Instruct","name":"Qwen/Qwen3-VL-30B-A3B-Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-05","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.29,"output":1}},"Qwen/Qwen3-Coder-30B-A3B-Instruct":{"id":"Qwen/Qwen3-Coder-30B-A3B-Instruct","name":"Qwen/Qwen3-Coder-30B-A3B-Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-01","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.07,"output":0.28}},"Qwen/Qwen3.5-122B-A10B":{"id":"Qwen/Qwen3.5-122B-A10B","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.26,"output":2.08},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"Qwen/Qwen2.5-7B-Instruct":{"id":"Qwen/Qwen2.5-7B-Instruct","name":"Qwen/Qwen2.5-7B-Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-09-18","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":33000,"output":4000},"cost":{"input":0.05,"output":0.05}},"Qwen/Qwen3-VL-32B-Instruct":{"id":"Qwen/Qwen3-VL-32B-Instruct","name":"Qwen/Qwen3-VL-32B-Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-21","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.2,"output":0.6}},"Qwen/Qwen3.6-27B":{"id":"Qwen/Qwen3.6-27B","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":3.2},"canonical_model_id":"alibaba/qwen3.6-27b"},"Qwen/Qwen3.5-9B":{"id":"Qwen/Qwen3.5-9B","name":"Qwen/Qwen3.5-9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-03","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.15}},"Qwen/Qwen3-32B":{"id":"Qwen/Qwen3-32B","name":"Qwen/Qwen3-32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-30","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":131000},"cost":{"input":0.14,"output":0.57}},"Qwen/Qwen3-8B":{"id":"Qwen/Qwen3-8B","name":"Qwen/Qwen3-8B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-30","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":131000},"cost":{"input":0.06,"output":0.06}},"Qwen/Qwen3-30B-A3B-Instruct-2507":{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507","name":"Qwen/Qwen3-30B-A3B-Instruct-2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":262000},"cost":{"input":0.09,"output":0.3}},"Qwen/Qwen3.5-27B":{"id":"Qwen/Qwen3.5-27B","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.25,"output":2},"canonical_model_id":"alibaba/qwen3.5-27b"},"Qwen/Qwen3.5-397B-A17B":{"id":"Qwen/Qwen3.5-397B-A17B","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.39,"output":2.34},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"Qwen/Qwen3-14B":{"id":"Qwen/Qwen3-14B","name":"Qwen/Qwen3-14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-30","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":131000},"cost":{"input":0.07,"output":0.28}},"Qwen/Qwen3.8-2.4T-A95B":{"id":"Qwen/Qwen3.8-2.4T-A95B","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1049000,"output":131000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"zai-org/GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-04-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":205000,"output":131000},"cost":{"input":1.19,"output":3.74,"cache_read":0.6,"cache_write":0}},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1049000,"output":262000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5":{"id":"zai-org/GLM-5","name":"zai-org/GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-06-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":205000,"output":205000},"cost":{"input":0.95,"output":2.55,"cache_read":0.2}},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1049000,"output":262000},"cost":{"input":1.302,"output":4.092,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5V-Turbo":{"id":"zai-org/GLM-5V-Turbo","name":"zai-org/GLM-5V-Turbo","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24,"cache_write":0}},"zai-org/GLM-4.5-Air":{"id":"zai-org/GLM-4.5-Air","name":"zai-org/GLM-4.5-Air","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-air","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":131000},"cost":{"input":0.14,"output":0.86}},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1049000,"output":262000},"cost":{"input":0.15,"output":0.5,"cache_read":0.03,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"google/gemma-4-12B-it":{"id":"google/gemma-4-12B-it","name":"Gemma 4 12B IT","description":"Compact Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"google/gemma-4-12b-it"},"google/gemma-4-26B-A4B-it":{"id":"google/gemma-4-26B-A4B-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.12,"output":0.4},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.13,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"tencent/Hy3":{"id":"tencent/Hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":262144},"cost":{"input":0.132,"output":0.528,"cache_read":0.033},"canonical_model_id":"tencent/hy3"},"tencent/Hunyuan-A13B-Instruct":{"id":"tencent/Hunyuan-A13B-Instruct","name":"tencent/Hunyuan-A13B-Instruct","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-06-30","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":131000},"cost":{"input":0.14,"output":0.57}},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"openai/gpt-oss-20b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-13","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":8000},"cost":{"input":0.04,"output":0.18}},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"openai/gpt-oss-120b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-13","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":8000},"cost":{"input":0.05,"output":0.45}}}},"prime":{"id":"prime","env":["PRIME_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.pinference.ai/api/v1","name":"Prime Inference","doc":"https://docs.primeintellect.ai/inference/overview","models":{"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50},"canonical_model_id":"anthropic/claude-fable-5-1"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro 0423","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.91,"output":3.83},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.44,"output":1.32},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash 0423","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943717},"cost":{"input":0.44,"output":1.32},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.006,"cache_write":0.3},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.3},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03,"cache_write":0.15},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":1.4},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":1.54,"output":4.84},"canonical_model_id":"zhipuai/glm-5.2"},"meta/muse-spark-open":{"id":"meta/muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25},"canonical_model_id":"meta/muse-spark-open"},"meta-llama/llama-3.3-70b-instruct":{"id":"meta-llama/llama-3.3-70b-instruct","name":"Llama 3.3 70B Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":1.04,"output":2.253},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"qwen/qwen3-vl-30b-a3b-instruct":{"id":"qwen/qwen3-vl-30b-a3b-instruct","name":"Qwen3 VL 30B A3B Instruct","description":"Open vision-language MoE model for image and video understanding, documents, and visual agents","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.2,"output":0.7},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"qwen/qwen3-vl-235b-a22b-instruct":{"id":"qwen/qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.3,"output":1.5},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0.99,"output":3.2},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3-coder-next":{"id":"qwen/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.3,"output":1.5},"canonical_model_id":"alibaba/qwen3-coder-next"},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 0905","description":"September 2025 Kimi K2 instruction update with a 256K context and improved agentic coding","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.6,"output":2.5},"canonical_model_id":"moonshotai/kimi-k2-0905"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":3.45,"output":17.25},"canonical_model_id":"moonshotai/kimi-k3"},"mistralai/mistral-nemo":{"id":"mistralai/mistral-nemo","name":"Mistral Nemo","description":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.165,"output":0.17},"canonical_model_id":"mistral/mistral-nemo"},"mistralai/mixtral-8x22b-instruct":{"id":"mistralai/mixtral-8x22b-instruct","name":"Mixtral 8x22B Instruct","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-01-31","release_date":"2024-04-17","last_updated":"2024-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":52428},"cost":{"input":2.2,"output":6.6},"canonical_model_id":"mistral/mixtral-8x22b-instruct"},"mistralai/mistral-large-2512":{"id":"mistralai/mistral-large-2512","name":"Mistral Large 3 2512","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/mistral-large-2512"},"mistralai/mistral-small-2603":{"id":"mistralai/mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.165,"output":0.66},"canonical_model_id":"mistral/mistral-small-2603"},"x-ai/grok-4.7":{"id":"x-ai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":1.6,"output":4.8},"canonical_model_id":"xai/grok-4.7"},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6},"canonical_model_id":"xai/grok-4.6"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.119,"output":0.3},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.35,"output":6.75},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["audio","image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["audio","image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":2,"output":10},"canonical_model_id":"google/gemini-4-argon"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"google/gemini-3.6-flash"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o-mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":5,"output":30},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-5.6-luna-pro":{"id":"openai/gpt-5.6-luna-pro","name":"GPT-5.6 Luna Pro","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":1,"output":6},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 Nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 Mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.75,"output":14},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0.1,"output":0.5},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.25,"output":10},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2.5,"output":15},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.25,"output":2},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0.2,"output":1.2},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2.5,"output":15},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"gpt-oss-20b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.07,"output":0.3},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-5.6-terra-pro":{"id":"openai/gpt-5.6-terra-pro","name":"GPT-5.6 Terra Pro","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2.5,"output":15},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":10,"output":50},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":5,"output":30},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.6-sol-pro":{"id":"openai/gpt-5.6-sol-pro","name":"GPT-5.6 Sol Pro","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":5,"output":30},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.2,"output":1.25},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"gpt-oss-120b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.35,"output":0.75},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.25,"output":10},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.75,"output":4.5},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.05,"output":0.4},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.75,"output":14},"canonical_model_id":"openai/gpt-5.3-codex"}}},"cloudferro-sherlock":{"id":"cloudferro-sherlock","env":["CLOUDFERRO_SHERLOCK_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api-sherlock.cloudferro.com/openai/v1/","name":"CloudFerro Sherlock","doc":"https://docs.sherlock.cloudferro.com/","models":{"meta-llama/Llama-3.3-70B-Instruct":{"id":"meta-llama/Llama-3.3-70B-Instruct","name":"Llama 3.3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10-09","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":70000,"output":70000},"cost":{"input":2.92,"output":2.92}},"speakleash/Bielik-11B-v3.0-Instruct":{"id":"speakleash/Bielik-11B-v3.0-Instruct","name":"Bielik 11B v3.0 Instruct","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000},"cost":{"input":0.67,"output":0.67}},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196000,"input":180000,"output":16000},"cost":{"input":0.3,"output":1.2}},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.34,"output":4.48},"canonical_model_id":"zhipuai/glm-5.3"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"OpenAI GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":2.92,"output":2.92},"canonical_model_id":"openai/gpt-oss-120b"}}},"inceptron":{"id":"inceptron","env":["INCEPTRON_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.inceptron.io/v1","name":"Inceptron","doc":"https://docs.inceptron.io","models":{"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.13,"output":0.28,"cache_read":0.03,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.53,"output":3.39,"cache_read":0.17,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.66,"output":3.4,"cache_read":0.18,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.71,"output":2.35,"cache_read":0.12,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"}}},"amd":{"id":"amd","env":["AMD_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://developer.amd.com.cn/radeon/api/v1","name":"AMD","doc":"https://developer.amd.com.cn/radeon/tokenfactory","models":{"Qwen3.8-Flash-Next":{"id":"Qwen3.8-Flash-Next","name":"Qwen3.8 Flash Next","description":"Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-flash-next"},"DeepSeek-V4-Flash-Vision-Exp":{"id":"DeepSeek-V4-Flash-Vision-Exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"MiniCPM5-2B":{"id":"MiniCPM5-2B","name":"MiniCPM5-2B","description":"Dense 2B-class open-source model for on-device and resource-constrained use, with native long-context support, tool calling, and agentic tasks","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-09-06","last_updated":"2026-09-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.124,"output":0.7425,"cache_read":0.124},"canonical_model_id":"openbmb/minicpm5-2b"},"Qwen3.8-27B":{"id":"Qwen3.8-27B","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.8-27b"},"DeepSeek-V4.1-Flash":{"id":"DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"DeepSeek-V4-Flash":{"id":"DeepSeek-V4-Flash","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"}}},"google-vertex-anthropic":{"id":"google-vertex-anthropic","env":["GOOGLE_VERTEX_PROJECT","GOOGLE_VERTEX_LOCATION","GOOGLE_APPLICATION_CREDENTIALS"],"npm":"@ai-sdk/google-vertex/anthropic","name":"Vertex (Anthropic)","doc":"https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude","models":{"claude-opus-5-5@default":{"id":"claude-opus-5-5@default","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"claude-sonnet-5-5@default":{"id":"claude-sonnet-5-5@default","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"claude-sonnet-5@default":{"id":"claude-sonnet-5@default","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"claude-sonnet-4-5@20250929":{"id":"claude-sonnet-4-5@20250929","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"claude-opus-4-8@default":{"id":"claude-opus-4-8@default","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"claude-opus-4-5@20251101":{"id":"claude-opus-4-5@20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-01","last_updated":"2025-11-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"claude-fable-5@default":{"id":"claude-fable-5@default","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"claude-opus-4-6@default":{"id":"claude-opus-4-6@default","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"claude-opus-5@default":{"id":"claude-opus-5@default","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"claude-sonnet-4-6@default":{"id":"claude-sonnet-4-6@default","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"claude-sonnet-4@20250514":{"id":"claude-sonnet-4@20250514","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"claude-opus-4-7@default":{"id":"claude-opus-4-7@default","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"claude-opus-4@20250514":{"id":"claude-opus-4@20250514","name":"Claude Opus 4","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-20250514"},"claude-fable-5-1@default":{"id":"claude-fable-5-1@default","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"claude-haiku-4-5@20251001":{"id":"claude-haiku-4-5@20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"claude-haiku-5-5@default":{"id":"claude-haiku-5-5@default","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-opus-4-1@20250805":{"id":"claude-opus-4-1@20250805","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"}}},"regolo-ai":{"id":"regolo-ai","env":["REGOLO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.regolo.ai/v1","name":"Regolo AI","doc":"https://docs.regolo.ai/","models":{"apertus-70b":{"id":"apertus-70b","name":"Apertus 70B","description":"Fully open 70B multilingual LLM supporting 1800+ languages with 65K context. Trained on 15T tokens of compliant open data. Apache 2.0, EU AI Act compliant.","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":30000,"output":30000},"cost":{"input":0.46,"output":2.42},"canonical_model_id":"swiss-ai/apertus-70b"},"qwen3.5-9b":{"id":"qwen3.5-9b","name":"Qwen3.5-9B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-01","last_updated":"2026-02-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.15,"output":0.6}},"brick-v1-beta":{"id":"brick-v1-beta","name":"Brick v1 Beta","description":"Semantic router by Regolo.ai that directs each request to the most suitable model, optimizing costs and performance","family":"model-router","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-02-06","last_updated":"2026-02-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":100000,"output":15000},"status":"beta","cost":{"input":0,"output":0}},"glm5.2":{"id":"glm5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":96000,"output":96000},"cost":{"input":2.31,"output":6},"canonical_model_id":"zhipuai/glm-5.2"},"brick-complexity-pro":{"id":"brick-complexity-pro","name":"Brick Complexity Pro","description":"Complexity classifier that powers the Brick semantic router by extracting query difficulty","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-02-06","last_updated":"2026-02-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":100000,"output":15000},"cost":{"input":0.12,"output":0.46}},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama 3.3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.6,"output":2.7}},"mistral-small-4-119b":{"id":"mistral-small-4-119b","name":"Mistral Small 4 119B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-15","last_updated":"2026-03-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":16384},"cost":{"input":0.75,"output":3}},"qwen-image":{"id":"qwen-image","name":"Qwen-Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-03-01","last_updated":"2026-03-01","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":4096},"cost":{"input":0.5,"output":2}},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT-OSS-20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-03-01","last_updated":"2026-03-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.4,"output":1.8}},"faster-whisper-large-v3":{"id":"faster-whisper-large-v3","type":"transcription","name":"Faster Whisper Large v3","description":"Open Whisper checkpoint for robust multilingual transcription and captioning","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-10-01","last_updated":"2024-10-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"openai/whisper-large-v3"},"qwen3-embedding-8b":{"id":"qwen3-embedding-8b","name":"Qwen3-Embedding-8B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-01","last_updated":"2026-02-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0.1,"output":0.1}},"qwen3.5-122b":{"id":"qwen3.5-122b","name":"Qwen3.5-122B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-01","last_updated":"2026-02-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.9,"output":3.6}},"deepseek-ocr-2":{"id":"deepseek-ocr-2","name":"DeepSeek OCR 2","description":"High-accuracy OCR model for extracting text from documents, screenshots, receipts, and natural scenes","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":4000,"output":4000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-ocr-2"},"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":120000,"output":120000},"cost":{"input":0.58,"output":2.42},"canonical_model_id":"alibaba/qwen3.8-27b"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT-OSS-120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1,"output":4.2}},"qwen3-coder-next":{"id":"qwen3-coder-next","name":"Qwen3-Coder-Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-01","last_updated":"2026-03-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.3,"output":1.2}},"qwen3-reranker-4b":{"id":"qwen3-reranker-4b","name":"Qwen3-Reranker-4B","description":"Reranking model for improving retrieval quality in search and recommendation systems","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-01","last_updated":"2026-02-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0.12,"output":0.12}},"gemma4-31b":{"id":"gemma4-31b","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":100000,"output":100000},"cost":{"input":0.46,"output":2.42},"canonical_model_id":"google/gemma-4-31b-it"}}},"inferx":{"id":"inferx","env":["INFERX_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://model.inferx.net/endpoints/v1","name":"InferX","doc":"https://model.inferx.net/endpoints","models":{"Qwen3-Embedding-8B":{"id":"Qwen3-Embedding-8B","name":"Qwen3-Embedding-8B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":0},"cost":{"input":0,"output":0}},"Devstral-2-123B-Instruct-2512-int4-AutoRound":{"id":"Devstral-2-123B-Instruct-2512-int4-AutoRound","name":"Devstral-2-123B-Instruct-2512-int4-AutoRound","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"mistral/devstral-2512"},"Ornith-1.0-35B-FP8":{"id":"Ornith-1.0-35B-FP8","name":"Ornith-1.0-35B-FP8","description":"Large coding-reasoning model for agentic software tasks and RL search","family":"ornith","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-06-25","last_updated":"2026-06-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":100000},"cost":{"input":0,"output":0},"canonical_model_id":"deepreinforce/ornith-1.0-35b"},"gemma-4-31B-it-fp8":{"id":"gemma-4-31B-it-fp8","name":"Gemma 4 31B IT FP8","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-31b-it"},"Qwen3.6-35B-A3B-FP8":{"id":"Qwen3.6-35B-A3B-FP8","name":"Qwen3.6 35B A3B FP8","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"mimo-v25":{"id":"mimo-v25","name":"mimo-v25","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":100000},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"Qwen3-Coder-Next-FP8":{"id":"Qwen3-Coder-Next-FP8","name":"Qwen3 Coder Next FP8","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256144,"output":65536},"cost":{"input":0,"output":0}},"Qwen3.6-35B-A3B-fp8-no-thinking":{"id":"Qwen3.6-35B-A3B-fp8-no-thinking","name":"Qwen3.6-35B-A3B-fp8-no-thinking","description":"Qwen3.6-35B-A3B-fp8 disable thinking","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen3-Coder-Next-FP8-no-thinking":{"id":"Qwen3-Coder-Next-FP8-no-thinking","name":"Qwen3-Coder-Next-FP8-no-thinking","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":260000,"output":65536},"cost":{"input":0,"output":0}},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"deepseek-v4-flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":100000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"Agents-A1":{"id":"Agents-A1","name":"Agents-A1","description":"35B MoE agentic model built for long-horizon search, engineering, and scientific reasoning tasks","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"release_date":"2026-06-26","last_updated":"2026-06-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":100000},"cost":{"input":0,"output":0}},"Qwen3.6-27B-FP8":{"id":"Qwen3.6-27B-FP8","name":"Qwen3.6 27B FP8","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.6-27b"}}},"zai":{"id":"zai","env":["ZHIPU_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.z.ai/api/paas/v4","name":"Z.AI","doc":"https://docs.z.ai/guides/overview/pricing","models":{"glm-4.6v":{"id":"glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.05,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6v"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-4.6v-flash":{"id":"glm-4.6v-flash","name":"GLM-4.6V-Flash","description":"Lightweight GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6v-flash"},"glm-4.5v":{"id":"glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":16384},"cost":{"input":0.6,"output":1.8,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5v"},"glm-4.5-air":{"id":"glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.2,"output":1.1,"cache_read":0.03,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-air"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"glm-4.7-flash":{"id":"glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"glm-5":{"id":"glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2,"cache_write":0},"canonical_model_id":"zhipuai/glm-5"},"glm-5v-turbo":{"id":"glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24,"cache_write":0},"canonical_model_id":"zhipuai/glm-5v-turbo"},"glm-5.3-flashx":{"id":"glm-5.3-flashx","name":"GLM-5.3-FlashX","description":"High-speed GLM-5.3-Flash serving option for coding and agent workflows","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.37,"output":1.25,"cache_read":0.075,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5-turbo":{"id":"glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24,"cache_write":0},"canonical_model_id":"zhipuai/glm-5-turbo"},"glm-4.6":{"id":"glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"glm-4.5":{"id":"glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5"},"glm-4.7-flashx":{"id":"glm-4.7-flashx","name":"GLM-4.7-FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.07,"output":0.4,"cache_read":0.01,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"glm-4.5-flash":{"id":"glm-4.5-flash","name":"GLM-4.5-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":98304},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-flash"}}},"commandcode":{"id":"commandcode","env":["CMD_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.commandcode.ai/provider/v1","name":"CommandCode","doc":"https://commandcode.ai/docs/provider","models":{"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":0},"canonical_model_id":"openai/gpt-5.5"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":0},"canonical_model_id":"openai/gpt-5.4"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":45,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075,"cache_write":0},"canonical_model_id":"openai/gpt-5.4-mini"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":2,"output":8,"cache_read":0.5,"cache_write":0},"canonical_model_id":"openai/gpt-5.3-codex"},"xai/grok-4.7":{"id":"xai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"xai/grok-4.5":{"id":"xai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.5"},"xai/grok-4.6":{"id":"xai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash-fast":{"id":"deepseek/deepseek-v4-flash-fast","name":"DeepSeek V4 Flash Fast","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.28,"output":0.56,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4.1-flash-fast":{"id":"deepseek/deepseek-v4.1-flash-fast","name":"DeepSeek V4.1 Flash Fast","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.16,"output":0.58,"cache_read":0.016},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"nvidia/nemotron-3-ultra-550b-a55b":{"id":"nvidia/nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.3-flashx":{"id":"z-ai/glm-5.3-flashx","name":"GLM-5.3-FlashX","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.37,"output":1.25,"cache_read":0.075},"canonical_model_id":"zhipuai/glm-5.3-flash"},"inclusionai/ling-3.1-flash:free":{"id":"inclusionai/ling-3.1-flash:free","name":"Ling 3.1 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"inclusionai/ling-3.1-flash"},"meta/muse-spark-1.3-contributor":{"id":"meta/muse-spark-1.3-contributor","name":"Muse Spark 1.3 (Contributor)","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.002},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.2-contributor":{"id":"meta/muse-spark-1.2-contributor","name":"Muse Spark 1.2 (Contributor)","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.002},"canonical_model_id":"meta/muse-spark-1.2"},"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-wonder-1.0":{"id":"meta/muse-wonder-1.0","name":"Muse Wonder","description":"Muse Wonder is a frontier multimodal reasoning model from Meta for personal and professional agents, long-horizon coding, and computer use. It combines Active Memory, which carries relevant context across conversations, with Active Focus, which keeps it working through blockers toward a goal.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.7-Code-Highspeed":{"id":"moonshotai/Kimi-K2.7-Code-Highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"MiniMaxAI/MiniMax-M2.7":{"id":"MiniMaxAI/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":1000000},"cost":{"input":0.5,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"stepfun/Step-3.5-Flash":{"id":"stepfun/Step-3.5-Flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.09,"output":0.3,"cache_read":0.02},"canonical_model_id":"stepfun/step-3.5-flash"},"stepfun/Step-5-Preview":{"id":"stepfun/Step-5-Preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":1000000},"cost":{"input":1,"output":2.7,"cache_read":0.05},"canonical_model_id":"stepfun/step-5-preview"},"stepfun/Step-3.7-Flash":{"id":"stepfun/Step-3.7-Flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"Qwen/Qwen3.8-Flash":{"id":"Qwen/Qwen3.8-Flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.16,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-flash"},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.4,"output":3,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3.8-Max":{"id":"Qwen/Qwen3.8-Max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"Qwen/Qwen3.8-Omni-Flash":{"id":"Qwen/Qwen3.8-Omni-Flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"Qwen/Qwen3.8-Max-0902":{"id":"Qwen/Qwen3.8-Max-0902","name":"Qwen3.8 Max 0902","description":"2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.2-Fast":{"id":"zai-org/GLM-5.2-Fast","name":"GLM-5.2-Fast","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3,"output":10.25,"cache_read":0.5},"canonical_model_id":"zhipuai/glm-5.2"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15,"cache_write":0.08334},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"google/gemini-4-argon"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.6-flash"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy3-paid":{"id":"tencent/hy3-paid","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0.14,"output":0.58,"cache_read":0.035},"canonical_model_id":"tencent/hy3"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"sakana/fugu-ultra"},"mistral/mistral-large-4":{"id":"mistral/mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":262144},"cost":{"input":1.36,"output":4.18,"cache_read":0.14},"canonical_model_id":"mistral/mistral-large-4"}}},"google":{"id":"google","env":["GOOGLE_API_KEY","GOOGLE_GENERATIVE_AI_API_KEY","GEMINI_API_KEY"],"npm":"@ai-sdk/google","name":"Google","doc":"https://ai.google.dev/gemini-api/docs/models","models":{"gemini-2.5-flash-preview-tts":{"id":"gemini-2.5-flash-preview-tts","type":"speech","name":"Gemini 2.5 Flash Preview TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gemini-flash","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-05-01","last_updated":"2025-05-01","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":16384},"cost":{"input":0.5,"output":10}},"gemini-3.1-flash-image":{"id":"gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","video","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":60},"canonical_model_id":"google/gemini-3.1-flash-image"},"gemini-flash-lite-latest":{"id":"gemini-flash-lite-latest","name":"Gemini Flash-Lite Latest","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03}},"gemini-3.1-flash-tts-preview":{"id":"gemini-3.1-flash-tts-preview","type":"speech","name":"Gemini 3.1 Flash TTS Preview","description":"Low-latency speech generation with steerable prompts and expressive audio tags","family":"gemini-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-15","last_updated":"2026-04-15","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":16384},"cost":{"input":1,"output":20},"canonical_model_id":"google/gemini-3.1-flash-tts-preview"},"gemini-3-pro-image-preview":{"id":"gemini-3-pro-image-preview","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":2,"output":120},"canonical_model_id":"google/gemini-3-pro-image-preview"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1}},"gemini-3.1-flash-image-preview":{"id":"gemini-3.1-flash-image-preview","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":65536},"cost":{"input":0.5,"output":60}},"gemini-2.5-computer-use-preview-10-2025":{"id":"gemini-2.5-computer-use-preview-10-2025","name":"Gemini 2.5 Computer Use Preview 10-2025","description":"Specialized Gemini 2.5 model for browser-control agents that automate UI tasks","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-10-07","last_updated":"2025-10-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"cost":{"input":1.25,"output":10,"tiers":[{"input":2.5,"output":15,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15}},"canonical_model_id":"google/gemini-2.5-computer-use-preview-10-2025"},"veo-3.1-generate-preview":{"id":"veo-3.1-generate-preview","type":"video","name":"Veo 3.1","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-15","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":8192},"status":"beta","canonical_model_id":"google/veo-3.1-generate-preview"},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3}},"gemini-3.5-live-translate-preview":{"id":"gemini-3.5-live-translate-preview","type":"realtime","name":"Gemini 3.5 Live Translate Preview","description":"Low-latency audio-to-audio model for real-time speech translation across 70+ languages","family":"gemini-pro","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["audio"],"output":["audio","text"]},"open_weights":false,"limit":{"context":16384,"output":32768},"cost":{"input":3.5,"output":21,"input_audio":3.5,"output_audio":21},"canonical_model_id":"google/gemini-3.5-live-translate-preview"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"gemini-3.1-pro-preview-customtools":{"id":"gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}}},"gemini-embedding-001":{"id":"gemini-embedding-001","type":"embedding","name":"Gemini Embedding 001","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-05","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2048,"output":1},"cost":{"input":0.15,"output":0}},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5}},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.7-flash"},"lyria-3-clip-preview":{"id":"lyria-3-clip-preview","name":"Lyria 3 Clip Preview","description":"Music generation model for short 30-second clips, loops, and previews from text or image prompts","family":"lyria","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-25","last_updated":"2026-03-25","modalities":{"input":["text","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/lyria-3-clip-preview"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}}},"gemini-embedding-2":{"id":"gemini-embedding-2","type":"embedding","name":"Gemini Embedding 2","description":"Multimodal embedding model mapping text, images, video, audio, and PDFs into a unified embedding space","family":"gemini","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-11","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1},"cost":{"input":0.2,"output":0,"input_audio":6.5},"canonical_model_id":"google/gemini-embedding-2"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.8-flash"},"gemini-3-pro-image":{"id":"gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":120},"canonical_model_id":"google/gemini-3-pro-image"},"deep-research-preview-04-2026":{"id":"deep-research-preview-04-2026","name":"Deep Research Preview (Apr-21-2026)","description":"Agentic model for autonomous multi-step research, synthesis, and cited reports","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/deep-research-preview-04-2026"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5}},"veo-3.1-fast-generate-preview":{"id":"veo-3.1-fast-generate-preview","type":"video","name":"Veo 3.1 fast","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-15","last_updated":"2026-01-01","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":8192},"canonical_model_id":"google/veo-3.1-fast-generate-preview"},"gemini-3.1-flash-lite-image":{"id":"gemini-3.1-flash-lite-image","type":"image","name":"Nano Banana 2 Lite","description":"Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":4096},"cost":{"input":0.25,"output":30},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":32768},"cost":{"input":2,"output":10,"cache_read":0.1,"input_audio":2},"canonical_model_id":"google/gemini-4-argon"},"gemma-4-26b-a4b-it":{"id":"gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768}},"gemini-3.1-flash-live-preview":{"id":"gemini-3.1-flash-live-preview","type":"realtime","name":"Gemini 3.1 Flash Live Preview","description":"High-quality, low-latency Live API model for real-time dialogue and voice-first AI applications","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-26","last_updated":"2026-03-26","modalities":{"input":["text","image","video","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":0.75,"output":4.5,"input_audio":3,"output_audio":12},"canonical_model_id":"google/gemini-3.1-flash-live-preview"},"veo-3.1-lite-generate-preview":{"id":"veo-3.1-lite-generate-preview","type":"video","name":"Veo 3.1 lite","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":8192},"canonical_model_id":"google/veo-3.1-lite-generate-preview"},"gemini-flash-latest":{"id":"gemini-flash-latest","name":"Gemini Flash Latest","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75}},"deep-research-max-preview-04-2026":{"id":"deep-research-max-preview-04-2026","name":"Deep Research Max Preview (Apr-21-2026)","description":"Maximum-comprehensiveness agentic researcher for multi-step investigation, synthesis, and cited reports","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/deep-research-max-preview-04-2026"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gemma-4-31b-it":{"id":"gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768}},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"gemini-3.1-flash-lite-preview":{"id":"gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Legacy model retained for compatibility with older integrations","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5}},"gemini-omni-flash-preview":{"id":"gemini-omni-flash-preview","type":"video","name":"Gemini Omni Flash Preview","description":"Video generation and editing model for fast, conversational text- and image-to-video workflows","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":1.5,"output":17.5},"canonical_model_id":"google/gemini-omni-flash-preview"},"gemini-2.5-flash-image":{"id":"gemini-2.5-flash-image","type":"image","name":"Nano Banana","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":30,"cache_read":0.075},"canonical_model_id":"google/gemini-2.5-flash-image"},"gemini-2.5-pro-preview-tts":{"id":"gemini-2.5-pro-preview-tts","type":"speech","name":"Gemini 2.5 Pro Preview TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gemini-flash","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-05-01","last_updated":"2025-05-01","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":16384},"cost":{"input":1,"output":20}},"lyria-3-pro-preview":{"id":"lyria-3-pro-preview","name":"Lyria 3 Pro Preview","description":"Music generation model for full-length songs from text or images with vocals and structure","family":"lyria","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-25","last_updated":"2026-03-25","modalities":{"input":["text","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/lyria-3-pro-preview"}}},"umans-ai":{"id":"umans-ai","env":["UMANS_AI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.code.umans.ai/v1","name":"Umans AI","doc":"https://app.umans.ai/offers/code/docs/orgs","models":{"umans-glm-5.3":{"id":"umans-glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131071},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"umans-flash":{"id":"umans-flash","name":"Umans Flash","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.15,"output":1,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"umans-deepseek-v4.1-flash":{"id":"umans-deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393215},"cost":{"input":0.15,"output":0.6,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"umans-kimi-k3":{"id":"umans-kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"umans-coder":{"id":"umans-coder","name":"Umans Coder","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131071},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"umans-glm-5.3-flash":{"id":"umans-glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131071},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"umans-deepseek-v4-flash-0731":{"id":"umans-deepseek-v4-flash-0731","name":"DeepSeek V4 Flash","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393215},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"}}},"standardcompute":{"id":"standardcompute","env":["STANDARDCOMPUTE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.stdcmpt.com/v1","name":"Standard Compute","doc":"https://docs.standardcompute.com/models","models":{"standardcompute":{"id":"standardcompute","name":"Standard Compute","description":"Flat-rate smart-routing gateway: one model id, each request routed across the customer's enabled models from a curated ~1M-context catalog (Claude, GPT, Gemini, DeepSeek, GLM, Qwen, Kimi, MiniMax, Grok, MiMo) or pinned to a user-selected model","family":"auto","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-03-01","last_updated":"2026-10-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"}},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":1.044,"output":2.088,"cache_read":0.2088},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":1.122,"output":3.366,"cache_read":0.1122},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.0826,"output":0.1652,"cache_read":0.01652},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.14,"output":0.42,"cache_read":0.0042},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.08,"output":0.5,"cache_read":0.012},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.5625,"output":1.8,"cache_read":0.105},"canonical_model_id":"zhipuai/glm-5.2"},"qwen/qwen3.7-flash":{"id":"qwen/qwen3.7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.038,"tiers":[{"input":0.1,"output":0.4,"cache_read":0.02,"cache_write":0.125,"tier":{"type":"context","size":32000}},{"input":0.2,"output":0.8,"cache_read":0.04,"cache_write":0.25,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.32,"output":1.28,"cache_read":0.064,"cache_write":0.4,"tiers":[{"input":0.96,"output":3.84,"cache_read":0.192,"cache_write":1.2,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.96,"output":3.84,"cache_read":0.192,"cache_write":1.2}},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.425,"output":2.55,"cache_read":0.085,"cache_write":0.53125},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":1.475,"output":4.425,"cache_read":0.295,"cache_write":1.84375},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.8-max-0902":{"id":"qwen/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"mistralai/mistral-large-4-0":{"id":"mistralai/mistral-large-4-0","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"xiaomi/mimo-v2.5-pro":{"id":"xiaomi/mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2.6-flash":{"id":"xiaomi/mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.43,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":30000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":1.25,"output":10,"cache_read":0.125,"cache_write":0.375,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":0.375,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.083333},"canonical_model_id":"google/gemini-2.5-flash"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"}}},"scaleway":{"id":"scaleway","env":["SCALEWAY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.scaleway.ai/v1","name":"Scaleway","doc":"https://www.scaleway.com/en/docs/generative-apis/","models":{"whisper-large-v3":{"id":"whisper-large-v3","name":"Whisper Large v3","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2023-09","release_date":"2023-09-01","last_updated":"2026-03-17","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":0,"output":8192},"cost":{"input":0.003,"output":0}},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"status":"beta","cost":{"input":0.468,"output":0.936,"reasoning":0.936,"cache_read":0.0936},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-05-01","last_updated":"2026-05-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"status":"beta","cost":{"input":0.25,"output":1.5},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2026-03-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":100000,"output":16384},"cost":{"input":0.9,"output":0.9}},"qwen3-235b-a22b-instruct-2507":{"id":"qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-01","last_updated":"2026-03-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":260000,"output":16384},"cost":{"input":0.75,"output":2.25,"reasoning":8.4},"canonical_model_id":"alibaba/qwen3-235b-a22b"},"bge-multilingual-gemma2":{"id":"bge-multilingual-gemma2","name":"BGE Multilingual Gemma2","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-07-26","last_updated":"2025-06-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8191,"output":3072},"cost":{"input":0.1,"output":0}},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"status":"beta","cost":{"input":2.01,"output":6.15},"canonical_model_id":"zhipuai/glm-5.3"},"qwen3-embedding-8b":{"id":"qwen3-embedding-8b","name":"Qwen3 Embedding 8B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-05","last_updated":"2026-03-17","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":4096},"cost":{"input":0.1,"output":0}},"gemma-4-26b-a4b-it":{"id":"gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-01","last_updated":"2026-05-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"status":"beta","cost":{"input":0.25,"output":0.5},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"mistral-small-3.2-24b-instruct-2506":{"id":"mistral-small-3.2-24b-instruct-2506","name":"Mistral Small 3.2 24B Instruct (2506)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.15,"output":0.35},"canonical_model_id":"mistral/mistral-small-2506"},"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"beta","cost":{"input":0.684,"output":3.762,"reasoning":3.762,"cache_read":0.137},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen3-coder-30b-a3b-instruct":{"id":"qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2026-03-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.2,"output":0.8},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5 397B A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT-OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2024-01-01","last_updated":"2026-03-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.15,"output":0.6}},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":1.8,"output":5.5},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"status":"beta","cost":{"input":0.672,"output":2.016,"reasoning":2.016,"cache_read":0.134},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"mistral-medium-3.5-128b":{"id":"mistral-medium-3.5-128b","name":"Mistral Medium 3.5 128B","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"mistral/mistral-medium-2604"},"pixtral-12b-2409":{"id":"pixtral-12b-2409","name":"Pixtral 12B 2409","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2024-09-25","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.2,"output":0.2},"canonical_model_id":"mistral/pixtral-12b"}}},"bailing":{"id":"bailing","env":["BAILING_API_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"https://api.tbox.cn/api/llm/v1","name":"Bailing","doc":"https://alipaytbox.yuque.com/sxs0ba/ling/intro","models":{"Ling-1T":{"id":"Ling-1T","name":"Ling-1T","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"ling","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-10","last_updated":"2025-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.57,"output":2.29}},"Ring-1T":{"id":"Ring-1T","name":"Ring-1T","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"ring","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-10","last_updated":"2025-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.57,"output":2.29}}}},"tencent-coding-plan":{"id":"tencent-coding-plan","env":["TENCENT_CODING_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.lkeap.cloud.tencent.com/coding/v3","name":"Tencent Coding Plan (China)","doc":"https://cloud.tencent.com/document/product/1772/128947","models":{"tc-code-latest":{"id":"tc-code-latest","name":"Auto","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-08","last_updated":"2026-03-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"hunyuan-2.0-instruct":{"id":"hunyuan-2.0-instruct","name":"Tencent HY 2.0 Instruct","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-08","last_updated":"2026-03-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"glm-5":{"id":"glm-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":16384},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"hunyuan-2.0-thinking":{"id":"hunyuan-2.0-thinking","name":"Tencent HY 2.0 Think","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-03-08","last_updated":"2026-03-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"hunyuan-turbos":{"id":"hunyuan-turbos","name":"Hunyuan-TurboS","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-08","last_updated":"2026-03-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"hunyuan-t1":{"id":"hunyuan-t1","name":"Hunyuan-T1","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-03-08","last_updated":"2026-03-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi-K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"minimax-m2.5":{"id":"minimax-m2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}}}},"zenmux":{"id":"zenmux","env":["ZENMUX_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://zenmux.ai/api/v1","name":"ZenMux","doc":"https://docs.zenmux.ai","models":{"sapiens-ai/agnes-1.5-pro":{"id":"sapiens-ai/agnes-1.5-pro","name":"Agnes 1.5 Pro","description":"Flagship model for demanding analysis, coding, and production agent workflows","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-21","last_updated":"2026-03-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.16,"output":0.8}},"sapiens-ai/agnes-1.5-lite":{"id":"sapiens-ai/agnes-1.5-lite","name":"Agnes 1.5 Lite","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-26","last_updated":"2026-03-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.12,"output":0.6}},"anthropic/claude-3.5-haiku":{"id":"anthropic/claude-3.5-haiku","name":"Claude 3.5 Haiku","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2024-11-04","last_updated":"2024-11-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.8,"output":4,"cache_read":0.08,"cache_write":1}},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-sonnet-5-free":{"id":"anthropic/claude-sonnet-5-free","name":"Claude Sonnet 5 (Free)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"anthropic/claude-sonnet-4":{"id":"anthropic/claude-sonnet-4","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-06","last_updated":"2026-02-06","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"anthropic/claude-opus-4.1":{"id":"anthropic/claude-opus-4.1","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25}},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":4},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-3.7-sonnet":{"id":"anthropic/claude-3.7-sonnet","name":"Claude 3.7 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-02-24","last_updated":"2025-02-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-18","last_updated":"2026-02-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-opus-4":{"id":"anthropic/claude-opus-4","name":"Claude Opus 4","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.435,"output":0.87,"cache_read":0.003625},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v3.2-exp":{"id":"deepseek/deepseek-v3.2-exp","name":"DeepSeek-V3.2-Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163000,"output":64000},"cost":{"input":0.22,"output":0.33}},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-chat":{"id":"deepseek/deepseek-chat","name":"DeepSeek-V3.2 (Non-thinking Mode)","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"status":"deprecated","cost":{"input":0.28,"output":0.42,"cache_read":0.03}},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-12-05","last_updated":"2025-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"cost":{"input":0.28,"output":0.43}},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"z-ai/glm-4.6v":{"id":"z-ai/glm-4.6v","name":"GLM 4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.1456,"output":0.4367,"cache_read":0.0291,"tiers":[{"input":0.2911,"output":0.8734,"cache_read":0.0582,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6v"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-4.7-flash-free":{"id":"z-ai/glm-4.7-flash-free","name":"GLM 4.7 Flash (Free)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"z-ai/glm-4.6v-flash":{"id":"z-ai/glm-4.6v-flash","name":"GLM 4.6V FlashX","description":"Lightweight GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.0218,"output":0.2184,"cache_read":0.0044,"tiers":[{"input":0.0437,"output":0.4367,"cache_read":0.0044,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6v-flash"},"z-ai/glm-4.6v-flash-free":{"id":"z-ai/glm-4.6v-flash-free","name":"GLM 4.6V Flash (Free)","description":"Lightweight GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"tiers":[{"input":0,"output":0,"cache_read":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6v-flash"},"z-ai/glm-4.5-air":{"id":"z-ai/glm-4.5-air","name":"GLM 4.5 Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":96000},"cost":{"input":0.1165,"output":0.2911,"cache_read":0.0233,"tiers":[{"input":0.1747,"output":1.1645,"cache_read":0.0349,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.5-air"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"GLM 4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.2911,"output":1.1645,"cache_read":0.0582,"tiers":[{"input":0.5823,"output":2.3291,"cache_read":0.1165,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM 5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.8781,"output":3.5126,"cache_read":0.1903,"tiers":[{"input":1.1709,"output":4.098,"cache_read":0.2927,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-image":{"id":"z-ai/glm-image","type":"image","name":"GLM-Image","description":"GLM-Image is an image generation model adopts a hybrid autoregressive + diffusion decoder architecture. In general image generation quality, GLM‑Image aligns with mainstream latent diffusion approaches, but it shows significant advantages in text-rendering and knowledge‑intensive generation scenarios. It performs especially well in tasks requiring precise semantic understanding and complex information expression, while maintaining strong capabilities in high‑fidelity and fine‑grained detail generation. In addition to text‑to‑image generation, GLM‑Image also supports a rich set of image‑to‑image tasks including image editing, style transfer, identity‑preserving generation, and multi‑subject consistency.","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-14","last_updated":"2026-01-14","modalities":{"input":["text"],"output":["image"]},"open_weights":true,"limit":{"context":10240,"output":0},"canonical_model_id":"zhipuai/glm-image"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM 5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.58,"output":2.6,"cache_read":0.14,"tiers":[{"input":0.87,"output":3.18,"cache_read":0.22,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-5v-turbo":{"id":"z-ai/glm-5v-turbo","name":"GLM 5V Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":0.726,"output":3.1946,"cache_read":0.1743,"tiers":[{"input":1.0165,"output":3.7754,"cache_read":0.2614,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5v-turbo"},"z-ai/glm-5.3-flashx":{"id":"z-ai/glm-5.3-flashx","name":"GLM 5.3 FlashX","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.375,"output":1.25,"cache_read":0.075},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM 5 Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":0.73,"output":3.19,"cache_read":0.174,"tiers":[{"input":1.02,"output":3.77,"cache_read":0.261,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5-turbo"},"z-ai/glm-4.6":{"id":"z-ai/glm-4.6","name":"GLM 4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.2911,"output":1.1645,"cache_read":0.0582,"tiers":[{"input":0.5823,"output":2.3291,"cache_read":0.1165,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.98,"output":3.08,"cache_read":0.182},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-4.5":{"id":"z-ai/glm-4.5","name":"GLM 4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":96000},"cost":{"input":0.2911,"output":1.1645,"cache_read":0.0582,"tiers":[{"input":0.5823,"output":2.3291,"cache_read":0.1165,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.5"},"z-ai/glm-4.7-flashx":{"id":"z-ai/glm-4.7-flashx","name":"GLM 4.7 FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.0728,"output":0.4367,"cache_read":0.0146},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"inclusionai/ring-1t":{"id":"inclusionai/ring-1t","name":"Ring-1T","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-10-12","last_updated":"2025-10-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"status":"deprecated","cost":{"input":0.56,"output":2.24,"cache_read":0.11}},"inclusionai/ring-2.6-1t":{"id":"inclusionai/ring-2.6-1t","name":"inclusionAI: Ring-2.6-1T","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-12-31","release_date":"2026-05-07","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":65000},"cost":{"input":0.3,"output":2.5,"cache_read":0.06}},"inclusionai/ling-1t":{"id":"inclusionai/ling-1t","name":"Ling-1T","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-10-09","last_updated":"2025-10-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"status":"deprecated","cost":{"input":0.56,"output":2.24,"cache_read":0.11}},"kuaishou/kat-coder-pro-v2":{"id":"kuaishou/kat-coder-pro-v2","name":"KAT-Coder-Pro-V2","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-30","last_updated":"2026-03-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":80000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06}},"qwen/qwen3-coder-plus":{"id":"qwen/qwen3-coder-plus","name":"Qwen3-Coder-Plus","description":"Qwen coding model for software agents, repository edits, and code reasoning","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25}},"qwen/qwen3.5-plus":{"id":"qwen/qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-03-20","last_updated":"2026-03-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.8,"output":4.8}},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6-Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-03-30","last_updated":"2026-03-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625,"tiers":[{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5}}},"qwen/qwen3.5-flash":{"id":"qwen/qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-03-20","last_updated":"2026-03-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1020000,"output":1020000},"cost":{"input":0.1,"output":0.4}},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.24,"cache_write":1.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.24,"cache_write":1.5}},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.5,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3-max":{"id":"qwen/qwen3-max","name":"Qwen3-Max-Thinking","description":"Qwen reasoning model for deliberate problem solving, math, and coding","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-01-23","last_updated":"2026-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":1.2,"output":6}},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 0905","description":"Kimi model for long-context chat, coding, and agentic reasoning","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-09-04","last_updated":"2025-09-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":64000},"status":"deprecated","cost":{"input":0.6,"output":2.5,"cache_read":0.15}},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2025-01-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262140,"output":262140},"cost":{"input":0.95,"output":4,"cache_read":0.16}},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2.7-code-highspeed":{"id":"moonshotai/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code HighSpeed","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k2.7-code-free":{"id":"moonshotai/kimi-k2.7-code-free","name":"Kimi K2.7 Code (Free)","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k2-thinking-turbo":{"id":"moonshotai/kimi-k2-thinking-turbo","name":"Kimi K2 Thinking Turbo","description":"Kimi reasoning model for long-horizon research, planning, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":64000},"status":"deprecated","cost":{"input":1.15,"output":8,"cache_read":0.15}},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2025-01-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":64000},"cost":{"input":0.58,"output":3.02,"cache_read":0.1}},"moonshotai/kimi-k3-free":{"id":"moonshotai/kimi-k3-free","name":"Kimi K3 (Free)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":64000},"status":"deprecated","cost":{"input":0.6,"output":2.5,"cache_read":0.15}},"baidu/ernie-5.0-thinking-preview":{"id":"baidu/ernie-5.0-thinking-preview","name":"ERNIE 5.0","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-01-22","last_updated":"2026-01-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"cost":{"input":0.84,"output":3.37}},"stepfun/step-3.7-flash":{"id":"stepfun/step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.2,"output":1.15},"canonical_model_id":"stepfun/step-3.7-flash"},"stepfun/step-3":{"id":"stepfun/step-3","name":"Step-3","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":64000},"status":"deprecated","cost":{"input":0.21,"output":0.57}},"stepfun/step-3.5-flash":{"id":"stepfun/step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-02-02","last_updated":"2026-02-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":0.1,"output":0.3}},"stepfun/step-3.7-flash-free":{"id":"stepfun/step-3.7-flash-free","name":"Step 3.7 Flash (Free)","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0,"output":0},"canonical_model_id":"stepfun/step-3.7-flash"},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-03-20","last_updated":"2026-03-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":204800,"output":131070},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.3055,"output":1.2219}},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax M2","description":"MiniMax model for chat, coding, office work, and agentic tasks","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":204000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.38}},"minimax/minimax-m2.5-lightning":{"id":"minimax/minimax-m2.5-lightning","name":"MiniMax M2.5 highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":204800,"output":131072},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.6,"output":4.8,"cache_read":0.06,"cache_write":0.75}},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":204000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.38}},"minimax/minimax-m3.1-flash":{"id":"minimax/minimax-m3.1-flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":512000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax M2.7 highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-03-20","last_updated":"2026-03-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":204800,"output":131070},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.611,"output":2.4439}},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.6,"output":2.4},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":204800,"output":131072},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375}},"volcengine/doubao-seed-code":{"id":"volcengine/doubao-seed-code","name":"Doubao-Seed-Code","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-11-11","last_updated":"2025-11-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0.17,"output":1.12,"cache_read":0.03}},"volcengine/doubao-seed-2.0-mini":{"id":"volcengine/doubao-seed-2.0-mini","name":"Doubao-Seed-2.0-mini","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-02-14","release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":0.03,"output":0.28,"cache_read":0.01,"cache_write":0.0024}},"volcengine/doubao-seed-2.0-code":{"id":"volcengine/doubao-seed-2.0-code","name":"Doubao Seed 2.0 Code","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-03-20","last_updated":"2026-03-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.9,"output":4.48}},"volcengine/doubao-seed-1.8":{"id":"volcengine/doubao-seed-1.8","name":"Doubao-Seed-1.8","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":0.11,"output":0.28,"cache_read":0.02,"cache_write":0.0024}},"volcengine/doubao-seed-2.0-lite":{"id":"volcengine/doubao-seed-2.0-lite","name":"Doubao-Seed-2.0-lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-02-14","release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":0.09,"output":0.51,"cache_read":0.02,"cache_write":0.0024}},"volcengine/doubao-seed-2.0-pro":{"id":"volcengine/doubao-seed-2.0-pro","name":"Doubao-Seed-2.0-pro","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-02-14","release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":0.45,"output":2.24,"cache_read":0.09,"cache_write":0.0024}},"xiaomi/mimo-v2.5-pro":{"id":"xiaomi/mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1,"output":3,"cache_read":0.2,"tiers":[{"input":2,"output":6,"cache_read":0.4,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.4}},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.4,"output":2,"cache_read":0.08,"tiers":[{"input":0.8,"output":4,"cache_read":0.16,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.8,"output":4,"cache_read":0.16}},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2-flash":{"id":"xiaomi/mimo-v2-flash","name":"MiMo-V2-Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-12-01","release_date":"2025-12-16","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.1,"output":0.3,"cache_read":0.01},"canonical_model_id":"xiaomi/mimo-v2-flash"},"xiaomi/mimo-v2-omni":{"id":"xiaomi/mimo-v2-omni","name":"MiMo V2 Omni","description":"MiMo omni model for text, image, video, audio, and agents","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":265000,"output":265000},"cost":{"input":0.4,"output":2,"cache_read":0.08},"canonical_model_id":"xiaomi/mimo-v2-omni"},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"xiaomi/mimo-v2-pro":{"id":"xiaomi/mimo-v2-pro","name":"MiMo V2 Pro","description":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":256000},"cost":{"input":1,"output":3,"cache_read":0.2,"tiers":[{"input":2,"output":6,"cache_read":0.4,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.4}},"canonical_model_id":"xiaomi/mimo-v2-pro"},"x-ai/grok-4.2-fast-non-reasoning":{"id":"x-ai/grok-4.2-fast-non-reasoning","name":"Grok 4.2 Fast Non Reasoning","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2,"tiers":[{"input":4,"output":12,"cache_read":0.2,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"x-ai/grok-voice-tts-1.0":{"id":"x-ai/grok-voice-tts-1.0","type":"speech","name":"Grok Voice TTS 1.0","description":"Convert text into spoken audio with a single API call. The API supports a rich set of expressive voices, inline speech tags for fine-grained delivery control, and output formats from high-fidelity MP3 to telephony-optimized μ-law.","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":15000,"output":15000},"canonical_model_id":"xai/grok-tts"},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"x-ai/grok-voice-stt-1.0":{"id":"x-ai/grok-voice-stt-1.0","type":"transcription","name":"Grok Voice STT 1.0","description":"Speech-to-text model with word-level timestamps, speaker diarization, multichannel audio, and streaming","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":15000,"output":15000},"canonical_model_id":"xai/grok-voice-transcribe-2.0"},"x-ai/grok-build-0.1":{"id":"x-ai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2},"canonical_model_id":"xai/grok-build-0.1"},"x-ai/grok-4.5":{"id":"x-ai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.5"},"x-ai/grok-imagine-image-2.0":{"id":"x-ai/grok-imagine-image-2.0","type":"image","name":"Grok Imagine Image 2.0","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-08-07","last_updated":"2026-08-07","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":66000,"output":0},"canonical_model_id":"xai/grok-imagine-image-2.0"},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"x-ai/grok-4.2-fast":{"id":"x-ai/grok-4.2-fast","name":"Grok 4.2 Fast","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2,"tiers":[{"input":4,"output":12,"cache_read":0.2,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048000,"output":64000},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":1}},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["pdf","image","text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048000,"output":64000},"cost":{"input":0.1,"output":0.4,"cache_read":0.03,"cache_write":1}},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["pdf","image","text","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048000,"output":64000},"cost":{"input":1.25,"output":10,"cache_read":0.31,"cache_write":4.5}},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-02-19","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","pdf","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048000,"output":64000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":4.5}},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["pdf","image","text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048000,"output":64000},"cost":{"input":0.3,"output":2.5,"cache_read":0.07,"cache_write":1}},"google/gemini-3.1-flash-lite-preview":{"id":"google/gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-03-20","last_updated":"2025-03-20","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":65530},"cost":{"input":0.25,"output":1.5}},"tencent/hy3-preview":{"id":"tencent/hy3-preview","name":"Hy3 preview","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.172,"output":0.572,"cache_read":0.058,"cache_write":0},"canonical_model_id":"tencent/hy3-preview"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-5.5-instant":{"id":"openai/gpt-5.5-instant","name":"GPT-5.5 Instant","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12-01","release_date":"2026-05-05","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5-instant"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-20","last_updated":"2026-03-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":45,"output":225}},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-01-01","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":64000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.17}},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":64000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.12}},"openai/gpt-5.3-chat":{"id":"openai/gpt-5.3-chat","name":"GPT-5.3 Chat","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-20","last_updated":"2026-03-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16380},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.75,"output":14}},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":3.125,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"cache_write":6.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5,"cache_write":6.25}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5-codex":{"id":"openai/gpt-5-codex","name":"GPT-5 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":64000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.12}},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":1,"output":6,"cache_read":0.1,"cache_write":1.25,"tiers":[{"input":2,"output":9,"cache_read":0.2,"cache_write":2.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":2,"output":9,"cache_read":0.2,"cache_write":2.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-20","last_updated":"2026-03-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":3.75,"output":18.75}},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-01-01","release_date":"2026-01-15","last_updated":"2026-01-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":64000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.17}},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1-Codex-Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":64000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":0.25,"output":2,"cache_read":0.03}},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT-5.1-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":64000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.12}},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2-Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":21,"output":168}},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":45,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-20","last_updated":"2026-03-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":0.2,"output":1.25}},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":64000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.12}},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-20","last_updated":"2026-03-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":0.75,"output":4.5}},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-5.1-chat":{"id":"openai/gpt-5.1-chat","name":"GPT-5.1 Chat","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"status":"deprecated","provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.12}},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-20","last_updated":"2026-03-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.75,"output":14}}}},"poe":{"id":"poe","env":["POE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.poe.com/v1","name":"Poe","doc":"https://creator.poe.com/docs/external-applications/openai-compatible-api","models":{"lumalabs/ray2":{"id":"lumalabs/ray2","name":"Ray2","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"ray","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-02-20","last_updated":"2025-02-20","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":5000,"output":0}},"anthropic/claude-sonnet-3.5-june":{"id":"anthropic/claude-sonnet-3.5-june","name":"Claude-Sonnet-3.5-June","description":"Legacy model retained for compatibility with older integrations","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-11-18","last_updated":"2024-11-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":189096,"output":8192},"status":"deprecated","cost":{"input":2.6,"output":13,"cache_read":0.26,"cache_write":3.2}},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude-Opus-4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":128000},"cost":{"input":4.2929,"output":21.4646},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-sonnet-3.5":{"id":"anthropic/claude-sonnet-3.5","name":"Claude-Sonnet-3.5","description":"Legacy model retained for compatibility with older integrations","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-06-05","last_updated":"2024-06-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":189096,"output":8192},"status":"deprecated","cost":{"input":2.6,"output":13,"cache_read":0.26,"cache_write":3.2}},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude-Opus-4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"release_date":"2026-04-15","last_updated":"2026-04-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":128000},"cost":{"input":4.3,"output":21,"cache_read":0.43,"cache_write":5.4}},"anthropic/claude-sonnet-4":{"id":"anthropic/claude-sonnet-4","name":"Claude-Sonnet-4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-05-21","last_updated":"2025-05-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":983040,"output":64000},"cost":{"input":2.6,"output":13,"cache_read":0.26,"cache_write":3.2}},"anthropic/claude-sonnet-3.7":{"id":"anthropic/claude-sonnet-3.7","name":"Claude-Sonnet-3.7","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":196608,"output":128000},"cost":{"input":2.6,"output":13,"cache_read":0.26,"cache_write":3.2}},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude-Opus-4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":0,"max":63999}],"tool_call":true,"temperature":false,"release_date":"2025-11-21","last_updated":"2025-11-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":196608,"output":64000},"cost":{"input":4.3,"output":21,"cache_read":0.43,"cache_write":5.3}},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude-Opus-4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"temperature":false,"release_date":"2026-02-04","last_updated":"2026-02-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":983040,"output":128000},"cost":{"input":4.3,"output":21,"cache_read":0.43,"cache_write":5.3}},"anthropic/claude-haiku-3.5":{"id":"anthropic/claude-haiku-3.5","name":"Claude-Haiku-3.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-10-01","last_updated":"2024-10-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":189096,"output":8192},"cost":{"input":0.68,"output":3.4,"cache_read":0.068,"cache_write":0.85}},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude-Sonnet-4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":31999}],"tool_call":true,"temperature":false,"release_date":"2025-09-26","last_updated":"2025-09-26","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":983040,"output":32768},"cost":{"input":2.6,"output":13,"cache_read":0.26,"cache_write":3.2}},"anthropic/claude-opus-4.1":{"id":"anthropic/claude-opus-4.1","name":"Claude-Opus-4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":31999}],"tool_call":true,"temperature":false,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":196608,"output":32000},"cost":{"input":13,"output":64,"cache_read":1.3,"cache_write":16}},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude-Haiku-4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":63999}],"tool_call":true,"temperature":false,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":192000,"output":64000},"cost":{"input":0.85,"output":4.3,"cache_read":0.085,"cache_write":1.1}},"anthropic/claude-haiku-3":{"id":"anthropic/claude-haiku-3","name":"Claude-Haiku-3","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-03-09","last_updated":"2024-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":189096,"output":8192},"cost":{"input":0.21,"output":1.1,"cache_read":0.021,"cache_write":0.26}},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude-Sonnet-4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"temperature":false,"release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":983040,"output":128000},"cost":{"input":2.6,"output":13,"cache_read":0.26,"cache_write":3.2}},"anthropic/claude-opus-4":{"id":"anthropic/claude-opus-4","name":"Claude-Opus-4","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-05-21","last_updated":"2025-05-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":192512,"output":28672},"cost":{"input":13,"output":64,"cache_read":1.3,"cache_write":16}},"xai/grok-4.20-multi-agent":{"id":"xai/grok-4.20-multi-agent","name":"Grok-4.20-Multi-Agent","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2026-03-13","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":0},"cost":{"input":2,"output":6,"cache_read":0.2}},"xai/grok-code-fast-1":{"id":"xai/grok-code-fast-1","name":"Grok Code Fast 1","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-08-22","last_updated":"2025-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.2,"output":1.5,"cache_read":0.02}},"xai/grok-4.1-fast-non-reasoning":{"id":"xai/grok-4.1-fast-non-reasoning","name":"Grok-4.1-Fast-Non-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000}},"xai/grok-4-fast-non-reasoning":{"id":"xai/grok-4-fast-non-reasoning","name":"Grok-4-Fast-Non-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-09-16","last_updated":"2025-09-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":128000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}},"xai/grok-4.1-fast-reasoning":{"id":"xai/grok-4.1-fast-reasoning","name":"Grok-4.1-Fast-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000}},"xai/grok-4-fast-reasoning":{"id":"xai/grok-4-fast-reasoning","name":"Grok-4-Fast-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-09-16","last_updated":"2025-09-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":128000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}},"xai/grok-3-mini":{"id":"xai/grok-3-mini","name":"Grok 3 Mini","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2025-04-11","last_updated":"2025-04-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.3,"output":0.5,"cache_read":0.075}},"xai/grok-4":{"id":"xai/grok-4","name":"Grok-4","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-07-10","last_updated":"2025-07-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.75}},"xai/grok-3":{"id":"xai/grok-3","name":"Grok 3","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-04-11","last_updated":"2025-04-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":3,"output":15,"cache_read":0.75}},"topazlabs-co/topazlabs":{"id":"topazlabs-co/topazlabs","name":"TopazLabs","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"topazlabs","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":204,"output":0}},"elevenlabs/elevenlabs-music":{"id":"elevenlabs/elevenlabs-music","name":"ElevenLabs-Music","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"elevenlabs","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-08-29","last_updated":"2025-08-29","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":2000,"output":0}},"elevenlabs/elevenlabs-v3":{"id":"elevenlabs/elevenlabs-v3","name":"ElevenLabs-v3","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"elevenlabs","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":128000,"output":0}},"elevenlabs/elevenlabs-v2.5-turbo":{"id":"elevenlabs/elevenlabs-v2.5-turbo","name":"ElevenLabs-v2.5-Turbo","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"elevenlabs","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-10-28","last_updated":"2024-10-28","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":128000,"output":0}},"fireworks-ai/kimi-k2.5-fw":{"id":"fireworks-ai/kimi-k2.5-fw","name":"Kimi-K2.5-FW","description":"Kimi multimodal agent model for visual understanding, coding, and planning","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":245760,"output":16384},"cost":{"input":0,"output":0}},"novita/glm-4.6v":{"id":"novita/glm-4.6v","name":"glm-4.6v","description":"GLM vision model for visual reasoning, documents, and multimodal agents","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":false,"release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":32768}},"novita/kimi-k2.6":{"id":"novita/kimi-k2.6","name":"Kimi-K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-20","last_updated":"2026-05-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":262144},"cost":{"input":0.96,"output":4.04,"cache_read":0.16}},"novita/glm-4.7":{"id":"novita/glm-4.7","name":"glm-4.7","description":"Legacy model retained for compatibility with older integrations","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":205000,"output":131072},"status":"deprecated"},"novita/minimax-m2.1":{"id":"novita/minimax-m2.1","name":"minimax-m2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":false,"release_date":"2025-12-26","last_updated":"2025-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":205000,"output":131072}},"novita/glm-4.7-flash":{"id":"novita/glm-4.7-flash","name":"glm-4.7-flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":false,"release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":65500}},"novita/glm-5":{"id":"novita/glm-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":205000,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2}},"novita/deepseek-v3.2":{"id":"novita/deepseek-v3.2","name":"DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":0},"cost":{"input":0.27,"output":0.4,"cache_read":0.13}},"novita/glm-4.6":{"id":"novita/glm-4.6","name":"GLM-4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0}},"novita/kimi-k2.5":{"id":"novita/kimi-k2.5","name":"Kimi-K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":262144},"cost":{"input":0.6,"output":3,"cache_read":0.1}},"novita/glm-4.7-n":{"id":"novita/glm-4.7-n","name":"glm-4.7-n","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":false,"release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":205000,"output":131072}},"novita/kimi-k2-thinking":{"id":"novita/kimi-k2-thinking","name":"kimi-k2-thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":false,"release_date":"2025-11-07","last_updated":"2025-11-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":0}},"stabilityai/stablediffusionxl":{"id":"stabilityai/stablediffusionxl","name":"StableDiffusionXL","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"stable-diffusion","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2023-07-09","last_updated":"2023-07-09","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":200,"output":0}},"trytako/tako":{"id":"trytako/tako","name":"Tako","description":"Tool-capable chat model for instruction following and agentic application workflows","family":"tako","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-08-15","last_updated":"2024-08-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2048,"output":0}},"empiriolabs/deepseek-v4-pro-el":{"id":"empiriolabs/deepseek-v4-pro-el","name":"DeepSeek-V4-Pro-EL","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"release_date":"2026-04-24","last_updated":"2026-05-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":384000},"cost":{"input":1.67,"output":3.33}},"empiriolabs/deepseek-v4-flash-el":{"id":"empiriolabs/deepseek-v4-flash-el","name":"DeepSeek-V4-Flash-EL","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"release_date":"2026-04-24","last_updated":"2026-05-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":384000},"cost":{"input":0.14,"output":0.28}},"google/gemini-deep-research":{"id":"google/gemini-deep-research","name":"gemini-deep-research","description":"Legacy model retained for compatibility with older integrations","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":0},"status":"deprecated","cost":{"input":1.6,"output":9.6}},"google/gemini-3-flash":{"id":"google/gemini-3-flash","name":"Gemini-3-Flash","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.404,"output":2.4242,"cache_read":0.0404},"canonical_model_id":"google/gemini-3-flash-preview"},"google/imagen-4":{"id":"google/imagen-4","name":"Imagen-4","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"imagen","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":480,"output":0}},"google/veo-2":{"id":"google/veo-2","name":"Veo-2","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-12-02","last_updated":"2024-12-02","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":0}},"google/imagen-4-fast":{"id":"google/imagen-4-fast","name":"Imagen-4-Fast","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"imagen","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-06-25","last_updated":"2025-06-25","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":480,"output":0}},"google/gemini-3-pro":{"id":"google/gemini-3-pro","name":"Gemini-3-Pro","description":"Legacy model retained for compatibility with older integrations","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-10-22","last_updated":"2025-10-22","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":1.6,"output":9.6,"cache_read":0.16}},"google/veo-3.1":{"id":"google/veo-3.1","name":"Veo-3.1","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":0}},"google/imagen-4-ultra":{"id":"google/imagen-4-ultra","name":"Imagen-4-Ultra","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"imagen","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-05-24","last_updated":"2025-05-24","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":480,"output":0}},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini-2.5-Flash-Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"temperature":false,"release_date":"2025-06-19","last_updated":"2025-06-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1024000,"output":64000},"cost":{"input":0.07,"output":0.28}},"google/lyria":{"id":"google/lyria","name":"Lyria","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"lyria","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-06-04","last_updated":"2025-06-04","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini-2.5-Pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":32768}],"tool_call":true,"temperature":false,"release_date":"2025-02-05","last_updated":"2025-02-05","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1065535,"output":65535},"cost":{"input":0.87,"output":7,"cache_read":0.087}},"google/gemma-4-31b":{"id":"google/gemma-4-31b","name":"Gemma-4-31B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":8192},"cost":{"input":0,"output":0}},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini-3.1-Flash-Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.2525,"output":1.5152},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.1-pro":{"id":"google/gemini-3.1-pro","name":"Gemini-3.1-Pro","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2.0202,"output":12.1212,"cache_read":0.202},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini-3.7-Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.7576,"output":3.7879,"cache_read":0.0758},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-omni-1.1-flash":{"id":"google/gemini-omni-1.1-flash","type":"video","name":"Gemini-Omni-1.1-Flash","description":"Video generation and editing model for scene extension, frame interpolation, and reference-guided multimodal control","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":false,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":131072,"output":57920},"cost":{"input":1.5152,"output":9.0909},"canonical_model_id":"google/gemini-omni-1.1-flash"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini-3.8-Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.7576,"output":3.7879,"cache_read":0.0758},"canonical_model_id":"google/gemini-3.8-flash"},"google/nano-banana-pro":{"id":"google/nano-banana-pro","name":"Nano-Banana-Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"nano-banana","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":65536,"output":0},"cost":{"input":2,"output":12,"cache_read":0.2}},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini-3.5-Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5152,"output":9.0909,"cache_read":0.1515},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-2.0-flash":{"id":"google/gemini-2.0-flash","name":"Gemini-2.0-Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":990000,"output":8192},"cost":{"input":0.1,"output":0.42}},"google/veo-3-fast":{"id":"google/veo-3-fast","name":"Veo-3-Fast","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-10-13","last_updated":"2025-10-13","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":0}},"google/imagen-3-fast":{"id":"google/imagen-3-fast","name":"Imagen-3-Fast","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"imagen","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-10-17","last_updated":"2024-10-17","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":480,"output":0}},"google/nano-banana":{"id":"google/nano-banana","name":"Nano-Banana","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"nano-banana","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":0},"cost":{"input":0.21,"output":1.8,"cache_read":0.021}},"google/imagen-3":{"id":"google/imagen-3","name":"Imagen-3","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"imagen","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-10-15","last_updated":"2024-10-15","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":480,"output":0}},"google/veo-3.1-fast":{"id":"google/veo-3.1-fast","name":"Veo-3.1-Fast","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":0}},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini-3.5-Flash-Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.303,"output":2.5253,"cache_read":0.0303},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini-2.5-Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"temperature":false,"release_date":"2025-04-26","last_updated":"2025-04-26","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1065535,"output":65535},"cost":{"input":0.21,"output":1.8,"cache_read":0.021}},"google/gemini-2.0-flash-lite":{"id":"google/gemini-2.0-flash-lite","name":"Gemini-2.0-Flash-Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-02-05","last_updated":"2025-02-05","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":990000,"output":8192},"cost":{"input":0.052,"output":0.21}},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini-3.6-Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.7576,"output":3.7879,"cache_read":0.0758},"canonical_model_id":"google/gemini-3.6-flash"},"google/veo-3":{"id":"google/veo-3","name":"Veo-3","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-05-21","last_updated":"2025-05-21","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":0}},"poetools/claude-code":{"id":"poetools/claude-code","name":"claude-code","description":"Claude model for careful reasoning, writing, coding, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-11-27","last_updated":"2025-11-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0}},"openai/o3-mini-high":{"id":"openai/o3-mini-high","name":"o3-mini-high","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-01-31","last_updated":"2025-01-31","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":0.99,"output":4}},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":0.99,"output":4,"cache_read":0.25}},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o-mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":124096,"output":4096},"cost":{"input":0.14,"output":0.54,"cache_read":0.068}},"openai/sora-2":{"id":"openai/sora-2","name":"Sora-2","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"sora","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-08","last_updated":"2026-04-08","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":4.5455,"output":27.2727,"cache_read":0.4545},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-3.5-turbo-instruct":{"id":"openai/gpt-3.5-turbo-instruct","name":"GPT-3.5-Turbo-Instruct","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2023-09-20","last_updated":"2023-09-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":3500,"output":1024},"cost":{"input":1.4,"output":1.8}},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1-nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.09,"output":0.36,"cache_read":0.022}},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1-mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.36,"output":1.4,"cache_read":0.09}},"openai/gpt-5-chat":{"id":"openai/gpt-5-chat","name":"GPT-5-Chat","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt-codex","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.1,"output":9,"cache_read":0.11}},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4-Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"temperature":false,"release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":27,"output":160}},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.6,"output":13,"cache_read":0.16}},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5-Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":14,"output":110}},"openai/chatgpt-4o-latest":{"id":"openai/chatgpt-4o-latest","name":"ChatGPT-4o-Latest","description":"Legacy model retained for compatibility with older integrations","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-08-14","last_updated":"2024-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":4.5,"output":14}},"openai/gpt-4o-aug":{"id":"openai/gpt-4o-aug","name":"GPT-4o-Aug","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-11-21","last_updated":"2024-11-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":2.2,"output":9,"cache_read":1.1}},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.1,"output":9,"cache_read":0.11}},"openai/gpt-4-classic-0314":{"id":"openai/gpt-4-classic-0314","name":"GPT-4-Classic-0314","description":"Legacy model retained for compatibility with older integrations","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-08-26","last_updated":"2024-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":4096},"status":"deprecated","cost":{"input":27,"output":54}},"openai/gpt-3.5-turbo-raw":{"id":"openai/gpt-3.5-turbo-raw","name":"GPT-3.5-Turbo-Raw","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2023-09-27","last_updated":"2023-09-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":4524,"output":2048},"cost":{"input":0.45,"output":1.4}},"openai/o3-pro":{"id":"openai/o3-pro","name":"o3-pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":18,"output":72}},"openai/o3-deep-research":{"id":"openai/o3-deep-research","name":"o3-deep-research","description":"Research model for long-horizon investigation, synthesis, and analytical reports","family":"o","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-06-27","last_updated":"2025-06-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":9,"output":36,"cache_read":2.2}},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2025-01-31","last_updated":"2025-01-31","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":0.99,"output":4}},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5-Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2023-09-13","last_updated":"2023-09-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":2048},"cost":{"input":0.45,"output":1.4}},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":1.8,"output":7.2,"cache_read":0.45}},"openai/gpt-image-1":{"id":"openai/gpt-image-1","name":"GPT-Image-1","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-03-31","last_updated":"2025-03-31","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":128000,"output":0}},"openai/gpt-5-codex":{"id":"openai/gpt-5-codex","name":"GPT-5-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.1,"output":9}},"openai/gpt-image-1.5":{"id":"openai/gpt-image-1.5","name":"gpt-image-1.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":128000,"output":0}},"openai/gpt-4-classic":{"id":"openai/gpt-4-classic","name":"GPT-4-Classic","description":"Legacy model retained for compatibility with older integrations","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-03-25","last_updated":"2024-03-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":4096},"status":"deprecated","cost":{"input":27,"output":54}},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5-mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2025-06-25","last_updated":"2025-06-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.22,"output":1.8,"cache_read":0.022}},"openai/gpt-4o-search":{"id":"openai/gpt-4o-search","name":"GPT-4o-Search","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-03-11","last_updated":"2025-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":2.2,"output":9}},"openai/gpt-image-2":{"id":"openai/gpt-image-2","name":"GPT-Image-2","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5.0505,"output":32.3232,"cache_read":1.2626}},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text","image","pdf"],"output":["image"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.2,"output":14,"cache_read":0.22}},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-05-13","last_updated":"2024-05-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192}},"openai/o4-mini-deep-research":{"id":"openai/o4-mini-deep-research","name":"o4-mini-deep-research","description":"Research model for long-horizon investigation, synthesis, and analytical reports","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-06-27","last_updated":"2025-06-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.8,"output":7.2,"cache_read":0.45}},"openai/gpt-image-1-mini":{"id":"openai/gpt-image-1-mini","name":"GPT-Image-1-Mini","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0}},"openai/gpt-5.3-codex-spark":{"id":"openai/gpt-5.3-codex-spark","name":"GPT-5.3-Codex-Spark","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2026-03-04","last_updated":"2026-03-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0,"output":0}},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"release_date":"2026-01-14","last_updated":"2026-01-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.6,"output":13,"cache_read":0.16}},"openai/o3":{"id":"openai/o3","name":"o3","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.8,"output":7.2,"cache_read":0.45}},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2024-12-18","last_updated":"2024-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":14,"output":54}},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1-Codex-Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2025-11-12","last_updated":"2025-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.22,"output":1.8,"cache_read":0.022}},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT-5.1-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2025-11-12","last_updated":"2025-11-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.1,"output":9,"cache_read":0.11}},"openai/gpt-5.3-instant":{"id":"openai/gpt-5.3-instant","name":"GPT-5.3-Instant","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":111616,"output":16384},"cost":{"input":1.6,"output":13,"cache_read":0.16}},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2-Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"temperature":false,"release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":19,"output":150}},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4-Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2023-09-13","last_updated":"2023-09-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":9,"output":27}},"openai/sora-2-pro":{"id":"openai/sora-2-pro","name":"Sora-2-Pro","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"sora","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0}},"openai/dall-e-3":{"id":"openai/dall-e-3","name":"DALL-E-3","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"dall-e","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":800,"output":0}},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4-Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.18,"output":1.1,"cache_read":0.018}},"openai/gpt-4o-mini-search":{"id":"openai/gpt-4o-mini-search","name":"GPT-4o-mini-Search","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-03-11","last_updated":"2025-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.14,"output":0.54}},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2025-11-12","last_updated":"2025-11-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.1,"output":9,"cache_read":0.11}},"openai/gpt-5.1-codex-max":{"id":"openai/gpt-5.1-codex-max","name":"GPT-5.1-Codex-Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.1,"output":9,"cache_read":0.11}},"openai/o1-pro":{"id":"openai/o1-pro","name":"o1-pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2025-03-19","last_updated":"2025-03-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":140,"output":540}},"openai/gpt-5.2-instant":{"id":"openai/gpt-5.2-instant","name":"GPT-5.2-Instant","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.6,"output":13,"cache_read":0.16}},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4-Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"release_date":"2026-03-12","last_updated":"2026-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.68,"output":4,"cache_read":0.068}},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5-Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-08","last_updated":"2026-04-08","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":27.2727,"output":163.6364},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5-nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.045,"output":0.36,"cache_read":0.0045}},"openai/gpt-5.1-instant":{"id":"openai/gpt-5.1-instant","name":"GPT-5.1-Instant","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-11-12","last_updated":"2025-11-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.1,"output":9,"cache_read":0.11}},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"release_date":"2026-02-10","last_updated":"2026-02-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.6,"output":13,"cache_read":0.16}},"runwayml/runway-gen-4-turbo":{"id":"runwayml/runway-gen-4-turbo","name":"Runway-Gen-4-Turbo","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"runway","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-05-09","last_updated":"2025-05-09","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":256,"output":0}},"runwayml/runway":{"id":"runwayml/runway","name":"Runway","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"runway","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-10-11","last_updated":"2024-10-11","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":256,"output":0}},"ideogramai/ideogram":{"id":"ideogramai/ideogram","name":"Ideogram","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"ideogram","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-04-03","last_updated":"2024-04-03","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":150,"output":0}},"ideogramai/ideogram-v2a":{"id":"ideogramai/ideogram-v2a","name":"Ideogram-v2a","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"ideogram","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-02-27","last_updated":"2025-02-27","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":150,"output":0}},"ideogramai/ideogram-v2":{"id":"ideogramai/ideogram-v2","name":"Ideogram-v2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"ideogram","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-08-21","last_updated":"2024-08-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":150,"output":0}},"ideogramai/ideogram-v2a-turbo":{"id":"ideogramai/ideogram-v2a-turbo","name":"Ideogram-v2a-Turbo","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"ideogram","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-02-27","last_updated":"2025-02-27","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":150,"output":0}},"cerebras/gpt-oss-120b-cs":{"id":"cerebras/gpt-oss-120b-cs","name":"GPT-OSS-120B-CS","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2025-08-06","last_updated":"2025-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":0},"cost":{"input":0.35,"output":0.75}},"cerebras/llama-3.3-70b-cs":{"id":"cerebras/llama-3.3-70b-cs","name":"llama-3.3-70b-cs","description":"Legacy model retained for compatibility with older integrations","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-05-13","last_updated":"2025-05-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"status":"deprecated"},"cerebras/qwen3-235b-2507-cs":{"id":"cerebras/qwen3-235b-2507-cs","name":"qwen3-235b-2507-cs","description":"Legacy model retained for compatibility with older integrations","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-08-06","last_updated":"2025-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"status":"deprecated"},"cerebras/llama-3.1-8b-cs":{"id":"cerebras/llama-3.1-8b-cs","name":"Llama-3.1-8B-CS","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-05-13","last_updated":"2025-05-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":0},"cost":{"input":0.1,"output":0.1}},"cerebras/qwen3-32b-cs":{"id":"cerebras/qwen3-32b-cs","name":"qwen3-32b-cs","description":"Legacy model retained for compatibility with older integrations","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-05-15","last_updated":"2025-05-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"status":"deprecated"}}},"oci":{"id":"oci","env":["OCI_GENAI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.generativeai.us-chicago-1.oci.oraclecloud.com/openai/v1","name":"OCI Generative AI","doc":"https://docs.oracle.com/en-us/iaas/Content/generative-ai/pretrained-models.htm","models":{"xai.grok-4.20-reasoning":{"id":"xai.grok-4.20-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"xai.grok-4.20-non-reasoning":{"id":"xai.grok-4.20-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"openai.gpt-oss-120b":{"id":"openai.gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"xai.grok-4.3":{"id":"xai.grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"xai.grok-4.6":{"id":"xai.grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"meta.llama-4-maverick-17b-128e-instruct-fp8":{"id":"meta.llama-4-maverick-17b-128e-instruct-fp8","name":"Llama 4 Maverick 17B Instruct","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":16384},"cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"openai.gpt-oss-20b":{"id":"openai.gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.07,"output":0.3},"canonical_model_id":"openai/gpt-oss-20b"},"meta.llama-4-scout-17b-16e-instruct":{"id":"meta.llama-4-scout-17b-16e-instruct","name":"Llama 4 Scout 17B Instruct","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":192000,"output":16384},"cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"meta.llama-3.3-70b-instruct":{"id":"meta.llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-3.3-70b-instruct"}}},"perplexity-agent":{"id":"perplexity-agent","env":["PERPLEXITY_API_KEY"],"npm":"@ai-sdk/openai","api":"https://api.perplexity.ai/v1","name":"Perplexity Agent","doc":"https://docs.perplexity.ai/docs/agent-api/models","models":{"moonshot-ai/kimi-k3":{"id":"moonshot-ai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-30","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshot-ai/kimi-k2.7-code":{"id":"moonshot-ai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-07-30","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"anthropic/claude-haiku-4-5":{"id":"anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1}},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5}},"anthropic/claude-sonnet-4-5":{"id":"anthropic/claude-sonnet-4-5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3}},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3}},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5}},"anthropic/claude-opus-4-5":{"id":"anthropic/claude-opus-4-5","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5}},"xai/grok-4.6":{"id":"xai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"xai/grok-4-1-fast-non-reasoning":{"id":"xai/grok-4-1-fast-non-reasoning","name":"Grok 4.1 Fast (Non-Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.13,"output":0.26,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"nvidia/nemotron-3-super-120b-a12b":{"id":"nvidia/nemotron-3-super-120b-a12b","name":"Nemotron 3 Super 120B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32000},"cost":{"input":0.25,"output":2.5},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"perplexity/sonar":{"id":"perplexity/sonar","name":"Sonar","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.25,"output":2.5,"cache_read":0.0625}},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"tiers":[{"input":0.5,"output":3,"cache_read":0.05,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":0.5,"output":3,"cache_read":0.05}}},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-03-20","last_updated":"2025-06-05","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}}},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}}},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-03-20","last_updated":"2025-06-05","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03}},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5}},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25}},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}}}},"302ai":{"id":"302ai","env":["302AI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.302.ai/v1","name":"302.AI","doc":"https://doc.302.ai","models":{"glm-4.6v":{"id":"glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.145,"output":0.43}},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.075,"output":0.25},"canonical_model_id":"zhipuai/glm-5.3-flash"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30},"canonical_model_id":"openai/gpt-5.5"},"gpt-5.6-luna-pro":{"id":"gpt-5.6-luna-pro","name":"gpt-5.6-luna-pro","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2},"canonical_model_id":"openai/gpt-5.6-luna"},"gpt-5.3-chat-latest":{"id":"gpt-5.3-chat-latest","name":"GPT-5.3 Chat (latest)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.75,"output":14},"canonical_model_id":"openai/gpt-5.3-chat-latest"},"gpt-5-thinking":{"id":"gpt-5-thinking","name":"gpt-5-thinking","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-08-08","last_updated":"2025-08-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.25,"output":10}},"claude-opus-4-1-20250805":{"id":"claude-opus-4-1-20250805","name":"claude-opus-4-1-20250805","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75}},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"claude-haiku-4-5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-16","last_updated":"2025-10-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5}},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"gpt-4.1-nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4}},"claude-sonnet-4-6-thinking":{"id":"claude-sonnet-4-6-thinking","name":"claude-sonnet-4-6-thinking","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-08","release_date":"2026-02-18","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15}},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"gpt-4.1-mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6}},"gemini-3-pro-image-preview":{"id":"gemini-3-pro-image-preview","name":"gemini-3-pro-image-preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-06","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":64000},"cost":{"input":2,"output":120}},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"gemini-3-flash-preview","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3}},"gemini-3.1-flash-image-preview":{"id":"gemini-3.1-flash-image-preview","name":"gemini-3.1-flash-image-preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-27","last_updated":"2026-02-27","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":60}},"gpt-5.2":{"id":"gpt-5.2","name":"gpt-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-12","last_updated":"2025-12-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14}},"gpt-5-pro":{"id":"gpt-5-pro","name":"gpt-5-pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-08","last_updated":"2025-10-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":15,"output":120}},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50},"canonical_model_id":"anthropic/claude-fable-5"},"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.72,"output":2.88},"canonical_model_id":"minimax/MiniMax-M3"},"claude-opus-4-7-thinking":{"id":"claude-opus-4-7-thinking","name":"claude-opus-4-7-thinking","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-7"},"qwen3.5-plus":{"id":"qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.12,"output":0.69},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.283,"output":1.705},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4},"canonical_model_id":"moonshotai/kimi-k2.6"},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5},"canonical_model_id":"xai/grok-4.3"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"openai/gpt-6-luna"},"gpt-5":{"id":"gpt-5","name":"gpt-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-08","last_updated":"2025-08-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10}},"gemini-2.5-flash-preview-09-2025":{"id":"gemini-2.5-flash-preview-09-2025","name":"gemini-2.5-flash-preview-09-2025","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-09-26","last_updated":"2025-09-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":2.5}},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15},"canonical_model_id":"moonshotai/kimi-k3"},"claude-opus-4-5-20251101":{"id":"claude-opus-4-5-20251101","name":"claude-opus-4-5-20251101","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25}},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":1.8},"canonical_model_id":"alibaba/qwen3.6-plus"},"deepseek-v3.2-thinking":{"id":"deepseek-v3.2-thinking","name":"DeepSeek-V3.2-Thinking","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.29,"output":0.43}},"qwen3-235b-a22b-instruct-2507":{"id":"qwen3-235b-a22b-instruct-2507","name":"qwen3-235b-a22b-instruct-2507","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":65536},"cost":{"input":0.29,"output":1.143}},"glm-4.5v":{"id":"glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-12","last_updated":"2025-08-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":16384},"cost":{"input":0.29,"output":0.86}},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12},"canonical_model_id":"openai/gpt-5.6-terra"},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.188,"output":1.133},"canonical_model_id":"alibaba/qwen3.6-flash"},"grok-4-fast-non-reasoning":{"id":"grok-4-fast-non-reasoning","name":"grok-4-fast-non-reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5}},"qwen3-235b-a22b":{"id":"qwen3-235b-a22b","name":"Qwen3-235B-A22B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.29,"output":2.86}},"doubao-seed-1-6-thinking-250715":{"id":"doubao-seed-1-6-thinking-250715","name":"doubao-seed-1-6-thinking-250715","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":16000},"cost":{"input":0.121,"output":1.21}},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"gemini-2.5-pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.25,"output":10}},"gpt-4.1":{"id":"gpt-4.1","name":"gpt-4.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8}},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.18,"output":0.564},"canonical_model_id":"alibaba/qwen3.8-flash"},"MiniMax-M2":{"id":"MiniMax-M2","name":"MiniMax-M2","description":"MiniMax model for chat, coding, office work, and agentic tasks","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-10-26","last_updated":"2025-10-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.33,"output":1.32}},"gemini-3.5-flash-thinking":{"id":"gemini-3.5-flash-thinking","name":"gemini-3.5-flash-thinking","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9},"canonical_model_id":"google/gemini-3.5-flash"},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.5"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50},"canonical_model_id":"anthropic/claude-fable-5-1"},"qwen3-30b-a3b":{"id":"qwen3-30b-a3b","name":"Qwen3-30B-A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.11,"output":1.08}},"claude-sonnet-4-5-20250929":{"id":"claude-sonnet-4-5-20250929","name":"claude-sonnet-4-5-20250929","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15}},"gpt-5-mini":{"id":"gpt-5-mini","name":"gpt-5-mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-08","last_updated":"2025-08-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2}},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"gemini-2.5-flash-nothink":{"id":"gemini-2.5-flash-nothink","name":"gemini-2.5-flash-nothink","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-24","last_updated":"2025-06-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":2.5}},"grok-4-fast-reasoning":{"id":"grok-4-fast-reasoning","name":"grok-4-fast-reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5}},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.285,"output":1.15},"canonical_model_id":"alibaba/qwen3.7-plus"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75},"canonical_model_id":"google/gemini-3.7-flash"},"claude-opus-5-thinking":{"id":"claude-opus-5-thinking","name":"claude-opus-5-thinking","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2},"canonical_model_id":"openai/gpt-5.6-luna"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12},"canonical_model_id":"google/gemini-3.1-pro-preview"},"gpt-5.1-chat-latest":{"id":"gpt-5.1-chat-latest","name":"gpt-5.1-chat-latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-14","last_updated":"2025-11-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.25,"output":10}},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4},"canonical_model_id":"zhipuai/glm-5.3"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75},"canonical_model_id":"google/gemini-3.8-flash"},"glm-4.7":{"id":"glm-4.7","name":"glm-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.286,"output":1.142}},"grok-4.20-beta-0309-reasoning":{"id":"grok-4.20-beta-0309-reasoning","name":"grok-4.20-beta-0309-reasoning","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6}},"gpt-5.4":{"id":"gpt-5.4","name":"gpt-5.4","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":0,"tiers":[{"input":5,"output":22.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5}}},"gpt-4o":{"id":"gpt-4o","name":"gpt-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-05-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10}},"claude-opus-4-1-20250805-thinking":{"id":"claude-opus-4-1-20250805-thinking","name":"claude-opus-4-1-20250805-thinking","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-05-27","last_updated":"2025-05-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75}},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4},"canonical_model_id":"zhipuai/glm-5.1"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9},"canonical_model_id":"google/gemini-3.5-flash"},"deepseek-flash":{"id":"deepseek-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-8"},"qwen3-max-2025-09-23":{"id":"qwen3-max-2025-09-23","name":"qwen3-max-2025-09-23","description":"Flagship model for demanding analysis, coding, and production agent workflows","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-24","last_updated":"2025-09-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":258048,"output":65536},"cost":{"input":0.86,"output":3.43}},"gemini-2.5-flash-lite-preview-09-2025":{"id":"gemini-2.5-flash-lite-preview-09-2025","name":"gemini-2.5-flash-lite-preview-09-2025","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-09-26","last_updated":"2025-09-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1,"output":0.4}},"o3":{"id":"o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8},"canonical_model_id":"openai/o3"},"gpt-5.6-terra-pro":{"id":"gpt-5.6-terra-pro","name":"gpt-5.6-terra-pro","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12},"canonical_model_id":"openai/gpt-5.6-terra"},"qwen3.5-35b-a3b":{"id":"qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.06,"output":0.46},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":4,"output":10,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-4-argon"},"qwen3-coder-480b-a35b-instruct":{"id":"qwen3-coder-480b-a35b-instruct","name":"qwen3-coder-480b-a35b-instruct","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.86,"output":3.43}},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"cache_write":0.5},"canonical_model_id":"xai/grok-4.7"},"MiniMax-M1":{"id":"MiniMax-M1","name":"MiniMax-M1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-06-16","last_updated":"2025-06-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.132,"output":1.254}},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50},"canonical_model_id":"openai/gpt-6-astra"},"mistral-large-2512":{"id":"mistral-large-2512","name":"mistral-large-2512","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":262144},"cost":{"input":1.1,"output":3.3}},"grok-4.1":{"id":"grok-4.1","name":"grok-4.1","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":2,"output":10}},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"claude-sonnet-4-6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-18","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15}},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6},"canonical_model_id":"xai/grok-4.5"},"doubao-seed-1-6-vision-250815":{"id":"doubao-seed-1-6-vision-250815","name":"doubao-seed-1-6-vision-250815","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.114,"output":1.143}},"glm-5":{"id":"glm-5","name":"glm-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.6}},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"claude-opus-4-7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}}},"gpt-5.6-sol-pro":{"id":"gpt-5.6-sol-pro","name":"gpt-5.6-sol-pro","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-5.2-chat-latest":{"id":"gpt-5.2-chat-latest","name":"gpt-5.2-chat-latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-12","last_updated":"2025-12-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.75,"output":14}},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"openai/gpt-6.1-sol"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"deepseek-v3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.29,"output":0.43}},"ministral-14b-2512":{"id":"ministral-14b-2512","name":"ministral-14b-2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.33,"output":0.33}},"glm-5v-turbo":{"id":"glm-5v-turbo","name":"GLM-5V-Turbo","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":0.72,"output":3.2}},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.7"},"doubao-seed-1-8-251215":{"id":"doubao-seed-1-8-251215","name":"doubao-seed-1-8-251215","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":224000,"output":64000},"cost":{"input":0.114,"output":0.286}},"glm-5-turbo":{"id":"glm-5-turbo","name":"glm-5-turbo","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":0.72,"output":3.2}},"kimi-k2-0905-preview":{"id":"kimi-k2-0905-preview","name":"kimi-k2-0905-preview","description":"Kimi model for long-context chat, coding, and agentic reasoning","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.632,"output":2.53}},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"gpt-5.4-nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-19","last_updated":"2026-03-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25}},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6},"canonical_model_id":"xai/grok-4.6"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gpt-5.1":{"id":"gpt-5.1","name":"gpt-5.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-14","last_updated":"2025-11-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10}},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"claude-haiku-4-5-20251001","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-16","last_updated":"2025-10-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5}},"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.8,"output":5.3},"canonical_model_id":"alibaba/qwen3.7-max"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-6-sol"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"anthropic/claude-haiku-5-5"},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"gemini-2.5-flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":2.5}},"gemini-2.0-flash-lite":{"id":"gemini-2.0-flash-lite","name":"gemini-2.0-flash-lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-11","release_date":"2025-06-16","last_updated":"2025-06-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":8192},"cost":{"input":0.075,"output":0.3}},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"gpt-5.4-mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-19","last_updated":"2026-03-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5}},"glm-4.6":{"id":"glm-4.6","name":"glm-4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.286,"output":1.142}},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4},"canonical_model_id":"zhipuai/glm-5.2"},"MiniMax-M2.1":{"id":"MiniMax-M2.1","name":"MiniMax-M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.1"},"glm-4.5":{"id":"glm-4.5","name":"GLM-4.5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-29","last_updated":"2025-07-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.286,"output":1.142}},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"google/gemini-3.6-flash"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.66,"output":3.3},"canonical_model_id":"moonshotai/kimi-k2.5"},"gemini-3.1-flash-lite-preview":{"id":"gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10},"canonical_model_id":"anthropic/claude-sonnet-5"},"qwen3.7-max-2026-06-08":{"id":"qwen3.7-max-2026-06-08","name":"qwen3.7-max-2026-06-08","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.8,"output":5.3},"canonical_model_id":"alibaba/qwen3.7-max"},"claude-sonnet-4-5-20250929-thinking":{"id":"claude-sonnet-4-5-20250929-thinking","name":"claude-sonnet-4-5-20250929-thinking","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15}},"grok-4-1-fast-reasoning":{"id":"grok-4-1-fast-reasoning","name":"grok-4-1-fast-reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5}},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.16,"output":6.36},"canonical_model_id":"alibaba/qwen3.8-max"},"gemini-2.5-flash-image":{"id":"gemini-2.5-flash-image","name":"gemini-2.5-flash-image","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-10-08","last_updated":"2025-10-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":30}},"kimi-k2-thinking":{"id":"kimi-k2-thinking","name":"kimi-k2-thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.575,"output":2.3}},"gemini-3-pro-preview":{"id":"gemini-3-pro-preview","name":"gemini-3-pro-preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2,"output":12}}}},"sakana":{"id":"sakana","env":["SAKANA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.sakana.ai/v1","name":"Sakana AI","doc":"https://console.sakana.ai/models","models":{"sakana-namazu":{"id":"sakana-namazu","name":"Sakana Namazu","description":"Japanese-specialized reasoning model based on Kimi K2.6 and tuned for Japanese language, culture, and business workflows","family":"sakana-namazu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.95,"output":4,"cache_read":0.15},"canonical_model_id":"sakana/sakana-namazu"},"fugu-ultra-20260615":{"id":"fugu-ultra-20260615","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"provider":{"shape":"responses"},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra"},"fugu":{"id":"fugu","name":"Fugu","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"provider":{"shape":"responses"},"canonical_model_id":"sakana/fugu"},"fugu-ultra":{"id":"fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"provider":{"shape":"responses"},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra"}}},"tinfoil":{"id":"tinfoil","env":["TINFOIL_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.tinfoil.sh/v1","name":"Tinfoil","doc":"https://docs.tinfoil.sh","models":{"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":4,"output":20,"cache_read":0.8},"canonical_model_id":"moonshotai/kimi-k3"},"deepseek-v4-1-flash":{"id":"deepseek-v4-1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"status":"beta","cost":{"input":0.65,"output":1.45,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"nomic-embed-text":{"id":"nomic-embed-text","name":"Nomic Embed Text v1.5","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2024-02","last_updated":"2024-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":768},"cost":{"input":0.05,"output":0}},"llama3-3-70b":{"id":"llama3-3-70b","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":1.75,"output":2.75},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"glm-5-3":{"id":"glm-5-3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.8,"output":5.75,"cache_read":0.45},"canonical_model_id":"zhipuai/glm-5.3"},"gpt-oss-safeguard-120b":{"id":"gpt-oss-safeguard-120b","name":"gpt-oss-safeguard-120b","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-safeguard-120b"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"gpt-oss-120b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"gemma4-31b":{"id":"gemma4-31b","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.4,"output":1},"canonical_model_id":"google/gemma-4-31b-it"}}},"azure-cognitive-services":{"id":"azure-cognitive-services","env":["AZURE_COGNITIVE_SERVICES_RESOURCE_NAME","AZURE_COGNITIVE_SERVICES_API_KEY"],"npm":"@ai-sdk/azure","name":"Azure Cognitive Services","doc":"https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models","models":{"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"gpt-4o-mini":{"id":"gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"gpt-3.5-turbo-instruct":{"id":"gpt-3.5-turbo-instruct","name":"GPT-3.5 Turbo Instruct","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2021-08","release_date":"2023-09-21","last_updated":"2023-09-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4096,"output":4096},"status":"deprecated","cost":{"input":1.5,"output":2}},"claude-opus-4-1":{"id":"claude-opus-4-1","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25}},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"phi-4-mini":{"id":"phi-4-mini","name":"Phi-4-mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"phi","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.075,"output":0.3}},"ministral-3b":{"id":"ministral-3b","name":"Ministral 3B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-03","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.04,"output":0.04}},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"deepseek-v3.2-speciale":{"id":"deepseek-v3.2-speciale","name":"DeepSeek-V3.2-Speciale","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.58,"output":1.68}},"gpt-image-2.5-sunburst":{"id":"gpt-image-2.5-sunburst","type":"image","name":"GPT Image 2.5 Sunburst","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-sunburst"},"gpt-5.4-pro":{"id":"gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek-V4-Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":1.74,"output":3.48},"canonical_model_id":"deepseek/deepseek-v4-pro"},"gpt-image-2.5-flare":{"id":"gpt-image-2.5-flare","type":"image","name":"GPT Image 2.5 Flare","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-flare"},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":28,"cache_read":0.35},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"gpt-5-pro":{"id":"gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"beta","provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"gpt-3.5-turbo-0125":{"id":"gpt-3.5-turbo-0125","name":"GPT-3.5 Turbo 0125","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2021-08","release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":4096},"status":"deprecated","cost":{"input":0.5,"output":1.5}},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.71,"output":0.71}},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.95,"output":4},"canonical_model_id":"moonshotai/kimi-k2.6"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"text-embedding-3-large":{"id":"text-embedding-3-large","name":"text-embedding-3-large","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":3072},"cost":{"input":0.13,"output":0}},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":4,"output":24,"cache_read":0.4,"cache_write":5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"o3-mini":{"id":"o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-07-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}}},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"gpt-image-1":{"id":"gpt-image-1","type":"image","name":"GPT-Image-1","description":"OpenAI image model for production generation, edits, and brand-safe visual workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-24","last_updated":"2025-04-24","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":40,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-1"},"gpt-5-codex":{"id":"gpt-5-codex","name":"GPT-5-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"gpt-image-1.5":{"id":"gpt-image-1.5","type":"image","name":"GPT-Image-1.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":32,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-1.5"},"phi-4-multimodal":{"id":"phi-4-multimodal","name":"Phi-4-multimodal","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"phi","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.08,"output":0.32,"input_audio":4}},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"gpt-chat-latest":{"id":"gpt-chat-latest","name":"GPT Chat Latest","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-05-05","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"beta","cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5-instant"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"llama-4-maverick-17b-128e-instruct-fp8":{"id":"llama-4-maverick-17b-128e-instruct-fp8","name":"Llama 4 Maverick 17B 128E Instruct FP8","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":16384},"cost":{"input":0.25,"output":1}},"deepseek-r1":{"id":"deepseek-r1","name":"DeepSeek-R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-01-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"status":"deprecated","cost":{"input":1.35,"output":5.4}},"phi-4-reasoning":{"id":"phi-4-reasoning","name":"Phi-4-reasoning","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"phi","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.125,"output":0.5}},"model-router":{"id":"model-router","name":"Model Router","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"model-router","attachment":true,"reasoning":false,"tool_call":true,"release_date":"2025-05-19","last_updated":"2025-11-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":16384},"provider":{"npm":"@ai-sdk/azure","shape":"completions"},"cost":{"input":0.14,"output":0}},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"gpt-image-2":{"id":"gpt-image-2","type":"image","name":"GPT-Image-2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-31","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"anthropic/claude-opus-4-8"},"gpt-3.5-turbo-1106":{"id":"gpt-3.5-turbo-1106","name":"GPT-3.5 Turbo 1106","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2021-08","release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":4096},"status":"deprecated","cost":{"input":1,"output":2}},"text-embedding-ada-002":{"id":"text-embedding-ada-002","name":"text-embedding-ada-002","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2022-12-15","last_updated":"2022-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536},"cost":{"input":0.1,"output":0}},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-01-14","last_updated":"2026-01-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"o3":{"id":"o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"llama-4-scout-17b-16e-instruct":{"id":"llama-4-scout-17b-16e-instruct","name":"Llama 4 Scout 17B 16E Instruct","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.2,"output":0.78}},"o1":{"id":"o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":15,"output":60,"cache_read":7.5}},"gpt-5.1-codex-mini":{"id":"gpt-5.1-codex-mini","name":"GPT-5.1 Codex Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-14","last_updated":"2025-11-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"codestral-2501":{"id":"codestral-2501","name":"Codestral 25.01","description":"Mistral coding model for code completion, generation, and developer workflows","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-03","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.3,"output":0.9}},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-14","last_updated":"2025-11-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"status":"beta","canonical_model_id":"xai/grok-4.7"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek-V4-Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.19,"output":0.51},"canonical_model_id":"deepseek/deepseek-v4-flash"},"gpt-4-turbo":{"id":"gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"claude-mythos-5":{"id":"claude-mythos-5","name":"Claude Mythos 5","description":"Restricted Claude model for advanced cybersecurity and biology research workflows","family":"claude-mythos","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"beta","provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-mythos-5"},"mistral-medium-2505":{"id":"mistral-medium-2505","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.4,"output":2}},"grok-4-20-reasoning":{"id":"grok-4-20-reasoning","name":"Grok 4.20 (Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-04-08","last_updated":"2026-04-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":8192},"status":"beta","cost":{"input":2,"output":6}},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":8,"output":40,"cache_read":0.8,"cache_write":10},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"cohere-embed-v3-english":{"id":"cohere-embed-v3-english","name":"Embed v3 English","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"cohere-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-11-07","last_updated":"2023-11-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":1024},"cost":{"input":0.1,"output":0}},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.58,"output":1.68}},"grok-4-20-non-reasoning":{"id":"grok-4-20-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-04-08","last_updated":"2026-04-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":8192},"status":"beta","cost":{"input":2,"output":6}},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"status":"beta","cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"phi-4-reasoning-plus":{"id":"phi-4-reasoning-plus","name":"Phi-4-reasoning-plus","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"phi","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.125,"output":0.5}},"cohere-embed-v3-multilingual":{"id":"cohere-embed-v3-multilingual","name":"Embed v3 Multilingual","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"cohere-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-11-07","last_updated":"2023-11-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":1024},"cost":{"input":0.1,"output":0}},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-14","last_updated":"2025-11-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":2.5,"output":20,"cache_read":0.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"grok-4-1-fast-non-reasoning":{"id":"grok-4-1-fast-non-reasoning","name":"Grok 4.1 Fast (Non-Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-06-27","last_updated":"2025-06-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}},"text-embedding-3-small":{"id":"text-embedding-3-small","name":"text-embedding-3-small","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536},"cost":{"input":0.02,"output":0}},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"gpt-5.1-codex-max":{"id":"gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex-max"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":1.5,"output":9,"cache_read":0.15},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek-V4.1-Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"status":"beta","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"cohere-command-a":{"id":"cohere-command-a","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":2.5,"output":10}},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.6,"output":3},"canonical_model_id":"moonshotai/kimi-k2.5"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-11-24","last_updated":"2025-08-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005}},"cohere-embed-v-4-0":{"id":"cohere-embed-v-4-0","name":"Embed v4","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"cohere-embed","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":1536},"cost":{"input":0.12,"output":0}},"grok-4-1-fast-reasoning":{"id":"grok-4-1-fast-reasoning","name":"Grok 4.1 Fast (Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-06-27","last_updated":"2025-06-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}},"gpt-4-turbo-vision":{"id":"gpt-4-turbo-vision","name":"GPT-4 Turbo Vision","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":10,"output":30}},"phi-4-mini-reasoning":{"id":"phi-4-mini-reasoning","name":"Phi-4-mini-reasoning","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"phi","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.075,"output":0.3}},"codex-mini":{"id":"codex-mini","name":"Codex Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-04","release_date":"2025-05-16","last_updated":"2025-05-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.5,"output":6,"cache_read":0.375}},"phi-4":{"id":"phi-4","name":"Phi-4","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"phi","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.125,"output":0.5}},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"mistral-small-2503":{"id":"mistral-small-2503","name":"Mistral Small 3.1","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2025-03-01","last_updated":"2025-03-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":0.1,"output":0.3}}}},"cline-pass":{"id":"cline-pass","env":["CLINE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.cline.bot/api/v1","name":"ClinePass","doc":"https://docs.cline.bot/getting-started/clinepass","models":{"cline-pass/glm-5.3-flash":{"id":"cline-pass/glm-5.3-flash","name":"cline-pass/glm-5.3-flash","description":"Latest natively multimodal model in the GLM-5 series","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03}},"cline-pass/muse-spark-1.3-contributor":{"id":"cline-pass/muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"canonical_model_id":"meta/muse-spark-1.3"},"cline-pass/deepseek-v4-pro":{"id":"cline-pass/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.74,"output":3.48,"cache_read":0.0145},"canonical_model_id":"deepseek/deepseek-v4-pro"},"cline-pass/mimo-v2.5-pro":{"id":"cline-pass/mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.74,"output":3.48,"cache_read":0.0145},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"cline-pass/kimi-k2.6":{"id":"cline-pass/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"cline-pass/kimi-k3":{"id":"cline-pass/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"cline-pass/mimo-v2.5":{"id":"cline-pass/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.5"},"cline-pass/qwen3.7-plus":{"id":"cline-pass/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":1.6,"cache_read":0.04,"cache_write":0.5,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.12,"cache_write":1.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.12,"cache_write":1.5}},"canonical_model_id":"alibaba/qwen3.7-plus"},"cline-pass/glm-5.3":{"id":"cline-pass/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"cline-pass/mimo-v2.6-flash":{"id":"cline-pass/mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"cline-pass/deepseek-v4-flash":{"id":"cline-pass/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"cline-pass/kimi-k2.7-code":{"id":"cline-pass/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"cline-pass/minimax-m3":{"id":"cline-pass/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"cline-pass/mimo-v2.6-pro":{"id":"cline-pass/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"cline-pass/qwen3.7-max":{"id":"cline-pass/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.5,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.7-max"},"cline-pass/glm-5.2":{"id":"cline-pass/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"cline-pass/deepseek-v4.1-flash":{"id":"cline-pass/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"cline-pass/qwen3.8-max":{"id":"cline-pass/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"}}},"tokenrouter":{"id":"tokenrouter","env":["TOKENROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.tokenrouter.com/v1","name":"TokenRouter","doc":"https://www.tokenrouter.com/docs/tokenrouter-feature-guide/","models":{"z-ai/glm-5.3-free":{"id":"z-ai/glm-5.3-free","name":"GLM-5.3 (free)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3"}}},"xpersona":{"id":"xpersona","env":["XPERSONA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://www.xpersona.co/v1","name":"Xpersona","doc":"https://www.xpersona.co/docs","models":{"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":1.5,"output":12,"reasoning":12,"cache_read":0.15},"canonical_model_id":"openai/gpt-5.5"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":0.6,"output":3.7,"reasoning":3.7,"cache_read":0.06},"canonical_model_id":"anthropic/claude-haiku-4-5"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":18.5,"reasoning":18.5,"cache_read":0.3},"canonical_model_id":"anthropic/claude-fable-5"},"xpersona-gpt-5.5":{"id":"xpersona-gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-12-30","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":18,"reasoning":18,"cache_read":0.3},"canonical_model_id":"openai/gpt-5.5"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":372000,"output":128000},"cost":{"input":1.5,"output":2,"reasoning":2,"cache_read":0.15},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0.75,"output":6,"reasoning":6,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":1.55,"output":12.2,"reasoning":12.2,"cache_read":0.155},"canonical_model_id":"google/gemini-3.5-flash"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":1.5,"output":9.25,"reasoning":9.25,"cache_read":0.15},"canonical_model_id":"anthropic/claude-opus-4-8"},"xpersona-frieren-coder":{"id":"xpersona-frieren-coder","name":"Xpersona Frieren 1","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-30","release_date":"2026-05-01","last_updated":"2026-05-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":384000},"cost":{"input":1.5,"output":6,"reasoning":6,"cache_read":0.15}},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":0.9,"output":5.55,"reasoning":5.55,"cache_read":0.09},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":372000,"output":128000},"cost":{"input":1.5,"output":12,"reasoning":12,"cache_read":0.15},"canonical_model_id":"openai/gpt-5.6-sol"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":0.375,"output":4,"reasoning":4,"cache_read":0.0375},"canonical_model_id":"openai/gpt-5.4-mini"},"gpt-5.6":{"id":"gpt-5.6","name":"GPT-5.6","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":372000,"output":128000},"cost":{"input":1.5,"output":12,"reasoning":12,"cache_read":0.15},"canonical_model_id":"openai/gpt-5.6-sol"}}},"routing-run":{"id":"routing-run","env":["ROUTING_RUN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.routing.run/v1","name":"routing.run","doc":"https://docs.routing.run/api-reference/models","models":{"qwen3.5-9b":{"id":"qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32000},"cost":{"input":0.16,"output":0.48},"canonical_model_id":"alibaba/qwen3.5-9b"},"kimi-k2.6-nitro":{"id":"kimi-k2.6-nitro","name":"Kimi K2.6 Nitro","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"cost":{"input":0.275,"output":1.1},"canonical_model_id":"moonshotai/kimi-k2.6"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":64000},"cost":{"input":0.348,"output":0.696},"canonical_model_id":"deepseek/deepseek-v4-pro"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"cost":{"input":0.275,"output":1.1},"canonical_model_id":"moonshotai/kimi-k2.6"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":1.5,"output":9},"canonical_model_id":"openai/gpt-5.6-terra"},"nemotron-3-ultra":{"id":"nemotron-3-ultra","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32000},"cost":{"input":0.1,"output":0.1},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":0.7,"output":4.2},"canonical_model_id":"openai/gpt-5.6-luna"},"glm-5.2-nitro":{"id":"glm-5.2-nitro","name":"GLM 5.2 Nitro","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"cost":{"input":0.8,"output":2.4},"canonical_model_id":"zhipuai/glm-5.2"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-8"},"kimi-k2.7-code-nitro":{"id":"kimi-k2.7-code-nitro","name":"Kimi K2.7 Code Nitro","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"cost":{"input":0.275,"output":1.1},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":64000},"cost":{"input":0.112,"output":0.224},"canonical_model_id":"deepseek/deepseek-v4-flash"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"cost":{"input":0.275,"output":1.1},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15},"canonical_model_id":"openai/gpt-5.6-sol"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"cost":{"input":0.8,"output":2.4},"canonical_model_id":"zhipuai/glm-5.2"}}},"openai":{"id":"openai","env":["OPENAI_API_KEY"],"npm":"@ai-sdk/openai","name":"OpenAI","doc":"https://platform.openai.com/docs/models","models":{"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","experimental":{"modes":{"fast":{"cost":{"input":2,"output":8,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.55,"output":2.2,"cache_read":0.1375},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":1.1,"output":4.4,"cache_read":0.275}},"gpt-4o-mini":{"id":"gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"experimental":{"modes":{"fast":{"cost":{"input":0.25,"output":1,"cache_read":0.125},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.15,"output":0.6,"cache_read":0.075}},"gpt-4":{"id":"gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"status":"deprecated","cost":{"input":30,"output":60}},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}}},"gpt-daybreak-blue-latest":{"id":"gpt-daybreak-blue-latest","name":"Daybreak Blue","description":"Alias for flagship general-purpose models with safeguards for defensive cybersecurity work","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-08-07","last_updated":"2026-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","experimental":{"modes":{"fast":{"cost":{"input":0.2,"output":0.8,"cache_read":0.05},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.1,"output":0.4,"cache_read":0.025}},"gpt-daybreak-red-latest":{"id":"gpt-daybreak-red-latest","name":"Daybreak Red","description":"Alias for advanced cybersecurity models for authorized vulnerability research and security testing","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-08-07","last_updated":"2026-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":12.5,"output":75,"cache_read":1.25,"cache_write":15.625},"canonical_model_id":"openai/gpt-5.6-cyber"},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"experimental":{"modes":{"fast":{"cost":{"input":0.7,"output":2.8,"cache_read":0.175},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.4,"output":1.6,"cache_read":0.1}},"gpt-5.4-pro":{"id":"gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}}},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":28,"cache_read":0.35},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.875,"output":7,"cache_read":0.0875},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"gpt-5-pro":{"id":"gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":15,"output":120}},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":0.2,"output":1,"cache_read":0.02,"cache_write":0.25},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}}}},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":2.5,"output":20,"cache_read":0.25},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.625,"output":5,"cache_read":0.0625},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"text-embedding-3-large":{"id":"text-embedding-3-large","type":"embedding","name":"text-embedding-3-large","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-01","release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8191,"output":3072},"cost":{"input":0.13,"output":0}},"o3-pro":{"id":"o3-pro","name":"o3-pro","description":"High-effort o3 tier for difficult technical reasoning and careful answers","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":20,"output":80}},"gpt-realtime-2.1":{"id":"gpt-realtime-2.1","type":"realtime","name":"GPT-Realtime-2.1","description":"Realtime speech-to-speech model with configurable reasoning, tool use, and robust voice-agent behavior","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2024-09-30","release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text","audio","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":128000,"input":96000,"output":32000},"cost":{"input":4,"output":24,"cache_read":0.4,"input_audio":32,"output_audio":64},"canonical_model_id":"openai/gpt-realtime-2.1"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":4,"output":24,"cache_read":0.4,"cache_write":5},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}}}},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"o3-mini":{"id":"o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.55}},"gpt-3.5-turbo":{"id":"gpt-3.5-turbo","name":"GPT-3.5-turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"status":"deprecated","cost":{"input":0.5,"output":1.5,"cache_read":0}},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":14,"cache_read":0.875},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2,"output":8,"cache_read":0.5}},"gpt-image-1":{"id":"gpt-image-1","type":"image","name":"gpt-image-1","description":"OpenAI image model for production generation, edits, and brand-safe visual workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-24","last_updated":"2025-04-24","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"input":0,"output":0},"status":"deprecated","canonical_model_id":"openai/gpt-image-1"},"gpt-image-1.5":{"id":"gpt-image-1.5","type":"image","name":"gpt-image-1.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":0,"input":0,"output":0},"canonical_model_id":"openai/gpt-image-1.5"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":0.45,"output":3.6,"cache_read":0.045},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.125,"output":1,"cache_read":0.0125},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"gpt-4o-2024-08-06":{"id":"gpt-4o-2024-08-06","name":"GPT-4o (2024-08-06)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25}},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":0.4,"output":2.4,"cache_read":0.04,"cache_write":0.5},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}}}},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"gpt-image-2":{"id":"gpt-image-2","type":"image","name":"gpt-image-2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"input":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}}},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"experimental":{"modes":{"fast":{"cost":{"input":4.25,"output":17,"cache_read":2.125},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":10,"cache_read":1.25}},"gpt-image-1-mini":{"id":"gpt-image-1-mini","type":"image","name":"gpt-image-1-mini","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-09-26","last_updated":"2025-09-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":0,"input":0,"output":0}},"gpt-5.3-codex-spark":{"id":"gpt-5.3-codex-spark","name":"GPT-5.3 Codex Spark","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex-spark","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":100000,"output":32000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"text-embedding-ada-002":{"id":"text-embedding-ada-002","type":"embedding","name":"text-embedding-ada-002","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2022-12","release_date":"2022-12-15","last_updated":"2022-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536},"cost":{"input":0.1,"output":0}},"o3":{"id":"o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":14,"cache_read":0.875},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":1,"output":4,"cache_read":0.25},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":2,"output":8,"cache_read":0.5}},"o1":{"id":"o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":15,"output":60,"cache_read":7.5}},"gpt-5.2-pro":{"id":"gpt-5.2-pro","name":"GPT-5.2 Pro","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":21,"output":168}},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":20,"output":100,"cache_read":2,"cache_write":25},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}},"ultrafast":{"cost":{"input":60,"output":300,"cache_read":6,"cache_write":75},"provider":{"body":{"service_tier":"ultrafast"}}}}},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"gpt-4-turbo":{"id":"gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":10,"output":30}},"gpt-4o-2024-05-13":{"id":"gpt-4o-2024-05-13","name":"GPT-4o (2024-05-13)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-05-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","experimental":{"modes":{"fast":{"cost":{"input":8.75,"output":26.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":15}},"chatgpt-image-latest":{"id":"chatgpt-image-latest","type":"image","name":"chatgpt-image-latest","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":0,"input":0,"output":0}},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":8,"output":40,"cache_read":0.8,"cache_write":10},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}}}},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}},"ultrafast":{"cost":{"input":12,"output":60,"cache_read":0.6,"cache_write":15},"provider":{"body":{"service_tier":"ultrafast"}}}}},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"flex":{"cost":{"input":0.1,"output":0.625,"cache_read":0.01},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":0.2,"output":1.25,"cache_read":0.02}},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":2.5,"output":20,"cache_read":0.25},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.625,"output":5,"cache_read":0.0625},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}}}},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"text-embedding-3-small":{"id":"text-embedding-3-small","type":"embedding","name":"text-embedding-3-small","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-01","release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8191,"output":1536},"cost":{"input":0.02,"output":0}},"o1-pro":{"id":"o1-pro","name":"o1-pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2025-03-19","last_updated":"2025-03-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":150,"output":600}},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":1.5,"output":9,"cache_read":0.15},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.375,"output":2.25,"cache_read":0.0375},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":0.75,"output":4.5,"cache_read":0.075}},"gpt-5.5-pro":{"id":"gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}}},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"flex":{"cost":{"input":0.025,"output":0.2,"cache_read":0.0025},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":0.05,"output":0.4,"cache_read":0.005}},"gpt-4o-2024-11-20":{"id":"gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25}},"gpt-5.6":{"id":"gpt-5.6","name":"GPT-5.6","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":8,"output":40,"cache_read":0.8,"cache_write":10},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}}}},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":28,"cache_read":0.35},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.75,"output":14,"cache_read":0.175}}}},"synthetic":{"id":"synthetic","env":["SYNTHETIC_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.synthetic.new/openai/v1","name":"Synthetic","doc":"https://synthetic.new/pricing","models":{"hf:moonshotai/Kimi-K2.7-Code":{"id":"hf:moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.95,"output":4,"cache_read":0.95},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"hf:moonshotai/Kimi-K3":{"id":"hf:moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"cost":{"input":3,"output":15,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"hf:deepseek-ai/DeepSeek-V4.1-Flash":{"id":"hf:deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"cost":{"input":0.6,"output":1.2,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"hf:Qwen/Qwen3.6-27B":{"id":"hf:Qwen/Qwen3.6-27B","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.45,"output":3.6,"cache_read":0.45},"canonical_model_id":"alibaba/qwen3.6-27b"},"hf:openai/gpt-oss-120b":{"id":"hf:openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.1,"output":0.1,"cache_read":0.1}},"hf:MiniMaxAI/MiniMax-M3":{"id":"hf:MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"cost":{"input":0.6,"output":1.2,"cache_read":0.6},"canonical_model_id":"minimax/MiniMax-M3"},"hf:nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4":{"id":"hf:nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":1,"cache_read":0.3},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"hf:zai-org/GLM-5.3":{"id":"hf:zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"hf:zai-org/GLM-5.2":{"id":"hf:zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"cost":{"input":1.4,"output":4.4,"cache_read":1.4},"canonical_model_id":"zhipuai/glm-5.2"},"hf:zai-org/GLM-4.7-Flash":{"id":"hf:zai-org/GLM-4.7-Flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":65536},"cost":{"input":0.1,"output":0.5,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-4.7-flash"},"hf:zai-org/GLM-5.3-Flash":{"id":"hf:zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"cost":{"input":0.15,"output":0.5,"cache_read":0.04},"canonical_model_id":"zhipuai/glm-5.3-flash"}}},"morph":{"id":"morph","env":["MORPH_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.morphllm.com/v1","name":"Morph","doc":"https://docs.morphllm.com/api-reference/introduction","models":{"auto":{"id":"auto","name":"Auto","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-06-01","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"cost":{"input":0.85,"output":1.55}},"morph-v3-fast":{"id":"morph-v3-fast","name":"Morph v3 Fast","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-08-15","last_updated":"2024-08-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":16000},"cost":{"input":0.8,"output":1.2}},"morph-v3-large":{"id":"morph-v3-large","name":"Morph v3 Large","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-08-15","last_updated":"2024-08-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"cost":{"input":0.9,"output":1.9}}}},"xiaomi-token-plan-cn":{"id":"xiaomi-token-plan-cn","env":["XIAOMI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://token-plan-cn.xiaomimimo.com/v1","name":"Xiaomi Token Plan (China)","doc":"https://platform.xiaomimimo.com/#/docs","models":{"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"mimo-v2.5-tts-voicedesign":{"id":"mimo-v2.5-tts-voicedesign","name":"MiMo-V2.5-TTS-VoiceDesign","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0}},"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"mimo-v2.5-tts-voiceclone":{"id":"mimo-v2.5-tts-voiceclone","name":"MiMo-V2.5-TTS-VoiceClone","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0}},"mimo-v2-tts":{"id":"mimo-v2-tts","name":"MiMo-V2-TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0}},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"mimo-v2.5-tts":{"id":"mimo-v2.5-tts","name":"MiMo-V2.5-TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0}},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"mimo-v2-pro":{"id":"mimo-v2-pro","name":"MiMo-V2-Pro","description":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2-pro"}}},"sarvam":{"id":"sarvam","env":["SARVAM_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.sarvam.ai/v1","name":"Sarvam AI","doc":"https://docs.sarvam.ai/api-reference-docs/getting-started/models","models":{"sarvam-105b":{"id":"sarvam-105b","name":"Sarvam-105B","description":"Flagship Indian-language reasoning model for enterprise multilingual applications","family":"sarvam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":[null,"low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-18","last_updated":"2026-03-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072}},"sarvam-30b":{"id":"sarvam-30b","name":"Sarvam-30B","description":"Efficient Indian-language reasoning model for chat, coding, and multilingual work","family":"sarvam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":[null,"low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-18","last_updated":"2026-03-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536}}}},"scx-ai":{"id":"scx-ai","env":["SCX_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.scx.ai/v1","name":"SCX.ai","doc":"https://platform.scx.ai/docs","models":{"GLM-5.2":{"id":"GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.55,"output":1.784,"cache_read":0.111},"canonical_model_id":"zhipuai/glm-5.2"},"Qwen3.8-Max":{"id":"Qwen3.8-Max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":983616,"output":131072},"cost":{"input":1.815,"output":5.4461,"cache_read":0.17,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":196608},"cost":{"input":0.48,"output":1.79,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M2.7"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.17,"output":0.55},"canonical_model_id":"openai/gpt-oss-120b"}}},"inco":{"id":"inco","env":["INCO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.inco.ai/v1","name":"Inco","doc":"https://platform.inco.ai/docs","models":{"minimax-m3:fast":{"id":"minimax-m3:fast","name":"MiniMax M3 Fast","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.6,"output":2.4},"canonical_model_id":"minimax/MiniMax-M3"},"glm-5.3-flash:fast":{"id":"glm-5.3-flash:fast","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4},"canonical_model_id":"zhipuai/glm-5.3"},"deepseek-v4.1-flash:fast":{"id":"deepseek-v4.1-flash:fast","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.6,"output":2.4},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"glm-5.3:fast":{"id":"glm-5.3:fast","name":"GLM-5.3 Fast","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2.8,"output":8.8},"canonical_model_id":"zhipuai/glm-5.3"},"kimi-k3:fast":{"id":"kimi-k3:fast","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":6,"output":30},"canonical_model_id":"moonshotai/kimi-k3"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M3"}}},"auriko":{"id":"auriko","env":["AURIKO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.auriko.ai/v1","name":"Auriko","doc":"https://docs.auriko.ai","models":{"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.435,"output":0.87,"cache_read":0.003625},"canonical_model_id":"deepseek/deepseek-v4-pro"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"minimax-m2-7":{"id":"minimax-m2-7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"qwen-3.6-plus":{"id":"qwen-3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.1,"tiers":[{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5}},"canonical_model_id":"alibaba/qwen3.6-plus"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-2.5-flash"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":2.8},"canonical_model_id":"moonshotai/kimi-k2.5"},"minimax-m2-7-highspeed":{"id":"minimax-m2-7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"}}},"helicone":{"id":"helicone","env":["HELICONE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://ai-gateway.helicone.ai/v1","name":"Helicone","doc":"https://helicone.ai/models","models":{"o4-mini":{"id":"o4-mini","name":"OpenAI o4 Mini","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2024-06","release_date":"2024-06-01","last_updated":"2024-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275}},"gpt-4o-mini":{"id":"gpt-4o-mini","name":"OpenAI GPT-4o-mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075}},"claude-3.5-haiku":{"id":"claude-3.5-haiku","name":"Anthropic: Claude 3.5 Haiku","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":0.7999999999999999,"output":4,"cache_read":0.08,"cache_write":1}},"claude-4.5-haiku":{"id":"claude-4.5-haiku","name":"Anthropic: Claude 4.5 Haiku","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-10","release_date":"2025-10-01","last_updated":"2025-10-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":1,"output":5,"cache_read":0.09999999999999999,"cache_write":1.25}},"sonar-deep-research":{"id":"sonar-deep-research","name":"Perplexity Sonar Deep Research","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar-deep-research","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-01-27","last_updated":"2025-01-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":127000,"output":4096},"cost":{"input":2,"output":8}},"llama-3.1-8b-instruct":{"id":"llama-3.1-8b-instruct","name":"Meta Llama 3.1 8B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":16384},"cost":{"input":0.02,"output":0.049999999999999996}},"grok-code-fast-1":{"id":"grok-code-fast-1","name":"xAI Grok Code Fast 1","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2024-08-25","last_updated":"2024-08-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":10000},"cost":{"input":0.19999999999999998,"output":1.5,"cache_read":0.02}},"mistral-small":{"id":"mistral-small","name":"Mistral Small 3.2","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.075,"output":0.2},"canonical_model_id":"mistral/mistral-small-2506"},"qwen3-235b-a22b-thinking":{"id":"qwen3-235b-a22b-thinking","name":"Qwen3 235B A22B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":81920},"cost":{"input":0.3,"output":2.9000000000000004}},"claude-opus-4-1":{"id":"claude-opus-4-1","name":"Anthropic: Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"temperature":true,"knowledge":"2025-08","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"mistral-nemo":{"id":"mistral-nemo","name":"Mistral Nemo","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16400},"cost":{"input":20,"output":40}},"qwen3-vl-235b-a22b-instruct":{"id":"qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":16384},"cost":{"input":0.3,"output":1.5}},"claude-opus-4-1-20250805":{"id":"claude-opus-4-1-20250805","name":"Anthropic: Claude Opus 4.1 (20250805)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"temperature":true,"knowledge":"2025-08","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"OpenAI GPT-4.1 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.09999999999999999,"output":0.39999999999999997,"cache_read":0.024999999999999998}},"llama-guard-4":{"id":"llama-guard-4","name":"Meta Llama Guard 4 12B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":1024},"cost":{"input":0.21,"output":0.21}},"gemma-3-12b-it":{"id":"gemma-3-12b-it","name":"Google Gemma 3 12B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-12","release_date":"2024-12-01","last_updated":"2024-12-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.049999999999999996,"output":0.09999999999999999}},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"OpenAI GPT-4.1 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.39999999999999997,"output":1.5999999999999999,"cache_read":0.09999999999999999}},"kimi-k2-0905":{"id":"kimi-k2-0905","name":"Kimi K2 (09/05)","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":16384},"cost":{"input":0.5,"output":2,"cache_read":0.39999999999999997}},"gpt-5-pro":{"id":"gpt-5-pro","name":"OpenAI: GPT-5 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":15,"output":120}},"chatgpt-4o-latest":{"id":"chatgpt-4o-latest","name":"OpenAI ChatGPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2024-08-14","last_updated":"2024-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":5,"output":20,"cache_read":2.5}},"claude-sonnet-4":{"id":"claude-sonnet-4","name":"Anthropic: Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-14","last_updated":"2025-05-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.30000000000000004,"cache_write":3.75}},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Meta Llama 3.3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16400},"cost":{"input":0.13,"output":0.39}},"sonar-reasoning":{"id":"sonar-reasoning","name":"Perplexity Sonar Reasoning","description":"Web-grounded reasoning model for multi-step research and cited answers","family":"sonar-reasoning","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-01-27","last_updated":"2025-01-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":127000,"output":4096},"cost":{"input":1,"output":5}},"gpt-5":{"id":"gpt-5","name":"OpenAI GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.12500000000000003}},"sonar-pro":{"id":"sonar-pro","name":"Perplexity Sonar Pro","description":"Advanced Sonar search model for deeper research and cited synthesis","family":"sonar-pro","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-01-27","last_updated":"2025-01-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":4096},"cost":{"input":3,"output":15}},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"Google Gemini 2.5 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.09999999999999999,"output":0.39999999999999997,"cache_read":0.024999999999999998,"cache_write":0.09999999999999999}},"o3-pro":{"id":"o3-pro","name":"OpenAI o3 Pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2024-06","release_date":"2024-06-01","last_updated":"2024-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":20,"output":80}},"ernie-4.5-21b-a3b-thinking":{"id":"ernie-4.5-21b-a3b-thinking","name":"Baidu Ernie 4.5 21B A3B Thinking","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"ernie","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-03","release_date":"2025-03-16","last_updated":"2025-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8000},"cost":{"input":0.07,"output":0.28}},"grok-4-fast-non-reasoning":{"id":"grok-4-fast-non-reasoning","name":"xAI Grok 4 Fast Non-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-19","last_updated":"2025-09-19","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":0.19999999999999998,"output":0.5,"cache_read":0.049999999999999996}},"o3-mini":{"id":"o3-mini","name":"OpenAI o3 Mini","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2023-10","release_date":"2023-10-01","last_updated":"2023-10-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55}},"deepseek-v3":{"id":"deepseek-v3","name":"DeepSeek V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.56,"output":1.68,"cache_read":0.07}},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Google Gemini 2.5 Pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.3125,"cache_write":1.25}},"gpt-4.1":{"id":"gpt-4.1","name":"OpenAI GPT-4.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5}},"gpt-5-codex":{"id":"gpt-5-codex","name":"OpenAI: GPT-5 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.12500000000000003}},"claude-4.5-opus":{"id":"claude-4.5-opus","name":"Anthropic: Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"llama-4-maverick":{"id":"llama-4-maverick","name":"Meta Llama 4 Maverick 17B 128E","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.15,"output":0.6}},"qwen3-30b-a3b":{"id":"qwen3-30b-a3b","name":"Qwen3 30B A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-06-01","last_updated":"2025-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":41000,"output":41000},"cost":{"input":0.08,"output":0.29}},"claude-3.5-sonnet-v2":{"id":"claude-3.5-sonnet-v2","name":"Anthropic: Claude 3.5 Sonnet v2","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":3,"output":15,"cache_read":0.30000000000000004,"cache_write":3.75}},"claude-sonnet-4-5-20250929":{"id":"claude-sonnet-4-5-20250929","name":"Anthropic: Claude Sonnet 4.5 (20250929)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.30000000000000004,"cache_write":3.75}},"gpt-5-mini":{"id":"gpt-5-mini","name":"OpenAI GPT-5 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.024999999999999998}},"grok-4-fast-reasoning":{"id":"grok-4-fast-reasoning","name":"xAI: Grok 4 Fast Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":0.19999999999999998,"output":0.5,"cache_read":0.049999999999999996}},"llama-4-scout":{"id":"llama-4-scout","name":"Meta Llama 4 Scout 17B 16E","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.08,"output":0.3}},"llama-3.1-8b-instant":{"id":"llama-3.1-8b-instant","name":"Meta Llama 3.1 8B Instant","description":"Compact Llama instruction model for fast chat and local deployment","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32678},"cost":{"input":0.049999999999999996,"output":0.08}},"deepseek-tng-r1t2-chimera":{"id":"deepseek-tng-r1t2-chimera","name":"DeepSeek TNG R1T2 Chimera","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek-thinking","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-02","last_updated":"2025-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":130000,"output":163840},"cost":{"input":0.3,"output":1.2}},"qwen3-next-80b-a3b-instruct":{"id":"qwen3-next-80b-a3b-instruct","name":"Qwen3 Next 80B A3B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":16384},"cost":{"input":0.14,"output":1.4}},"gpt-5.1-chat-latest":{"id":"gpt-5.1-chat-latest","name":"OpenAI GPT-5.1 Chat","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-codex","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.25,"output":10,"cache_read":0.12500000000000003}},"qwen2.5-coder-7b-fast":{"id":"qwen2.5-coder-7b-fast","name":"Qwen2.5 Coder 7B fast","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-09","release_date":"2024-09-15","last_updated":"2024-09-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":8192},"cost":{"input":0.03,"output":0.09}},"gpt-4o":{"id":"gpt-4o","name":"OpenAI GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-05","release_date":"2024-05-13","last_updated":"2024-05-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25}},"gpt-4.1-mini-2025-04-14":{"id":"gpt-4.1-mini-2025-04-14","name":"OpenAI GPT-4.1 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.39999999999999997,"output":1.5999999999999999,"cache_read":0.09999999999999999}},"grok-3-mini":{"id":"grok-3-mini","name":"xAI Grok 3 Mini","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2024-06-01","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":0.3,"output":0.5,"cache_read":0.075}},"o1-mini":{"id":"o1-mini","name":"OpenAI: o1-mini","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":65536},"cost":{"input":1.1,"output":4.4,"cache_read":0.55}},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"OpenAI GPT-OSS 20b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2024-06-01","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":0.049999999999999996,"output":0.19999999999999998}},"llama-3.1-8b-instruct-turbo":{"id":"llama-3.1-8b-instruct-turbo","name":"Meta Llama 3.1 8B Instruct Turbo","description":"Compact Llama instruction model for fast chat and local deployment","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.02,"output":0.03}},"o3":{"id":"o3","name":"OpenAI o3","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2024-06","release_date":"2024-06-01","last_updated":"2024-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5}},"o1":{"id":"o1","name":"OpenAI: o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5}},"gpt-5.1-codex-mini":{"id":"gpt-5.1-codex-mini","name":"OpenAI: GPT-5.1 Codex Mini","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-codex","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.024999999999999998}},"kimi-k2-0711":{"id":"kimi-k2-0711","name":"Kimi K2 (07/11)","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"cost":{"input":0.5700000000000001,"output":2.3}},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"OpenAI: GPT-5.1 Codex","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-codex","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.12500000000000003}},"deepseek-v3.1-terminus":{"id":"deepseek-v3.1-terminus","name":"DeepSeek V3.1 Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-22","last_updated":"2025-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.27,"output":1,"cache_read":0.21600000000000003}},"sonar":{"id":"sonar","name":"Perplexity Sonar","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-01-27","last_updated":"2025-01-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":127000,"output":4096},"cost":{"input":1,"output":1}},"qwen3-32b":{"id":"qwen3-32b","name":"Qwen3 32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":40960},"cost":{"input":0.29,"output":0.59}},"llama-3.3-70b-versatile":{"id":"llama-3.3-70b-versatile","name":"Meta Llama 3.3 70B Versatile","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32678},"cost":{"input":0.59,"output":0.7899999999999999}},"claude-4.5-sonnet":{"id":"claude-4.5-sonnet","name":"Anthropic: Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.30000000000000004,"cache_write":3.75}},"deepseek-r1-distill-llama-70b":{"id":"deepseek-r1-distill-llama-70b","name":"DeepSeek R1 Distill Llama 70B","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-01-20","last_updated":"2025-01-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":0.03,"output":0.13}},"qwen3-coder-30b-a3b-instruct":{"id":"qwen3-coder-30b-a3b-instruct","name":"Qwen3 Coder 30B A3B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.09999999999999999,"output":0.3}},"gemma2-9b-it":{"id":"gemma2-9b-it","name":"Google Gemma 2","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-06","release_date":"2024-06-25","last_updated":"2024-06-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0.01,"output":0.03}},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-22","last_updated":"2025-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":65536},"cost":{"input":0.27,"output":0.41}},"grok-4":{"id":"grok-4","name":"xAI Grok 4","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-09","last_updated":"2024-07-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":3,"output":15,"cache_read":0.75}},"llama-prompt-guard-2-86m":{"id":"llama-prompt-guard-2-86m","name":"Meta Llama Prompt Guard 2 86M","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-10-01","last_updated":"2024-10-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":512,"output":2},"cost":{"input":0.01,"output":0.01}},"mistral-large-2411":{"id":"mistral-large-2411","name":"Mistral-Large","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-24","last_updated":"2024-07-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":2,"output":6}},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"OpenAI GPT-OSS 120b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2024-06-01","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":0.04,"output":0.16}},"gpt-5.1":{"id":"gpt-5.1","name":"OpenAI GPT-5.1","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.12500000000000003}},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Anthropic: Claude 4.5 Haiku (20251001)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-10","release_date":"2025-10-01","last_updated":"2025-10-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":1,"output":5,"cache_read":0.09999999999999999,"cache_write":1.25}},"hermes-2-pro-llama-3-8b":{"id":"hermes-2-pro-llama-3-8b","name":"Hermes 2 Pro Llama 3 8B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-05","release_date":"2024-05-27","last_updated":"2024-05-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":0.14,"output":0.14}},"grok-4-1-fast-non-reasoning":{"id":"grok-4-1-fast-non-reasoning","name":"xAI Grok 4.1 Fast Non-Reasoning","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-17","last_updated":"2025-11-17","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.19999999999999998,"output":0.5,"cache_read":0.049999999999999996}},"grok-3":{"id":"grok-3","name":"xAI Grok 3","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2024-06-01","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.75}},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Google Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.3,"output":2.5,"cache_read":0.075,"cache_write":0.3}},"qwen3-coder":{"id":"qwen3-coder","name":"Qwen3 Coder 480B A35B Instruct Turbo","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":16384},"cost":{"input":0.22,"output":0.95}},"glm-4.6":{"id":"glm-4.6","name":"Zai GLM-4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":204800,"output":131072},"cost":{"input":0.44999999999999996,"output":1.5}},"sonar-reasoning-pro":{"id":"sonar-reasoning-pro","name":"Perplexity Sonar Reasoning Pro","description":"Web-grounded reasoning model for multi-step research and cited answers","family":"sonar-reasoning","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-01-27","last_updated":"2025-01-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":127000,"output":4096},"cost":{"input":2,"output":8}},"llama-prompt-guard-2-22m":{"id":"llama-prompt-guard-2-22m","name":"Meta Llama Prompt Guard 2 22M","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-10-01","last_updated":"2024-10-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":512,"output":2},"cost":{"input":0.01,"output":0.01}},"gpt-5-nano":{"id":"gpt-5-nano","name":"OpenAI GPT-5 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.049999999999999996,"output":0.39999999999999997,"cache_read":0.005}},"claude-3-haiku-20240307":{"id":"claude-3-haiku-20240307","name":"Anthropic: Claude 3 Haiku","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-03","release_date":"2024-03-07","last_updated":"2024-03-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":4096},"cost":{"input":0.25,"output":1.25,"cache_read":0.03,"cache_write":0.3}},"grok-4-1-fast-reasoning":{"id":"grok-4-1-fast-reasoning","name":"xAI Grok 4.1 Fast Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-17","last_updated":"2025-11-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":0.19999999999999998,"output":0.5,"cache_read":0.049999999999999996}},"claude-3.7-sonnet":{"id":"claude-3.7-sonnet","name":"Anthropic: Claude 3.7 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-02","release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.30000000000000004,"cache_write":3.75}},"deepseek-reasoner":{"id":"deepseek-reasoner","name":"DeepSeek Reasoner","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-01-20","last_updated":"2025-01-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"cost":{"input":0.56,"output":1.68,"cache_read":0.07}},"gpt-5-chat-latest":{"id":"gpt-5-chat-latest","name":"OpenAI GPT-5 Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt-codex","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2024-09","release_date":"2024-09-30","last_updated":"2024-09-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.25,"output":10,"cache_read":0.12500000000000003}},"claude-opus-4":{"id":"claude-opus-4","name":"Anthropic: Claude Opus 4","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-14","last_updated":"2025-05-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"kimi-k2-thinking":{"id":"kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":262144},"cost":{"input":0.48,"output":2}},"gemini-3-pro-preview":{"id":"gemini-3-pro-preview","name":"Google Gemini 3 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.19999999999999998}}}},"evroc":{"id":"evroc","env":["EVROC_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://models.think.evroc.com/v1","name":"evroc","doc":"https://docs.evroc.com/products/think/overview.html","models":{"intfloat/multilingual-e5-large-instruct":{"id":"intfloat/multilingual-e5-large-instruct","name":"E5 Multi-Lingual Large Embeddings 0.6B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2024-06-01","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":512},"cost":{"input":0.114,"output":0.114}},"nvidia/Llama-3.3-70B-Instruct-FP8":{"id":"nvidia/Llama-3.3-70B-Instruct-FP8","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":1.15,"output":1.15},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.4375,"output":5.75},"canonical_model_id":"moonshotai/kimi-k2.6"},"mistralai/Mistral-Medium-3.5-128B":{"id":"mistralai/Mistral-Medium-3.5-128B","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.725,"output":6.9},"canonical_model_id":"mistral/mistral-medium-2604"},"mistralai/Voxtral-Small-24B-2507":{"id":"mistralai/Voxtral-Small-24B-2507","name":"Voxtral Small 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"voxtral","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2025-03-01","last_updated":"2025-03-01","modalities":{"input":["audio","text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000},"cost":{"input":0.0023,"output":0.0023,"output_audio":2.3}},"Qwen/Qwen3-Embedding-8B":{"id":"Qwen/Qwen3-Embedding-8B","name":"Qwen3 Embedding 8B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":4096},"cost":{"input":0.115,"output":0.115}},"Qwen/Qwen3.6-35B-A3B":{"id":"Qwen/Qwen3.6-35B-A3B","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.345,"output":1.38},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8-27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.87,"output":3.5},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3-Reranker-4B":{"id":"Qwen/Qwen3-Reranker-4B","name":"Qwen3 Reranker 4B","description":"Reranking model for improving retrieval quality in search and recommendation systems","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.0575,"output":0}},"KBLab/kb-whisper-large":{"id":"KBLab/kb-whisper-large","name":"KB Whisper","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2024-10-01","last_updated":"2024-10-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":448},"cost":{"input":0.0023,"output":0.0023,"output_audio":2.3}},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":131072},"cost":{"input":1.4375,"output":5.75},"canonical_model_id":"zhipuai/glm-5.2"},"google/gemma-4-26B-A4B-it":{"id":"google/gemma-4-26B-A4B-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.144,"output":0.575},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"openai/whisper-large-v3":{"id":"openai/whisper-large-v3","type":"transcription","name":"Whisper 3 Large","description":"Open Whisper checkpoint for robust multilingual transcription and captioning","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-10-01","last_updated":"2024-10-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":4096},"cost":{"input":0.0023,"output":0.0023,"output_audio":2.3},"canonical_model_id":"openai/whisper-large-v3"},"openai/whisper-large-v3-turbo":{"id":"openai/whisper-large-v3-turbo","type":"transcription","name":"Whisper Large v3 Turbo","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-10-01","last_updated":"2024-10-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":448},"cost":{"input":0.0023,"output":0.0023,"output_audio":2.3},"canonical_model_id":"openai/whisper-large-v3-turbo"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.23,"output":0.92}},"evroc/roc":{"id":"evroc/roc","name":"roc","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-06-06","last_updated":"2026-06-06","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":2.875,"output":11.516}}}},"opencode":{"id":"opencode","env":["OPENCODE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://opencode.ai/inference/openai/v1","name":"OpenCode","doc":"https://opencode.ai/v2/docs/console/models/","models":{"longcat-2.0-free":{"id":"longcat-2.0-free","name":"LongCat-2.0 Free","description":"Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"meituan/longcat-2.0"},"deepseek-v4.1-flash-free":{"id":"deepseek-v4.1-flash-free","name":"DeepSeek V4.1 Flash Free","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"minimax-m2.1-free":{"id":"minimax-m2.1-free","name":"MiniMax-M2.1 Free","description":"Legacy model retained for compatibility with older integrations","family":"minimax-free","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"minimax-m3-free":{"id":"minimax-m3-free","name":"MiniMax-M3 Free","description":"Legacy model retained for compatibility with older integrations","family":"minimax-m3-free","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-31","last_updated":"2026-05-31","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"gemini-3-flash":{"id":"gemini-3-flash","name":"Gemini 3 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":0.5,"output":3,"cache_read":0.05}},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"claude-opus-4-1":{"id":"claude-opus-4-1","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.06}},"nemotron-3-super-free":{"id":"nemotron-3-super-free","name":"Nemotron 3 Super Free","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron-free","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2026-02","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25}},"nemotron-3-ultra-free":{"id":"nemotron-3-ultra-free","name":"Nemotron 3 Ultra Free","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron-free","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2026-02","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"glm-4.7-free":{"id":"glm-4.7-free","name":"GLM-4.7 Free","description":"Legacy model retained for compatibility with older integrations","family":"glm-free","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"gpt-5.4-pro":{"id":"gpt-5.4-pro","name":"GPT-5.4 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":30,"output":180,"cache_read":30}},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.74,"output":3.84,"cache_read":0.145},"canonical_model_id":"deepseek/deepseek-v4-pro"},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"gemini-3-pro":{"id":"gemini-3-pro","name":"Gemini 3 Pro","description":"Legacy model retained for compatibility with older integrations","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}}},"kimi-k2.5-free":{"id":"kimi-k2.5-free","name":"Kimi K2.5 Free","description":"Legacy model retained for compatibility with older integrations","family":"kimi-free","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-10","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"cat-jam-free":{"id":"cat-jam-free","name":"Cat Jam Free","description":"A text-only reasoning model for coding and agentic workflows, with tool calling and structured JSON outputs.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"ling-2.6-flash-free":{"id":"ling-2.6-flash-free","name":"Ling 2.6 Flash Free","description":"Legacy model retained for compatibility with older integrations","family":"ling-flash-free","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262100,"output":32800},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0}},"ll-alpha-free":{"id":"ll-alpha-free","name":"LL Alpha Free","description":"Best in class local model","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-06","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"step-5-preview-free":{"id":"step-5-preview-free","name":"Step 5 Preview Free","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"stepfun/step-5-preview"},"claude-sonnet-4":{"id":"claude-sonnet-4","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}}},"ling-3.0-flash-free":{"id":"ling-3.0-flash-free","name":"Ling-3.0-flash Free","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"qwen3.5-plus":{"id":"qwen3.5-plus","name":"Qwen3.5 Plus","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25}},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-10","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.95,"output":4,"cache_read":0.16}},"muse-spark-1.3-contributor-free":{"id":"muse-spark-1.3-contributor-free","name":"Muse Spark 1.3 Free","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for coding and agentic workflows.","family":"muse-free","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.07,"output":8.5,"cache_read":0.107}},"north-mini-code-free":{"id":"north-mini-code-free","name":"North Mini Code Free","description":"Cohere coding model for practical software engineering and agentic edits","family":"north-free","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-09-23","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0},"canonical_model_id":"cohere/north-mini-code-1-0"},"muse-spark-1.2-contributor-free":{"id":"muse-spark-1.2-contributor-free","name":"Muse Spark 1.2 Free","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse-free","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"meta/muse-spark-1.2"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625}},"mimo-v2-omni-free":{"id":"mimo-v2-omni-free","name":"MiMo V2 Omni Free","description":"Legacy model retained for compatibility with older integrations","family":"mimo-omni-free","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":64000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":3.125,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"cache_write":6.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5,"cache_write":6.25}},"canonical_model_id":"openai/gpt-5.6-terra"},"exo-free":{"id":"exo-free","name":"Exo Free","description":"Free-tier preview reasoning model with text and image input and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"temperature":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0}},"longcat-2.5-preview-free":{"id":"longcat-2.5-preview-free","name":"LongCat 2.5 Preview Free","description":"Meituan's multimodal reasoning model for coding and agent tasks, with image understanding and a 1M-token context window","family":"longcat","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"meituan/longcat-2.5-preview"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"gpt-5-codex":{"id":"gpt-5-codex","name":"GPT-5 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.07,"output":8.5,"cache_read":0.107}},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"beam-free":{"id":"beam-free","name":"Beam Free","description":"Sparse mixture-of-experts reasoning model with 23B active parameters for coding and agentic workflows","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"kimi-k2":{"id":"kimi-k2","name":"Kimi K2","description":"Legacy model retained for compatibility with older integrations","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.4,"output":2.5,"cache_read":0.4}},"mimo-v2.5-free":{"id":"mimo-v2.5-free","name":"MiMo V2.5 Free","description":"MiMo omni model for text, image, video, audio, and agents","family":"mimo-v2.5-free","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"mimo-v2.6-flash-free":{"id":"mimo-v2.6-flash-free","name":"MiMo-V2.6-Flash Free","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"gemini-3.1-pro":{"id":"gemini-3.1-pro","name":"Gemini 3.1 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}}},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":1.5,"output":7.5,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.7-flash"},"glm-5-free":{"id":"glm-5-free","name":"GLM-5 Free","description":"Legacy model retained for compatibility with older integrations","family":"glm-free","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"hy3-free":{"id":"hy3-free","name":"Hy3 Free","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hy3-free","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":190000,"input":192000,"output":64000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"tencent/hy3"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":1.5,"output":7.5,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.8-flash"},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Legacy model retained for compatibility with older integrations","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.6,"output":2.2,"cache_read":0.1}},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"ling-3.1-flash-free":{"id":"ling-3.1-flash-free","name":"Ling 3.1 Flash Free","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"inclusionai/ling-3.1-flash"},"fledge-alpha-free":{"id":"fledge-alpha-free","name":"Fledge Alpha Free","description":"Free-tier preview reasoning model with text and image input and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0}},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.4,"output":4.4,"cache_read":0.26}},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5}},"mimo-v2-pro-free":{"id":"mimo-v2-pro-free","name":"MiMo V2 Pro Free","description":"Legacy model retained for compatibility with older integrations","family":"mimo-pro-free","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"minimax-m2.1":{"id":"minimax-m2.1","name":"MiniMax-M2.1","description":"Legacy model retained for compatibility with older integrations","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.1}},"gpt-5.3-codex-spark":{"id":"gpt-5.3-codex-spark","name":"GPT-5.3 Codex Spark","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex-spark","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"claude-3-5-haiku":{"id":"claude-3-5-haiku","name":"Claude Haiku 3.5","description":"Legacy model retained for compatibility with older integrations","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07-31","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0.8,"output":4,"cache_read":0.08,"cache_write":1}},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-01-14","last_updated":"2026-01-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"gpt-5.1-codex-mini":{"id":"gpt-5.1-codex-mini","name":"GPT-5.1 Codex Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"laguna-s-2.1-free":{"id":"laguna-s-2.1-free","name":"Laguna S 2.1 Free","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"poolside/laguna-s-2.1"},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.07,"output":8.5,"cache_read":0.107}},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":2,"output":10,"cache_read":0.1,"input_audio":2},"canonical_model_id":"google/gemini-4-argon"},"deepseek-v4-flash-vision-exp":{"id":"deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}}},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"grok-build-0.1":{"id":"grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1,"output":2,"cache_read":0.2},"canonical_model_id":"xai/grok-build-0.1"},"trinity-large-preview-free":{"id":"trinity-large-preview-free","name":"Trinity Large Preview","description":"Legacy model retained for compatibility with older integrations","family":"trinity","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-01-27","last_updated":"2026-01-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0},"canonical_model_id":"arcee-ai/trinity-large-preview"},"muse-spark-1.3":{"id":"muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"nemotron-3.5-super-vl-free":{"id":"nemotron-3.5-super-vl-free","name":"Nemotron 3.5 Super VL Free","description":"NVIDIA vision-language reasoning model for text, image, and video understanding and agentic tool use","family":"nemotron-free","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"hy3-preview-free":{"id":"hy3-preview-free","name":"Hy3 preview Free","description":"Legacy model retained for compatibility with older integrations","family":"hy3-free","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"glm-5.3-flash-free":{"id":"glm-5.3-flash-free","name":"GLM-5.3-Flash Free","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5":{"id":"glm-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1,"output":3.2,"cache_read":0.2}},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"ling-3.0-tiny-free":{"id":"ling-3.0-tiny-free","name":"Ling-3.0-tiny Free","description":"Compact MoE model for responsive agents, instruction following, and multi-turn conversations","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0}},"big-pickle":{"id":"big-pickle","name":"Big Pickle","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"big-pickle","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-10-17","last_updated":"2025-10-17","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":160000,"output":32000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"ring-2.6-1t-free":{"id":"ring-2.6-1t-free","name":"Ring 2.6 1T Free","description":"Legacy model retained for compatibility with older integrations","family":"ring-1t-free","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-06","release_date":"2026-05-08","last_updated":"2026-05-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":66000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0}},"nemotron-3.5-lightning-free":{"id":"nemotron-3.5-lightning-free","name":"Nemotron 3.5 Lightning Free","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron-free","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"x-preview-f-free":{"id":"x-preview-f-free","name":"Ox Alpha Free (Unlimited)","description":"Stealth reasoning model for coding, agentic tasks, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-08-21","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"ling-3.0-flash-fin-free":{"id":"ling-3.0-flash-fin-free","name":"Ling 3.0 Flash Fin Free","description":"Finance-enhanced model for financial research, multi-step investment workflows, and long-horizon planning and execution","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-09-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"inclusionai/ling-3.0-flash-fin"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.2,"output":1.25,"cache_read":0.02}},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":128000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"qwen3.6-plus-free":{"id":"qwen3.6-plus-free","name":"Qwen3.6 Plus Free","description":"Legacy model retained for compatibility with older integrations","family":"qwen-free","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"grok-code":{"id":"grok-code","name":"Grok Code Fast 1","description":"Legacy model retained for compatibility with older integrations","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-08-20","last_updated":"2025-08-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"mistral-large-4":{"id":"mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":262144},"provider":{"npm":"@ai-sdk/mistral","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.07,"output":8.5,"cache_read":0.107}},"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"deepseek-v4-flash-free":{"id":"deepseek-v4-flash-free","name":"DeepSeek V4 Flash Free","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"gpt-5.1-codex-max":{"id":"gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"qwen3-coder":{"id":"qwen3-coder","name":"Qwen3 Coder","description":"Legacy model retained for compatibility with older integrations","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.45,"output":1.8}},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075}},"glm-4.6":{"id":"glm-4.6","name":"GLM-4.6","description":"Legacy model retained for compatibility with older integrations","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.6,"output":2.2,"cache_read":0.1}},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"space-bunny-free":{"id":"space-bunny-free","name":"Space Bunny Free","description":"Anonymous preview reasoning model for coding, agentic tasks, tool use, and multimodal input","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":524288,"output":524288},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"minimax-m2.5-free":{"id":"minimax-m2.5-free","name":"MiniMax-M2.5 Free","description":"Legacy model retained for compatibility with older integrations","family":"minimax-free","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":1.5,"output":7.5,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.6-flash"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-10","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.6,"output":3,"cache_read":0.08}},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-5.5-pro":{"id":"gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":30,"output":180,"cache_read":30}},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.05,"output":0.4,"cache_read":0.005}},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"minimax-m2.5":{"id":"minimax-m2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.06}},"mimo-v2-flash-free":{"id":"mimo-v2-flash-free","name":"MiMo V2 Flash Free","description":"Legacy model retained for compatibility with older integrations","family":"mimo-flash-free","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"minimax-m3.1-flash-free":{"id":"minimax-m3.1-flash-free","name":"MiniMax M3.1 Flash Free","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":512000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"kimi-k2-thinking":{"id":"kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Legacy model retained for compatibility with older integrations","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-10","release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.4,"output":2.5,"cache_read":0.4}},"muse-spark-1.2":{"id":"muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.175}}}},"stepfun-step-plan":{"id":"stepfun-step-plan","env":["STEPFUN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.stepfun.com/step_plan/v1","name":"StepFun Step Plan (China)","doc":"https://platform.stepfun.com/docs/zh/step-plan/integrations/reasoning-api","models":{"step-router-v1":{"id":"step-router-v1","name":"Step Router v1","description":"StepFun routing model that dispatches requests to the appropriate Step model.","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":256000}},"step-3.7-flash":{"id":"step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"canonical_model_id":"stepfun/step-3.7-flash"},"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"canonical_model_id":"stepfun/step-5-preview"},"step-3.5-flash-2603":{"id":"step-3.5-flash-2603","name":"Step 3.5 Flash 2603","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"canonical_model_id":"stepfun/step-3.5-flash-2603"},"step-3.5-flash":{"id":"step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"canonical_model_id":"stepfun/step-3.5-flash"}}},"cerebras":{"id":"cerebras","env":["CEREBRAS_API_KEY"],"npm":"@ai-sdk/cerebras","name":"Cerebras","doc":"https://inference-docs.cerebras.ai/models/overview","models":{"qwen-3.8-27b":{"id":"qwen-3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-09-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":40960},"cost":{"input":0.99,"output":1.49,"cache_read":0.99},"canonical_model_id":"alibaba/qwen3.8-27b"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2026-06-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":40960},"cost":{"input":0.35,"output":0.75},"canonical_model_id":"openai/gpt-oss-120b"}}},"alibaba":{"id":"alibaba","env":["DASHSCOPE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://dashscope-intl.aliyuncs.com/compatible-mode/v1","name":"Alibaba","doc":"https://www.alibabacloud.com/help/en/model-studio/models","models":{"qwen3.7-flash":{"id":"qwen3.7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":131072},"cost":{"input":0.03,"output":0.13,"cache_read":0.003,"cache_write":0.0375,"tiers":[{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":0.125,"tier":{"type":"context","size":32000}},{"input":0.2,"output":0.8,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen3-vl-235b-a22b":{"id":"qwen3-vl-235b-a22b","name":"Qwen3-VL 235B-A22B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.7,"output":2.8,"reasoning":8.4}},"qwen2-5-omni-7b":{"id":"qwen2-5-omni-7b","name":"Qwen2.5-Omni 7B","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-12","last_updated":"2024-12","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":true,"limit":{"context":32768,"output":2048},"status":"deprecated","cost":{"input":0.1,"output":0.4,"input_audio":6.76}},"qwen-plus":{"id":"qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.4,"output":1.2,"reasoning":4,"tiers":[{"input":1.2,"output":3.6,"reasoning":12,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.2,"output":3.6,"reasoning":12}}},"qwen-flash":{"id":"qwen-flash","name":"Qwen Flash","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.05,"output":0.4,"tiers":[{"input":0.25,"output":2,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.25,"output":2}}},"qwen3-coder-plus":{"id":"qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Hosted Qwen coder for software agents, repo edits, and long-context code","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":1,"output":5}},"qwen3.5-122b-a10b":{"id":"qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":3.2}},"qwen-plus-character-ja":{"id":"qwen-plus-character-ja","name":"Qwen Plus Character (Japanese)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-01","last_updated":"2024-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":512},"cost":{"input":0.5,"output":1.4}},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.44,"output":1.32,"cache_read":0.088},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"qwen3-8b":{"id":"qwen3-8b","name":"Qwen3 8B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.18,"output":0.7,"reasoning":2.1}},"qwen3.5-plus":{"id":"qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.4,"output":2.4,"reasoning":2.4,"tiers":[{"input":0.5,"output":3,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.5,"output":3}}},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.375,"output":2.25}},"qwen3.8-omni-flash":{"id":"qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"qwen-vl-plus":{"id":"qwen-vl-plus","name":"Qwen-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-08-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.21,"output":0.63}},"qwen-max":{"id":"qwen-max","name":"Qwen Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":1.6,"output":6.4}},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625,"tiers":[{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5}}},"qwq-plus":{"id":"qwq-plus","name":"QwQ Plus","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-03-05","last_updated":"2025-03-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.8,"output":2.4}},"qwen3-asr-flash":{"id":"qwen3-asr-flash","type":"transcription","name":"Qwen3-ASR Flash","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-04","release_date":"2025-09-08","last_updated":"2025-09-08","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":53248,"output":4096},"cost":{"input":0.035,"output":0.035}},"qwen3-omni-flash":{"id":"qwen3-omni-flash","name":"Qwen3-Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":false,"limit":{"context":65536,"output":16384},"cost":{"input":0.43,"output":1.66,"input_audio":3.81,"output_audio":15.11}},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.3125,"tiers":[{"input":1,"output":4,"cache_read":0.1,"cache_write":1.25,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1,"output":4,"cache_read":0.1,"cache_write":1.25}}},"qwen3.5-flash":{"id":"qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"alibaba/qwen3.5-flash"},"qwen3-235b-a22b":{"id":"qwen3-235b-a22b","name":"Qwen3 235B-A22B","description":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.7,"output":2.8,"reasoning":8.4}},"qwen3-coder-flash":{"id":"qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":1.5,"tiers":[{"input":0.5,"output":2.5,"tier":{"type":"context","size":32000}},{"input":0.8,"output":4,"tier":{"type":"context","size":128000}},{"input":1.6,"output":9.6,"tier":{"type":"context","size":256000}}]}},"qwen2-5-vl-72b-instruct":{"id":"qwen2-5-vl-72b-instruct","name":"Qwen2.5-VL 72B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":2.8,"output":8.4}},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen-omni-turbo-realtime":{"id":"qwen-omni-turbo-realtime","type":"realtime","name":"Qwen-Omni Turbo Realtime","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-05-08","last_updated":"2025-05-08","modalities":{"input":["text","image","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":32768,"output":2048},"status":"deprecated","cost":{"input":0.27,"output":1.07,"input_audio":4.44,"output_audio":8.89}},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-04","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.4,"output":1.6,"cache_read":0.04,"cache_write":0.5,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.12,"cache_write":1.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.12,"cache_write":1.5}}},"qwen-vl-max":{"id":"qwen-vl-max","name":"Qwen-VL Max","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-08","last_updated":"2025-08-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.8,"output":3.2}},"qwen-turbo":{"id":"qwen-turbo","name":"Qwen Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-11-01","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":16384},"status":"deprecated","cost":{"input":0.05,"output":0.2,"reasoning":0.5}},"qwen3-next-80b-a3b-instruct":{"id":"qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.5,"output":2}},"qvq-max":{"id":"qvq-max","name":"QVQ Max","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qvq","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-03-25","last_updated":"2025-03-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":1.2,"output":4.8}},"qwen3.5-27b":{"id":"qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":2.4}},"qwen3-next-80b-a3b-thinking":{"id":"qwen3-next-80b-a3b-thinking","name":"Qwen3-Next 80B-A3B (Thinking)","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.5,"output":6}},"qwen3-vl-30b-a3b":{"id":"qwen3-vl-30b-a3b","name":"Qwen3-VL 30B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.2,"output":0.8,"reasoning":2.4}},"qwen-omni-turbo":{"id":"qwen-omni-turbo","name":"Qwen-Omni Turbo","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-01-19","last_updated":"2025-03-26","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":false,"limit":{"context":32768,"output":2048},"status":"deprecated","cost":{"input":0.07,"output":0.27,"input_audio":4.44,"output_audio":8.89}},"qwen3.5-35b-a3b":{"id":"qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2}},"qwen3-coder-480b-a35b-instruct":{"id":"qwen3-coder-480b-a35b-instruct","name":"Qwen3-Coder 480B-A35B Instruct","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":1.5,"output":7.5,"tiers":[{"input":2.7,"output":13.5,"tier":{"type":"context","size":32000}},{"input":4.5,"output":22.5,"tier":{"type":"context","size":128000}}]}},"qwen2-5-72b-instruct":{"id":"qwen2-5-72b-instruct","name":"Qwen2.5 72B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":1.4,"output":5.6}},"qwen3-32b":{"id":"qwen3-32b","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.7,"output":2.8,"reasoning":8.4}},"qwen3-omni-flash-realtime":{"id":"qwen3-omni-flash-realtime","type":"realtime","name":"Qwen3-Omni Flash Realtime","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":false,"limit":{"context":65536,"output":16384},"status":"deprecated","cost":{"input":0.52,"output":1.99,"input_audio":4.57,"output_audio":18.13}},"qwen2-5-14b-instruct":{"id":"qwen2-5-14b-instruct","name":"Qwen2.5 14B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.35,"output":1.4}},"qwen3-coder-30b-a3b-instruct":{"id":"qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.45,"output":2.25,"tiers":[{"input":0.75,"output":3.75,"tier":{"type":"context","size":32000}},{"input":1.2,"output":6,"tier":{"type":"context","size":128000}}]}},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6}},"qwen2-5-vl-7b-instruct":{"id":"qwen2-5-vl-7b-instruct","name":"Qwen2.5-VL 7B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.35,"output":1.05}},"qwen-vl-ocr":{"id":"qwen-vl-ocr","name":"Qwen-VL OCR","description":"OCR model for extracting structured text from documents and screenshots","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-10-28","last_updated":"2025-04-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":34096,"output":4096},"cost":{"input":0.07,"output":0.16}},"qwen3.6-max-preview":{"id":"qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":1.3,"output":7.8,"cache_read":0.13,"cache_write":1.625}},"qwen-mt-turbo":{"id":"qwen-mt-turbo","name":"Qwen-MT Turbo","description":"Translation model for multilingual conversion, localization, and cross-language workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-01","last_updated":"2025-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":8192},"status":"deprecated","cost":{"input":0.16,"output":0.49}},"qwen2-5-32b-instruct":{"id":"qwen2-5-32b-instruct","name":"Qwen2.5 32B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.7,"output":2.8}},"qwen3.6-27b":{"id":"qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6}},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.5,"cache_write":3.125}},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.28,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"qwen3-livetranslate-flash-realtime":{"id":"qwen3-livetranslate-flash-realtime","type":"realtime","name":"Qwen3-LiveTranslate Flash Realtime","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-09-22","last_updated":"2025-09-22","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":false,"limit":{"context":53248,"output":4096},"status":"deprecated","cost":{"input":10,"output":10,"input_audio":10,"output_audio":38}},"qwen-mt-plus":{"id":"qwen-mt-plus","name":"Qwen-MT Plus","description":"Translation model for multilingual conversion, localization, and cross-language workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-01","last_updated":"2025-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":8192},"cost":{"input":2.46,"output":7.37}},"qwen3-vl-plus":{"id":"qwen3-vl-plus","name":"Qwen3-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":1.6,"tiers":[{"input":0.3,"output":2.4,"tier":{"type":"context","size":32000}},{"input":0.6,"output":4.8,"tier":{"type":"context","size":128000}}]}},"qwen2-5-7b-instruct":{"id":"qwen2-5-7b-instruct","name":"Qwen2.5 7B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.175,"output":0.7}},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen3-max":{"id":"qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":1.2,"output":6}},"qwen3-14b":{"id":"qwen3-14b","name":"Qwen3 14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.35,"output":1.4,"reasoning":4.2}}}},"blueclaw":{"id":"blueclaw","env":["BLUECLAW_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://openai.blueclaw.network/v1","name":"Blue Claw","doc":"https://blueclaw.network","models":{"Qwen3.8-27B":{"id":"Qwen3.8-27B","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"beta","canonical_model_id":"alibaba/qwen3.8-27b"}}},"crossmodel":{"id":"crossmodel","env":["CROSSMODEL_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.crossmodel.ai/v1","name":"CrossModel","doc":"https://www.crossmodel.ai/docs","models":{"anthropic/claude-haiku-4-5":{"id":"anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-fable-5-1":{"id":"anthropic/claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-opus-4-8":{"id":"anthropic/claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-sonnet-5-5":{"id":"anthropic/claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-haiku-5-5":{"id":"anthropic/claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-5-5":{"id":"anthropic/claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"moonshot/kimi-k2.6":{"id":"moonshot/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":1,"output":4.16,"cache_read":0.18,"cache_write":1},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshot/kimi-k3":{"id":"moonshot/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshot/kimi-k2.7-code":{"id":"moonshot/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":1,"output":4.16,"cache_read":0.18,"cache_write":1},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.35,"output":4.05,"cache_read":0.045,"cache_write":1.35},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006,"cache_write":0.3},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006,"cache_write":0.3},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006,"cache_write":0.3},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.15,"output":0.5,"cache_read":0.03,"cache_write":0.15},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.2,"output":4.4,"cache_read":0.3,"cache_write":1.2},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.47,"output":2.16,"cache_read":0.1,"cache_write":0.47,"tiers":[{"input":0.62,"output":2.47,"cache_read":0.13,"cache_write":0.62,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":1,"output":3.8,"cache_read":0.2,"cache_write":1,"tiers":[{"input":1.2,"output":4.4,"cache_read":0.3,"cache_write":1.2,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.6,"output":3,"cache_read":0.16,"cache_write":0.6,"tiers":[{"input":0.8,"output":3.4,"cache_read":0.2,"cache_write":0.8,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":0.9,"output":3.7,"cache_read":0.18,"cache_write":0.9,"tiers":[{"input":1.1,"output":4.3,"cache_read":0.27,"cache_write":1.1,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5-turbo"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.2,"output":4.4,"cache_read":0.3,"cache_write":1.2},"canonical_model_id":"zhipuai/glm-5.2"},"qwen/qwen3.7-flash":{"id":"qwen/qwen3.7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":65536},"cost":{"input":0.04,"output":0.13,"cache_read":0.01,"cache_write":0.04,"tiers":[{"input":0.1,"output":0.37,"cache_read":0.02,"cache_write":0.12,"tier":{"type":"context","size":32000}},{"input":0.19,"output":0.74,"cache_read":0.04,"cache_write":0.24,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen/qwen3.8-omni-flash":{"id":"qwen/qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.13,"output":0.43,"cache_read":0.016,"cache_write":0.13},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.32,"output":1.88,"cache_read":0.032,"cache_write":0.4,"tiers":[{"input":1.25,"output":7.5,"cache_read":0.124,"cache_write":1.57,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.25,"output":7.5,"cache_read":0.124,"cache_write":1.57}},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.19,"output":1.13,"cache_read":0.019,"cache_write":0.24,"tiers":[{"input":0.75,"output":4.5,"cache_read":0.075,"cache_write":0.94,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.75,"output":4.5,"cache_read":0.075,"cache_write":0.94}},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.13,"output":0.43,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.32,"output":1.25,"cache_read":0.032,"cache_write":0.4,"tiers":[{"input":0.96,"output":3.75,"cache_read":0.096,"cache_write":1.2,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.96,"output":3.75,"cache_read":0.096,"cache_write":1.2}},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.88,"output":5.63,"cache_read":0.375,"cache_write":2.35},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.8-max":{"id":"qwen/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.88,"output":5.63,"cache_read":0.23,"cache_write":2.35},"canonical_model_id":"alibaba/qwen3.8-max"},"gemini/gemini-3-flash-preview":{"id":"gemini/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.5},"canonical_model_id":"google/gemini-3-flash-preview"},"gemini/gemini-2.5-flash-lite":{"id":"gemini/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":0.1},"canonical_model_id":"google/gemini-2.5-flash-lite"},"gemini/gemini-2.5-pro":{"id":"gemini/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"cache_write":1.25,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"cache_write":2.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":2.5}},"canonical_model_id":"google/gemini-2.5-pro"},"gemini/gemini-3.7-flash":{"id":"gemini/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.75},"canonical_model_id":"google/gemini-3.7-flash"},"gemini/gemini-3.1-pro-preview":{"id":"gemini/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"gemini/gemini-3.8-flash":{"id":"gemini/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.75},"canonical_model_id":"google/gemini-3.8-flash"},"gemini/gemini-3.5-flash":{"id":"gemini/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"gemini/gemini-4-argon":{"id":"gemini/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2},"canonical_model_id":"google/gemini-4-argon"},"gemini/gemini-3.5-flash-lite":{"id":"gemini/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.3},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gemini/gemini-2.5-flash":{"id":"gemini/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.3},"canonical_model_id":"google/gemini-2.5-flash"},"gemini/gemini-3.6-flash":{"id":"gemini/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.33,"output":1.32,"cache_read":0.066,"cache_write":0.42},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":512000},"cost":{"input":0.33,"output":1.32,"cache_read":0.066,"cache_write":0.33,"tiers":[{"input":0.66,"output":2.63,"cache_read":0.132,"cache_write":0.66,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":0.66,"output":2.63,"cache_read":0.132,"cache_write":0.66}},"canonical_model_id":"minimax/MiniMax-M3"},"xiaomi/mimo-v2.5-pro":{"id":"xiaomi/mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.47,"output":0.94,"cache_read":0.005,"cache_write":0.47},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.16,"output":0.32,"cache_read":0.004,"cache_write":0.16},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2.6-flash":{"id":"xiaomi/mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.16,"output":0.32,"cache_read":0.004,"cache_write":0.16},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.47,"output":0.94,"cache_read":0.005,"cache_write":0.47},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"cache_write":1.25,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"cache_write":2.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4,"cache_write":2.5}},"canonical_model_id":"xai/grok-4.3"},"x-ai/grok-4.7":{"id":"x-ai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"cache_write":2,"tiers":[{"input":4,"output":12,"cache_read":1,"cache_write":4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1,"cache_write":4}},"canonical_model_id":"xai/grok-4.7"},"x-ai/grok-build-0.1":{"id":"x-ai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"cache_write":1,"tiers":[{"input":2,"output":4,"cache_read":0.4,"cache_write":2,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4,"cache_write":2}},"canonical_model_id":"xai/grok-build-0.1"},"x-ai/grok-4.5":{"id":"x-ai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3,"cache_write":2,"tiers":[{"input":4,"output":12,"cache_read":0.6,"cache_write":4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6,"cache_write":4}},"canonical_model_id":"xai/grok-4.5"},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"cache_write":2,"tiers":[{"input":4,"output":12,"cache_read":1,"cache_write":4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1,"cache_write":4}},"canonical_model_id":"xai/grok-4.6"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0.96,"output":2.88,"cache_read":0.048,"cache_write":0.96},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":131072},"cost":{"input":0.16,"output":0.64,"cache_read":0.04,"cache_write":0.16},"canonical_model_id":"tencent/hy3"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075,"cache_write":0.15},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":5,"tiers":[{"input":10,"output":45,"cache_read":1,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1,"cache_write":10}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":2.5,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5,"cache_write":5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02,"cache_write":0.2},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075,"cache_write":0.75},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"}}},"mistral":{"id":"mistral","env":["MISTRAL_API_KEY"],"npm":"@ai-sdk/mistral","name":"Mistral","doc":"https://docs.mistral.ai/getting-started/models/","models":{"mistral-medium-2604":{"id":"mistral-medium-2604","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5,"cache_read":0.15}},"devstral-small-2507":{"id":"devstral-small-2507","name":"Devstral Small","description":"Legacy model retained for compatibility with older integrations","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-07-10","last_updated":"2025-07-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.1,"output":0.3}},"zai-glm-5-3":{"id":"zai-glm-5-3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.3"},"mistral-medium-2508":{"id":"mistral-medium-2508","name":"Mistral Medium 3.1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-08-12","last_updated":"2025-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.4,"output":2}},"mistral-nemo":{"id":"mistral-nemo","name":"Mistral Nemo","description":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.15,"output":0.15}},"ministral-3b-latest":{"id":"ministral-3b-latest","name":"Ministral 3B (latest)","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-10-01","last_updated":"2024-10-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":128000},"cost":{"input":0.1,"output":0.1,"cache_read":0.01}},"labs-devstral-small-2512":{"id":"labs-devstral-small-2512","name":"Devstral Small 2","description":"Legacy model retained for compatibility with older integrations","family":"devstral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"status":"deprecated","cost":{"input":0,"output":0}},"magistral-medium-latest":{"id":"magistral-medium-latest","name":"Magistral Medium (latest)","description":"Mistral reasoning model for transparent analysis, math, and complex decisions","family":"magistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-03-17","last_updated":"2025-03-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"status":"deprecated","cost":{"input":2,"output":5}},"zai-glm-5-2":{"id":"zai-glm-5-2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"devstral-small-2505":{"id":"devstral-small-2505","name":"Devstral Small 2505","description":"Legacy model retained for compatibility with older integrations","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.1,"output":0.3}},"open-mistral-nemo":{"id":"open-mistral-nemo","name":"Open Mistral Nemo","description":"Legacy model retained for compatibility with older integrations","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.15,"output":0.15}},"pixtral-12b":{"id":"pixtral-12b","name":"Pixtral 12B","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2024-09-01","last_updated":"2024-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.15,"output":0.15}},"codestral-latest":{"id":"codestral-latest","name":"Codestral (latest)","description":"Mistral code model for completions, refactors, and developer IDE workflows","family":"codestral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-05-29","last_updated":"2025-01-04","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":4096},"cost":{"input":0.3,"output":0.9,"cache_read":0.03}},"mistral-small-latest":{"id":"mistral-small-latest","name":"Mistral Small (latest)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015}},"mistral-embed":{"id":"mistral-embed","type":"embedding","name":"Mistral Embed","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"mistral-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-12-11","last_updated":"2023-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":3072},"cost":{"input":0.1,"output":0}},"ministral-8b-2512":{"id":"ministral-8b-2512","name":"Ministral 3 8B","description":"Compact open vision-language model for edge deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.15,"output":0.15,"cache_read":0.015},"canonical_model_id":"mistral/ministral-3-8b-instruct-2512"},"labs-leanstral-1-5-1":{"id":"labs-leanstral-1-5-1","name":"Leanstral 1.5","description":"An updated Lean 4 formal proof engineering model optimised for automated theorem proving and autoformalization. 119B total parameters, 6.5B active.","family":"leanstral","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"status":"deprecated","cost":{"input":0,"output":0}},"devstral-2512":{"id":"devstral-2512","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.4,"output":2}},"ministral-3b-2512":{"id":"ministral-3b-2512","name":"Ministral 3 3B","description":"Compact open vision-language model for edge deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":262144},"cost":{"input":0.1,"output":0.1,"cache_read":0.01},"canonical_model_id":"mistral/ministral-3-3b-instruct-2512"},"mistral-large-2512":{"id":"mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5,"cache_read":0.05}},"voxtral-small-latest":{"id":"voxtral-small-latest","name":"Voxtral Small (latest)","description":"Instruct model with native audio input for speech understanding and tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32000},"cost":{"input":0.1,"output":0.4}},"mistral-small-2603":{"id":"mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015}},"mistral-medium-2505":{"id":"mistral-medium-2505","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":0.4,"output":2}},"devstral-latest":{"id":"devstral-latest","name":"Devstral 2","description":"Legacy model retained for compatibility with older integrations","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.4,"output":2}},"devstral-medium-latest":{"id":"devstral-medium-latest","name":"Devstral 2 (latest)","description":"Legacy model retained for compatibility with older integrations","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.4,"output":2}},"ministral-8b-latest":{"id":"ministral-8b-latest","name":"Ministral 8B (latest)","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-10-01","last_updated":"2024-10-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"cost":{"input":0.15,"output":0.15,"cache_read":0.015}},"ministral-14b-2512":{"id":"ministral-14b-2512","name":"Ministral 3 14B","description":"Open vision-language model for efficient local deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.2,"output":0.2,"cache_read":0.02},"canonical_model_id":"mistral/ministral-3-14b-instruct-2512"},"pixtral-large-latest":{"id":"pixtral-large-latest","name":"Pixtral Large (latest)","description":"Mistral's larger vision model for document-heavy image understanding and chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2024-11-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":2,"output":6}},"glm-5-2":{"id":"glm-5-2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"mistral-large-2411":{"id":"mistral-large-2411","name":"Mistral Large 2.1","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-18","last_updated":"2024-11-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":2,"output":6},"canonical_model_id":"mistral/mistral-large-2411"},"mistral-large-4":{"id":"mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":262144},"status":"beta","cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"mistral-small-2506":{"id":"mistral-small-2506","name":"Mistral Small 3.2","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":0.1,"output":0.3}},"devstral-medium-2507":{"id":"devstral-medium-2507","name":"Devstral Medium","description":"Legacy model retained for compatibility with older integrations","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-07-10","last_updated":"2025-07-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.4,"output":2}},"codestral-2508":{"id":"codestral-2508","name":"Codestral 25.08","description":"Mistral code model for low-latency completion, fill-in-the-middle, and code generation","family":"codestral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":8192},"cost":{"input":0.3,"output":0.9,"cache_read":0.03},"canonical_model_id":"mistral/codestral-2508"},"mistral-large-latest":{"id":"mistral-large-latest","name":"Mistral Large (latest)","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5,"cache_read":0.05}},"mistral-medium-latest":{"id":"mistral-medium-latest","name":"Mistral Medium (latest)","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"mistral/mistral-medium-latest"},"voxtral-mini-tts-latest":{"id":"voxtral-mini-tts-latest","type":"speech","name":"Voxtral Mini TTS (latest)","description":"Multilingual text-to-speech model with zero-shot voice cloning","family":"voxtral","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-23","last_updated":"2026-03-23","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":0,"output":0}},"voxtral-mini-latest":{"id":"voxtral-mini-latest","type":"transcription","name":"Voxtral Mini (latest)","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"voxtral","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-04","last_updated":"2026-02-04","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0}},"voxtral-small-2507":{"id":"voxtral-small-2507","name":"Voxtral Small","description":"Instruct model with native audio input for speech understanding and tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32000},"cost":{"input":0.1,"output":0.4}}}},"boundless":{"id":"boundless","env":["BOUNDLESS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.inference.boundless.network/v1","name":"Boundless","doc":"https://inference.boundless.network/models","models":{"qwen3.6":{"id":"qwen3.6","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.04,"output":0.35,"cache_read":0.02},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.12,"output":0.4,"cache_read":0.024},"canonical_model_id":"zhipuai/glm-5.3-flash"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01},"canonical_model_id":"openai/gpt-6-luna"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":2.3,"output":11.4,"cache_read":0.23},"canonical_model_id":"moonshotai/kimi-k3"},"claude-opus-5.5":{"id":"claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"anthropic/claude-opus-5-5"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-flash"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.04,"output":2.08,"cache_read":0.0352},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.6-luna"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.12,"output":3.52,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.3"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"nemotron3-super":{"id":"nemotron3-super","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.12,"output":0.52,"cache_read":0.06},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"dsv4":{"id":"dsv4","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.04,"output":0.21,"cache_read":0.02},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-5"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.28,"output":1.1,"cache_read":0.056},"canonical_model_id":"minimax/MiniMax-M3"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2},"canonical_model_id":"openai/gpt-6-sol"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.34,"output":2.3,"cache_read":0.17},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.2,"output":1,"cache_read":0.01},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"claude-fable-5.1":{"id":"claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25},"canonical_model_id":"anthropic/claude-fable-5-1"}}}} \ No newline at end of file +{"ai-router":{"id":"ai-router","env":["AI_ROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.ai-router.dev/v1","name":"AI-ROUTER","doc":"https://ai-router.dev/openai-compatible-api-gateway/","models":{"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"openai/gpt-5.6-sol"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":1,"output":6,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"openai/gpt-5.6-luna"}}},"volcengine":{"id":"volcengine","env":["ARK_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://ark.cn-beijing.volces.com/api/v3","name":"Volcengine Ark","doc":"https://www.volcengine.com/docs/82379/1330310","models":{"doubao-seed-2-0-code-preview-260215":{"id":"doubao-seed-2-0-code-preview-260215","name":"Seed 2.0 Code","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"status":"deprecated","cost":{"input":0.47499,"output":2.37494,"cache_read":0.095,"tiers":[{"input":0.71248,"output":3.56241,"cache_read":0.1425,"tier":{"type":"context","size":32000}},{"input":1.42496,"output":7.12482,"cache_read":0.28499,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"deepseek-v4-pro-ga-260813":{"id":"deepseek-v4-pro-ga-260813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.3359,"output":4.00771,"cache_read":0.04453},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"doubao-seed-character-260628":{"id":"doubao-seed-character-260628","name":"Seed Character","description":"ByteDance Seed model optimized for character-driven dialogue and consistent conversational behavior","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.11875,"output":0.29687,"cache_read":0.02375,"tiers":[{"input":0.17812,"output":0.8906,"cache_read":0.02375,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"bytedance-seed/seed-character"},"glm-5-2-260617":{"id":"glm-5-2-260617","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.18747,"output":4.15615,"cache_read":0.29687},"canonical_model_id":"zhipuai/glm-5.2"},"doubao-seed-2-0-mini-260428":{"id":"doubao-seed-2-0-mini-260428","name":"Seed 2.0 Mini","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":131072},"cost":{"input":0.02969,"output":0.29687,"cache_read":0.00594,"tiers":[{"input":0.05937,"output":0.59374,"cache_read":0.01187,"tier":{"type":"context","size":32000}},{"input":0.11875,"output":1.18747,"cache_read":0.02375,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"doubao-seed-2-1-pro-260628":{"id":"doubao-seed-2-1-pro-260628","name":"Seed 2.1 Pro","description":"Flagship ByteDance Seed 2.1 model for complex multimodal reasoning, coding, and agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.8906,"output":4.45301,"cache_read":0.17812},"canonical_model_id":"bytedance-seed/seed-2.1-pro"},"doubao-seed-2-0-pro-260215":{"id":"doubao-seed-2-0-pro-260215","name":"Seed 2.0 Pro","description":"Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"status":"deprecated","cost":{"input":0.47499,"output":2.37494,"cache_read":0.095,"tiers":[{"input":0.71248,"output":3.56241,"cache_read":0.1425,"tier":{"type":"context","size":32000}},{"input":1.42496,"output":7.12482,"cache_read":0.28499,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-pro"},"deepseek-v4-flash-ga-260731":{"id":"deepseek-v4-flash-ga-260731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.4453,"output":1.3359,"cache_read":0.01484},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"doubao-seed-evolving":{"id":"doubao-seed-evolving","name":"Seed Evolving","description":"Rolling ByteDance Seed model for rapidly updated reasoning, coding, and agent capabilities","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.8906,"output":4.45301,"cache_read":0.17812},"canonical_model_id":"bytedance-seed/seed-evolving"},"doubao-seed-1-6-251015":{"id":"doubao-seed-1-6-251015","name":"Seed 1.6","description":"ByteDance Seed model for long-context reasoning, instruction following, and tool-assisted tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0.11875,"output":1.18747,"cache_read":0.02375,"tiers":[{"input":0.17812,"output":2.37494,"cache_read":0.02375,"tier":{"type":"context","size":32000}},{"input":0.35624,"output":3.56241,"cache_read":0.02375,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-6"},"doubao-seed-2-1-turbo-260628":{"id":"doubao-seed-2-1-turbo-260628","name":"Seed 2.1 Turbo","description":"Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.4453,"output":2.22651,"cache_read":0.08906},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"doubao-seed-1-8-251228":{"id":"doubao-seed-1-8-251228","name":"Seed 1.8","description":"ByteDance Seed model for multimodal reasoning, long-context analysis, and agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-12-28","last_updated":"2025-12-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0.11875,"output":1.18747,"cache_read":0.02375,"tiers":[{"input":0.17812,"output":2.37494,"cache_read":0.02375,"tier":{"type":"context","size":32000}},{"input":0.35624,"output":3.56241,"cache_read":0.02375,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-8"},"doubao-seed-2-0-lite-260428":{"id":"doubao-seed-2-0-lite-260428","name":"Seed 2.0 Lite","description":"Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":131072},"cost":{"input":0.08906,"output":0.53436,"cache_read":0.01781,"tiers":[{"input":0.13359,"output":0.80154,"cache_read":0.02672,"tier":{"type":"context","size":32000}},{"input":0.26718,"output":1.60308,"cache_read":0.05344,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"doubao-seed-1-6-vision-250815":{"id":"doubao-seed-1-6-vision-250815","name":"Seed 1.6 Vision","description":"ByteDance Seed multimodal model for image understanding, visual reasoning, and tool-assisted tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-08-15","last_updated":"2025-08-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"status":"deprecated","cost":{"input":0.11875,"output":1.18747,"cache_read":0.02375,"tiers":[{"input":0.17812,"output":2.37494,"cache_read":0.02375,"tier":{"type":"context","size":32000}},{"input":0.35624,"output":3.56241,"cache_read":0.02375,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-6-vision"},"glm-5-3-flash-260828":{"id":"glm-5-3-flash-260828","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.11875,"output":0.41563,"cache_read":0.03414},"canonical_model_id":"zhipuai/glm-5.3-flash"},"doubao-seed-1-6-flash-250828":{"id":"doubao-seed-1-6-flash-250828","name":"Seed 1.6 Flash","description":"Low-latency ByteDance Seed model for high-throughput chat, extraction, and lightweight tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"status":"deprecated","cost":{"input":0.02227,"output":0.22265,"cache_read":0.00445,"tiers":[{"input":0.04453,"output":0.4453,"cache_read":0.00445,"tier":{"type":"context","size":32000}},{"input":0.08906,"output":0.8906,"cache_read":0.00445,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-6-flash"}}},"meta":{"id":"meta","env":["META_MODEL_API_KEY"],"npm":"@ai-sdk/openai","api":"https://api.meta.ai/v1","name":"Meta","doc":"https://dev.meta.ai/docs","models":{"muse-wonder-1.0":{"id":"muse-wonder-1.0","name":"Muse Wonder","description":"Muse Wonder is a frontier multimodal reasoning model from Meta for personal and professional agents, long-horizon coding, and computer use. It combines Active Memory, which carries relevant context across conversations, with Active Focus, which keeps it working through blockers toward a goal.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"muse-spark-1.3":{"id":"muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"muse-spark-1.3-contributor":{"id":"muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.002},"canonical_model_id":"meta/muse-spark-1.3"},"muse-spark-1.1":{"id":"muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"muse-spark-1.2":{"id":"muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"muse-spark-1.2-contributor":{"id":"muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.002},"canonical_model_id":"meta/muse-spark-1.2"}}},"amazon-bedrock":{"id":"amazon-bedrock","env":["AWS_ACCESS_KEY_ID","AWS_SECRET_ACCESS_KEY","AWS_REGION","AWS_BEARER_TOKEN_BEDROCK"],"npm":"@ai-sdk/amazon-bedrock","name":"Amazon Bedrock","doc":"https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html","models":{"qwen.qwen3-coder-next":{"id":"qwen.qwen3-coder-next","name":"Qwen3 Coder Next","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.5,"output":1.2},"canonical_model_id":"alibaba/qwen3-coder-next"},"anthropic.claude-opus-4-1-20250805-v1:0":{"id":"anthropic.claude-opus-4-1-20250805-v1:0","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"anthropic.claude-sonnet-5":{"id":"anthropic.claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"in.anthropic.claude-opus-5":{"id":"in.anthropic.claude-opus-5","name":"Claude Opus 5 (India)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-5"},"meta.llama4-maverick-17b-instruct-v1:0":{"id":"meta.llama4-maverick-17b-instruct-v1:0","name":"Llama 4 Maverick 17B Instruct","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":8192},"status":"deprecated","cost":{"input":0.24,"output":0.97},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"global.anthropic.claude-sonnet-4-6":{"id":"global.anthropic.claude-sonnet-4-6","name":"Claude Sonnet 4.6 (Global)","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"jp.anthropic.claude-opus-5-5":{"id":"jp.anthropic.claude-opus-5-5","name":"Claude Opus 5.5 (JP)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.22,"cache_write":5.5},"canonical_model_id":"anthropic/claude-opus-5-5"},"deepseek.v3-v1:0":{"id":"deepseek.v3-v1:0","name":"DeepSeek-V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":81920},"status":"deprecated","cost":{"input":0.58,"output":1.68},"canonical_model_id":"deepseek/deepseek-v3.1"},"anthropic.claude-opus-5-5":{"id":"anthropic.claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"in.openai.gpt-5.6-luna":{"id":"in.openai.gpt-5.6-luna","name":"GPT-5.6 Luna (India)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.22,"output":1.32,"cache_read":0.022,"cache_write":0.275,"tiers":[{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai.gpt-5.6-luna":{"id":"openai.gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":0.22,"output":1.32,"cache_read":0.022,"cache_write":0.275,"tiers":[{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55}},"canonical_model_id":"openai/gpt-5.6-luna"},"au.anthropic.claude-opus-4-7":{"id":"au.anthropic.claude-opus-4-7","name":"Claude Opus 4.7 (AU)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-7"},"openai.gpt-6.1-sol":{"id":"openai.gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"ultrafast":{"cost":{"input":13.2,"output":66,"cache_read":0.66,"cache_write":16.5},"provider":{"body":{"service_tier":"ultrafast"}}}}},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":2.2,"output":11,"cache_read":0.11,"cache_write":2.75,"tiers":[{"input":4.4,"output":16.5,"cache_read":0.22,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":16.5,"cache_read":0.22,"cache_write":5.5}},"canonical_model_id":"openai/gpt-6.1-sol"},"anthropic.claude-haiku-4-5-20251001-v1:0":{"id":"anthropic.claude-haiku-4-5-20251001-v1:0","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"global.anthropic.claude-fable-5-1":{"id":"global.anthropic.claude-fable-5-1","name":"Claude Fable 5.1 (Global)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"openai.gpt-6-sol":{"id":"openai.gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75,"tiers":[{"input":4.4,"output":16.5,"cache_read":0.44,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":16.5,"cache_read":0.44,"cache_write":5.5}},"canonical_model_id":"openai/gpt-6-sol"},"meta.llama3-1-70b-instruct-v1:0":{"id":"meta.llama3-1-70b-instruct-v1:0","name":"Llama 3.1 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-3.1-70b-instruct"},"global.openai.gpt-6-luna":{"id":"global.openai.gpt-6-luna","name":"GPT-6 Luna (Global)","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"moonshot.kimi-k2-thinking":{"id":"moonshot.kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16000},"status":"deprecated","cost":{"input":0.6,"output":2.5},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"minimax.minimax-m2.5":{"id":"minimax.minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":98304},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.5"},"us.anthropic.claude-haiku-4-5-20251001-v1:0":{"id":"us.anthropic.claude-haiku-4-5-20251001-v1:0","name":"Claude Haiku 4.5 (US)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"jp.anthropic.claude-opus-5":{"id":"jp.anthropic.claude-opus-5","name":"Claude Opus 5 (JP)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic.claude-haiku-5-5":{"id":"anthropic.claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"us.anthropic.claude-sonnet-5-5":{"id":"us.anthropic.claude-sonnet-5-5","name":"Claude Sonnet 5.5 (US)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.11,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"us.openai.gpt-6-luna":{"id":"us.openai.gpt-6-luna","name":"GPT-6 Luna (US)","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.11,"output":0.55,"cache_read":0.011,"cache_write":0.1375,"tiers":[{"input":0.22,"output":0.825,"cache_read":0.022,"cache_write":0.275,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.22,"output":0.825,"cache_read":0.022,"cache_write":0.275}},"canonical_model_id":"openai/gpt-6-luna"},"meta.llama3-3-70b-instruct-v1:0":{"id":"meta.llama3-3-70b-instruct-v1:0","name":"Llama 3.3 70B Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"us.anthropic.claude-sonnet-4-5-20250929-v1:0":{"id":"us.anthropic.claude-sonnet-4-5-20250929-v1:0","name":"Claude Sonnet 4.5 (US)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"mistral.mistral-large-3-675b-instruct":{"id":"mistral.mistral-large-3-675b-instruct","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/mistral-large-2512"},"jp.amazon.nova-2-lite-v1:0":{"id":"jp.amazon.nova-2-lite-v1:0","name":"Nova 2 Lite (JP)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.396,"output":3.311,"cache_read":0.099,"cache_write":0.396},"canonical_model_id":"amazon/nova-2-lite"},"global.anthropic.claude-opus-4-7":{"id":"global.anthropic.claude-opus-4-7","name":"Claude Opus 4.7 (Global)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"us.anthropic.claude-opus-4-6-v1":{"id":"us.anthropic.claude-opus-4-6-v1","name":"Claude Opus 4.6 (US)","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-6"},"eu.anthropic.claude-sonnet-4-6":{"id":"eu.anthropic.claude-sonnet-4-6","name":"Claude Sonnet 4.6 (EU)","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"eu.anthropic.claude-opus-4-6-v1":{"id":"eu.anthropic.claude-opus-4-6-v1","name":"Claude Opus 4.6 (EU)","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-6"},"us.meta.llama4-maverick-17b-instruct-v1:0":{"id":"us.meta.llama4-maverick-17b-instruct-v1:0","name":"Llama 4 Maverick 17B Instruct (US)","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":8192},"status":"deprecated","cost":{"input":0.24,"output":0.97},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"us.meta.llama3-1-8b-instruct-v1:0":{"id":"us.meta.llama3-1-8b-instruct-v1:0","name":"Llama 3.1 8B Instruct (US)","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.22,"output":0.22},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"eu.amazon.nova-micro-v1:0":{"id":"eu.amazon.nova-micro-v1:0","name":"Nova Micro (EU)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.04,"output":0.16,"cache_read":0.01,"cache_write":0.04},"canonical_model_id":"amazon/nova-micro"},"apac.amazon.nova-pro-v1:0":{"id":"apac.amazon.nova-pro-v1:0","name":"Nova Pro (APAC)","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.84,"output":3.36,"cache_read":0.21,"cache_write":0.84},"canonical_model_id":"amazon/nova-pro"},"mistral.ministral-3-8b-instruct":{"id":"mistral.ministral-3-8b-instruct","name":"Ministral 3 8B","description":"Compact open vision-language model for edge deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.15,"output":0.15},"canonical_model_id":"mistral/ministral-3-8b-instruct-2512"},"us.anthropic.claude-opus-5-5":{"id":"us.anthropic.claude-opus-5-5","name":"Claude Opus 5.5 (US)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.22,"cache_write":5.5},"canonical_model_id":"anthropic/claude-opus-5-5"},"global.openai.gpt-5.6-luna":{"id":"global.openai.gpt-5.6-luna","name":"GPT-5.6 Luna (Global)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"us.writer.palmyra-x4-v1:0":{"id":"us.writer.palmyra-x4-v1:0","name":"Palmyra X4 (US)","description":"Enterprise language model for workflow automation, coding, data analysis, and tool use","family":"palmyra","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2024-10-09","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":122880,"output":8192},"status":"deprecated","cost":{"input":2.5,"output":10},"canonical_model_id":"writer/palmyra-x4"},"minimax.minimax-m2":{"id":"minimax.minimax-m2","name":"MiniMax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204608,"output":128000},"status":"deprecated","cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2"},"us.meta.llama4-scout-17b-instruct-v1:0":{"id":"us.meta.llama4-scout-17b-instruct-v1:0","name":"Llama 4 Scout 17B Instruct (US)","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":10000000,"output":8192},"status":"deprecated","cost":{"input":0.17,"output":0.66},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"us.mistral.pixtral-large-2502-v1:0":{"id":"us.mistral.pixtral-large-2502-v1:0","name":"Pixtral Large (25.02) (US)","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-08","last_updated":"2025-04-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/pixtral-large-2502"},"us.anthropic.claude-sonnet-4-20250514-v1:0":{"id":"us.anthropic.claude-sonnet-4-20250514-v1:0","name":"Claude Sonnet 4 (US)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"anthropic.claude-opus-4-6-v1":{"id":"anthropic.claude-opus-4-6-v1","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-6"},"eu.amazon.nova-pro-v1:0":{"id":"eu.amazon.nova-pro-v1:0","name":"Nova Pro (EU)","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.92,"output":3.68,"cache_read":0.23,"cache_write":0.92},"canonical_model_id":"amazon/nova-pro"},"us.anthropic.claude-opus-4-1-20250805-v1:0":{"id":"us.anthropic.claude-opus-4-1-20250805-v1:0","name":"Claude Opus 4.1 (US)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"jp.anthropic.claude-opus-4-7":{"id":"jp.anthropic.claude-opus-4-7","name":"Claude Opus 4.7 (JP)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-7"},"us.openai.gpt-5.6-terra":{"id":"us.openai.gpt-5.6-terra","name":"GPT-5.6 Terra (US)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.2,"output":13.2,"cache_read":0.22,"cache_write":2.75,"tiers":[{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5}},"canonical_model_id":"openai/gpt-5.6-terra"},"global.amazon.nova-2-lite-v1:0":{"id":"global.amazon.nova-2-lite-v1:0","name":"Nova 2 Lite (Global)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.3,"output":2.5,"cache_read":0.075,"cache_write":0.3},"canonical_model_id":"amazon/nova-2-lite"},"mistral.voxtral-mini-3b-2507":{"id":"mistral.voxtral-mini-3b-2507","name":"Voxtral Mini 3B 2507","description":"Open audio-language model for speech transcription, audio understanding, and voice-driven tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":4096},"status":"deprecated","cost":{"input":0.04,"output":0.04},"canonical_model_id":"mistral/voxtral-mini-3b-2507"},"anthropic.claude-opus-4-5-20251101-v1:0":{"id":"anthropic.claude-opus-4-5-20251101-v1:0","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"global.zai.glm-5.3":{"id":"global.zai.glm-5.3","name":"GLM-5.3 (Global)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.68,"output":5.28,"cache_read":0.312,"cache_write":2.1},"canonical_model_id":"zhipuai/glm-5.3"},"eu.anthropic.claude-opus-5-5":{"id":"eu.anthropic.claude-opus-5-5","name":"Claude Opus 5.5 (EU)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.22,"cache_write":5.5},"canonical_model_id":"anthropic/claude-opus-5-5"},"amazon.nova-pro-v1:0":{"id":"amazon.nova-pro-v1:0","name":"Nova Pro","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.8,"output":3.2,"cache_read":0.2,"cache_write":0.8},"canonical_model_id":"amazon/nova-pro"},"anthropic.claude-sonnet-4-6":{"id":"anthropic.claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"google.gemma-3-27b-it":{"id":"google.gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.23,"output":0.38},"canonical_model_id":"google/gemma-3-27b-it"},"xai.grok-4.6":{"id":"xai.grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":2.2,"output":6.6,"cache_read":0.55},"canonical_model_id":"xai/grok-4.6"},"us.openai.gpt-5.6-luna":{"id":"us.openai.gpt-5.6-luna","name":"GPT-5.6 Luna (US)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.22,"output":1.32,"cache_read":0.022,"cache_write":0.275,"tiers":[{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55}},"canonical_model_id":"openai/gpt-5.6-luna"},"global.anthropic.claude-haiku-4-5-20251001-v1:0":{"id":"global.anthropic.claude-haiku-4-5-20251001-v1:0","name":"Claude Haiku 4.5 (Global)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"global.openai.gpt-5.6-terra":{"id":"global.openai.gpt-5.6-terra","name":"GPT-5.6 Terra (Global)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"us.amazon.nova-lite-v1:0":{"id":"us.amazon.nova-lite-v1:0","name":"Nova Lite (US)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.06,"output":0.24,"cache_read":0.015,"cache_write":0.06},"canonical_model_id":"amazon/nova-lite"},"minimax.minimax-m2.1":{"id":"minimax.minimax-m2.1","name":"MiniMax-M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"status":"deprecated","cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.1"},"au.anthropic.claude-sonnet-5":{"id":"au.anthropic.claude-sonnet-5","name":"Claude Sonnet 5 (AU)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5"},"us.writer.palmyra-x5-v1:0":{"id":"us.writer.palmyra-x5-v1:0","name":"Palmyra X5 (US)","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"palmyra","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1040000,"input":1040000,"output":8192},"status":"deprecated","cost":{"input":0.6,"output":6},"canonical_model_id":"writer/palmyra-x5"},"amazon.nova-2-lite-v1:0":{"id":"amazon.nova-2-lite-v1:0","name":"Nova 2 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.33,"output":2.75,"cache_read":0.0825,"cache_write":0.33},"canonical_model_id":"amazon/nova-2-lite"},"jp.anthropic.claude-sonnet-5":{"id":"jp.anthropic.claude-sonnet-5","name":"Claude Sonnet 5 (JP)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5"},"us.xai.grok-4.6":{"id":"us.xai.grok-4.6","name":"Grok 4.6 (US)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2.2,"output":6.6,"cache_read":0.55},"canonical_model_id":"xai/grok-4.6"},"anthropic.claude-sonnet-5-5":{"id":"anthropic.claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"eu.anthropic.claude-opus-4-5-20251101-v1:0":{"id":"eu.anthropic.claude-opus-4-5-20251101-v1:0","name":"Claude Opus 4.5 (EU)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"apac.anthropic.claude-sonnet-4-20250514-v1:0":{"id":"apac.anthropic.claude-sonnet-4-20250514-v1:0","name":"Claude Sonnet 4 (APAC)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"global.openai.gpt-6-sol":{"id":"global.openai.gpt-6-sol","name":"GPT-6 Sol (Global)","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"us.openai.gpt-6.1-sol":{"id":"us.openai.gpt-6.1-sol","name":"GPT-6.1 Sol (US)","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"ultrafast":{"cost":{"input":13.2,"output":66,"cache_read":0.66,"cache_write":16.5},"provider":{"body":{"additionalModelRequestFields":{"service_tier":"ultrafast"}}}}}},"cost":{"input":2.2,"output":11,"cache_read":0.11,"cache_write":2.75,"tiers":[{"input":4.4,"output":16.5,"cache_read":0.22,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":16.5,"cache_read":0.22,"cache_write":5.5}},"canonical_model_id":"openai/gpt-6.1-sol"},"us.xai.grok-4.7":{"id":"us.xai.grok-4.7","name":"Grok 4.7 (US)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2.2,"output":6.6,"cache_read":0.55},"canonical_model_id":"xai/grok-4.7"},"jp.anthropic.claude-haiku-4-5-20251001-v1:0":{"id":"jp.anthropic.claude-haiku-4-5-20251001-v1:0","name":"Claude Haiku 4.5 (JP)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"qwen.qwen3-32b-v1:0":{"id":"qwen.qwen3-32b-v1:0","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-32b"},"in.anthropic.claude-sonnet-5":{"id":"in.anthropic.claude-sonnet-5","name":"Claude Sonnet 5 (India)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5"},"us.anthropic.claude-haiku-5-5":{"id":"us.anthropic.claude-haiku-5-5","name":"Claude Haiku 5.5 (US)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.11,"output":0.55,"cache_read":0.011,"cache_write":0.1375,"tiers":[{"input":0.55,"output":2.75,"cache_read":0.055,"cache_write":0.6875,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"au.anthropic.claude-haiku-4-5-20251001-v1:0":{"id":"au.anthropic.claude-haiku-4-5-20251001-v1:0","name":"Claude Haiku 4.5 (AU)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"qwen.qwen3-coder-30b-a3b-v1:0":{"id":"qwen.qwen3-coder-30b-a3b-v1:0","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"status":"deprecated","cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"au.anthropic.claude-sonnet-4-6":{"id":"au.anthropic.claude-sonnet-4-6","name":"AU Anthropic Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"openai.gpt-6-luna":{"id":"openai.gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":0.11,"output":0.55,"cache_read":0.011,"cache_write":0.1375,"tiers":[{"input":0.22,"output":0.825,"cache_read":0.022,"cache_write":0.275,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.22,"output":0.825,"cache_read":0.022,"cache_write":0.275}},"canonical_model_id":"openai/gpt-6-luna"},"global.anthropic.claude-opus-5-5":{"id":"global.anthropic.claude-opus-5-5","name":"Claude Opus 5.5 (Global)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"global.xai.grok-4.7":{"id":"global.xai.grok-4.7","name":"Grok 4.7 (Global)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"eu.amazon.nova-2-lite-v1:0":{"id":"eu.amazon.nova-2-lite-v1:0","name":"Nova 2 Lite (EU)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.374,"output":3.157,"cache_read":0.0935,"cache_write":0.374},"canonical_model_id":"amazon/nova-2-lite"},"zai.glm-4.7-flash":{"id":"zai.glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"status":"deprecated","cost":{"input":0.07,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"openai.gpt-5.6-terra":{"id":"openai.gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":2.2,"output":13.2,"cache_read":0.22,"cache_write":2.75,"tiers":[{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5}},"canonical_model_id":"openai/gpt-5.6-terra"},"us.openai.gpt-5.6-sol":{"id":"us.openai.gpt-5.6-sol","name":"GPT-5.6 Sol (US)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.44,"cache_write":5.5,"tiers":[{"input":8.8,"output":33,"cache_read":0.88,"cache_write":11,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8.8,"output":33,"cache_read":0.88,"cache_write":11}},"canonical_model_id":"openai/gpt-5.6-sol"},"us.amazon.nova-premier-v1:0":{"id":"us.amazon.nova-premier-v1:0","name":"Nova Premier (US)","description":"Multimodal model for complex analysis, long-context understanding, tool use, and model distillation","family":"nova","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2025-04-30","last_updated":"2025-04-30","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":10000},"status":"deprecated","cost":{"input":2.5,"output":12.5,"cache_read":0.625,"cache_write":2.5},"canonical_model_id":"amazon/nova-premier"},"anthropic.claude-opus-4-7":{"id":"anthropic.claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"nvidia.nemotron-super-3-120b":{"id":"nvidia.nemotron-super-3-120b","name":"NVIDIA Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.15,"output":0.65},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"qwen.qwen3-235b-a22b-2507-v1:0":{"id":"qwen.qwen3-235b-a22b-2507-v1:0","name":"Qwen3 235B-A22B Instruct 2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.22,"output":0.88},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"writer.palmyra-x4-v1:0":{"id":"writer.palmyra-x4-v1:0","name":"Palmyra X4","description":"Enterprise language model for workflow automation, coding, data analysis, and tool use","family":"palmyra","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2024-10-09","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":122880,"output":8192},"status":"deprecated","cost":{"input":2.5,"output":10},"canonical_model_id":"writer/palmyra-x4"},"apac.amazon.nova-lite-v1:0":{"id":"apac.amazon.nova-lite-v1:0","name":"Nova Lite (APAC)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.063,"output":0.252,"cache_read":0.01575,"cache_write":0.063},"canonical_model_id":"amazon/nova-lite"},"global.xai.grok-4.6":{"id":"global.xai.grok-4.6","name":"Grok 4.6 (Global)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"jp.anthropic.claude-opus-4-8":{"id":"jp.anthropic.claude-opus-4-8","name":"Claude Opus 4.8 (JP)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-8"},"mistral.voxtral-small-24b-2507":{"id":"mistral.voxtral-small-24b-2507","name":"Voxtral Small 24B 2507","description":"Open audio-language model for speech transcription, audio understanding, and voice-driven tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"status":"deprecated","cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/voxtral-small-24b-2507"},"global.moonshotai.kimi-k3":{"id":"global.moonshotai.kimi-k3","name":"Kimi K3 (Global)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"moonshotai/kimi-k3"},"meta.llama3-1-8b-instruct-v1:0":{"id":"meta.llama3-1-8b-instruct-v1:0","name":"Llama 3.1 8B Instruct","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.22,"output":0.22},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"anthropic.claude-opus-4-8":{"id":"anthropic.claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"global.anthropic.claude-opus-4-5-20251101-v1:0":{"id":"global.anthropic.claude-opus-4-5-20251101-v1:0","name":"Claude Opus 4.5 (Global)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"us.anthropic.claude-sonnet-5":{"id":"us.anthropic.claude-sonnet-5","name":"Claude Sonnet 5 (US)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5"},"us.meta.llama3-1-70b-instruct-v1:0":{"id":"us.meta.llama3-1-70b-instruct-v1:0","name":"Llama 3.1 70B Instruct (US)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-3.1-70b-instruct"},"global.anthropic.claude-sonnet-5":{"id":"global.anthropic.claude-sonnet-5","name":"Claude Sonnet 5 (Global)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"eu.mistral.pixtral-large-2502-v1:0":{"id":"eu.mistral.pixtral-large-2502-v1:0","name":"Pixtral Large (25.02) (EU)","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-08","last_updated":"2025-04-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/pixtral-large-2502"},"eu.anthropic.claude-sonnet-4-20250514-v1:0":{"id":"eu.anthropic.claude-sonnet-4-20250514-v1:0","name":"Claude Sonnet 4 (EU)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"global.anthropic.claude-opus-5":{"id":"global.anthropic.claude-opus-5","name":"Claude Opus 5 (Global)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"us.anthropic.claude-opus-4-7":{"id":"us.anthropic.claude-opus-4-7","name":"Claude Opus 4.7 (US)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-7"},"au.anthropic.claude-haiku-5-5":{"id":"au.anthropic.claude-haiku-5-5","name":"Claude Haiku 5.5 (AU)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.11,"output":0.55,"cache_read":0.011,"cache_write":0.1375,"tiers":[{"input":0.55,"output":2.75,"cache_read":0.055,"cache_write":0.6875,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"eu.anthropic.claude-haiku-4-5-20251001-v1:0":{"id":"eu.anthropic.claude-haiku-4-5-20251001-v1:0","name":"Claude Haiku 4.5 (EU)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"eu.anthropic.claude-fable-5":{"id":"eu.anthropic.claude-fable-5","name":"Claude Fable 5 (EU)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5"},"openai.gpt-5.4":{"id":"openai.gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-06-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":2.75,"output":16.5,"cache_read":0.275},"canonical_model_id":"openai/gpt-5.4"},"us.anthropic.claude-fable-5":{"id":"us.anthropic.claude-fable-5","name":"Claude Fable 5 (US)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5"},"au.anthropic.claude-opus-5-5":{"id":"au.anthropic.claude-opus-5-5","name":"Claude Opus 5.5 (AU)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.22,"cache_write":5.5},"canonical_model_id":"anthropic/claude-opus-5-5"},"openai.gpt-5.5":{"id":"openai.gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-06-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":5.5,"output":33,"cache_read":0.55},"canonical_model_id":"openai/gpt-5.5"},"eu.anthropic.claude-sonnet-5-5":{"id":"eu.anthropic.claude-sonnet-5-5","name":"Claude Sonnet 5.5 (EU)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.11,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"eu.amazon.nova-lite-v1:0":{"id":"eu.amazon.nova-lite-v1:0","name":"Nova Lite (EU)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.069,"output":0.276,"cache_read":0.01725,"cache_write":0.069},"canonical_model_id":"amazon/nova-lite"},"openai.gpt-oss-20b":{"id":"openai.gpt-oss-20b","name":"gpt-oss-20b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/v1","shape":"responses"},"cost":{"input":0.07,"output":0.3},"canonical_model_id":"openai/gpt-oss-20b"},"us.moonshotai.kimi-k3":{"id":"us.moonshotai.kimi-k3","name":"Kimi K3 (US)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"moonshotai/kimi-k3"},"google.gemma-3-12b-it":{"id":"google.gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.09,"output":0.29},"canonical_model_id":"google/gemma-3-12b-it"},"us.anthropic.claude-opus-4-5-20251101-v1:0":{"id":"us.anthropic.claude-opus-4-5-20251101-v1:0","name":"Claude Opus 4.5 (US)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"global.anthropic.claude-sonnet-4-5-20250929-v1:0":{"id":"global.anthropic.claude-sonnet-4-5-20250929-v1:0","name":"Claude Sonnet 4.5 (Global)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"zai.glm-4.7":{"id":"zai.glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"status":"deprecated","cost":{"input":0.6,"output":2.2},"canonical_model_id":"zhipuai/glm-4.7"},"google.gemma-3-4b-it":{"id":"google.gemma-3-4b-it","name":"Gemma 3 4B IT","description":"Open multimodal Gemma instruction model for efficient text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"status":"deprecated","cost":{"input":0.04,"output":0.08},"canonical_model_id":"google/gemma-3-4b-it"},"global.openai.gpt-5.6-sol":{"id":"global.openai.gpt-5.6-sol","name":"GPT-5.6 Sol (Global)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"us.amazon.nova-pro-v1:0":{"id":"us.amazon.nova-pro-v1:0","name":"Nova Pro (US)","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.8,"output":3.2,"cache_read":0.2,"cache_write":0.8},"canonical_model_id":"amazon/nova-pro"},"deepseek.v3.2":{"id":"deepseek.v3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2026-02-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":81920},"cost":{"input":0.62,"output":1.85},"canonical_model_id":"deepseek/deepseek-v3.2"},"openai.gpt-oss-safeguard-120b":{"id":"openai.gpt-oss-safeguard-120b","name":"GPT OSS Safeguard 120B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-safeguard-120b"},"ca.amazon.nova-lite-v1:0":{"id":"ca.amazon.nova-lite-v1:0","name":"Nova Lite (CA)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.064,"output":0.256,"cache_read":0.016,"cache_write":0.064},"canonical_model_id":"amazon/nova-lite"},"us.amazon.nova-2-lite-v1:0":{"id":"us.amazon.nova-2-lite-v1:0","name":"Nova 2 Lite (US)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.33,"output":2.75,"cache_read":0.0825,"cache_write":0.33},"canonical_model_id":"amazon/nova-2-lite"},"us.openai.gpt-6-sol":{"id":"us.openai.gpt-6-sol","name":"GPT-6 Sol (US)","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75,"tiers":[{"input":4.4,"output":16.5,"cache_read":0.44,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":16.5,"cache_read":0.44,"cache_write":5.5}},"canonical_model_id":"openai/gpt-6-sol"},"us.anthropic.claude-opus-5":{"id":"us.anthropic.claude-opus-5","name":"Claude Opus 5 (US)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-5"},"global.openai.gpt-6.1-sol":{"id":"global.openai.gpt-6.1-sol","name":"GPT-6.1 Sol (Global)","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"ultrafast":{"cost":{"input":12,"output":60,"cache_read":0.6,"cache_write":15},"provider":{"body":{"additionalModelRequestFields":{"service_tier":"ultrafast"}}}}}},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"au.anthropic.claude-opus-5":{"id":"au.anthropic.claude-opus-5","name":"Claude Opus 5 (AU)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-5"},"global.anthropic.claude-opus-4-6-v1":{"id":"global.anthropic.claude-opus-4-6-v1","name":"Claude Opus 4.6 (Global)","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"us-gov.openai.gpt-oss-120b-1:0":{"id":"us-gov.openai.gpt-oss-120b-1:0","name":"gpt-oss-120b (GovCloud)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.18,"output":0.72},"canonical_model_id":"openai/gpt-oss-120b"},"openai.gpt-oss-120b":{"id":"openai.gpt-oss-120b","name":"gpt-oss-120b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/v1","shape":"responses"},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"moonshotai.kimi-k2.5":{"id":"moonshotai.kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-02-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.6,"output":3},"canonical_model_id":"moonshotai/kimi-k2.5"},"jp.anthropic.claude-haiku-5-5":{"id":"jp.anthropic.claude-haiku-5-5","name":"Claude Haiku 5.5 (JP)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.11,"output":0.55,"cache_read":0.011,"cache_write":0.1375,"tiers":[{"input":0.55,"output":2.75,"cache_read":0.055,"cache_write":0.6875,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic.claude-fable-5":{"id":"anthropic.claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"openai.gpt-6-astra":{"id":"openai.gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"ultrafast":{"cost":{"input":66,"output":330,"cache_read":6.6,"cache_write":82.5},"provider":{"body":{"service_tier":"ultrafast"}}}}},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75,"tiers":[{"input":22,"output":82.5,"cache_read":2.2,"cache_write":27.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":22,"output":82.5,"cache_read":2.2,"cache_write":27.5}},"canonical_model_id":"openai/gpt-6-astra"},"google.gemma-4-e2b":{"id":"google.gemma-4-e2b","name":"Gemma 4 E2B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":0.04,"output":0.08},"canonical_model_id":"google/gemma-4-E2B-it"},"us.anthropic.claude-opus-4-8":{"id":"us.anthropic.claude-opus-4-8","name":"Claude Opus 4.8 (US)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-8"},"mistral.pixtral-large-2502-v1:0":{"id":"mistral.pixtral-large-2502-v1:0","name":"Pixtral Large (25.02)","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-08","last_updated":"2025-04-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/pixtral-large-2502"},"global.anthropic.claude-haiku-5-5":{"id":"global.anthropic.claude-haiku-5-5","name":"Claude Haiku 5.5 (Global)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"eu.anthropic.claude-sonnet-4-5-20250929-v1:0":{"id":"eu.anthropic.claude-sonnet-4-5-20250929-v1:0","name":"Claude Sonnet 4.5 (EU)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"eu.anthropic.claude-sonnet-5":{"id":"eu.anthropic.claude-sonnet-5","name":"Claude Sonnet 5 (EU)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5"},"eu.anthropic.claude-haiku-5-5":{"id":"eu.anthropic.claude-haiku-5-5","name":"Claude Haiku 5.5 (EU)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.11,"output":0.55,"cache_read":0.011,"cache_write":0.1375,"tiers":[{"input":0.55,"output":2.75,"cache_read":0.055,"cache_write":0.6875,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"meta.llama4-scout-17b-instruct-v1:0":{"id":"meta.llama4-scout-17b-instruct-v1:0","name":"Llama 4 Scout 17B Instruct","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":10000000,"output":8192},"status":"deprecated","cost":{"input":0.17,"output":0.66},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"openai.gpt-oss-120b-1:0":{"id":"openai.gpt-oss-120b-1:0","name":"gpt-oss-120b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":128000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"au.anthropic.claude-opus-4-6-v1":{"id":"au.anthropic.claude-opus-4-6-v1","name":"AU Anthropic Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-6"},"us.amazon.nova-micro-v1:0":{"id":"us.amazon.nova-micro-v1:0","name":"Nova Micro (US)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.035,"output":0.14,"cache_read":0.00875,"cache_write":0.035},"canonical_model_id":"amazon/nova-micro"},"nvidia.nemotron-nano-3-30b":{"id":"nvidia.nemotron-nano-3-30b","name":"NVIDIA Nemotron Nano 3 30B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.06,"output":0.24},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"qwen.qwen3-next-80b-a3b":{"id":"qwen.qwen3-next-80b-a3b","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-11","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262000},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"us.deepseek.r1-v1:0":{"id":"us.deepseek.r1-v1:0","name":"DeepSeek-R1 (US)","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"status":"deprecated","cost":{"input":1.35,"output":5.4},"canonical_model_id":"deepseek/deepseek-r1"},"anthropic.claude-fable-5-1":{"id":"anthropic.claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"jp.anthropic.claude-sonnet-4-5-20250929-v1:0":{"id":"jp.anthropic.claude-sonnet-4-5-20250929-v1:0","name":"Claude Sonnet 4.5 (JP)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"amazon.nova-micro-v1:0":{"id":"amazon.nova-micro-v1:0","name":"Nova Micro","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.035,"output":0.14,"cache_read":0.00875,"cache_write":0.035},"canonical_model_id":"amazon/nova-micro"},"us.openai.gpt-6-astra":{"id":"us.openai.gpt-6-astra","name":"GPT-6 Astra (US)","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"ultrafast":{"cost":{"input":66,"output":330,"cache_read":6.6,"cache_write":82.5},"provider":{"body":{"additionalModelRequestFields":{"service_tier":"ultrafast"}}}}}},"cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75,"tiers":[{"input":22,"output":82.5,"cache_read":2.2,"cache_write":27.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":22,"output":82.5,"cache_read":2.2,"cache_write":27.5}},"canonical_model_id":"openai/gpt-6-astra"},"global.openai.gpt-6-astra":{"id":"global.openai.gpt-6-astra","name":"GPT-6 Astra (Global)","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"ultrafast":{"cost":{"input":60,"output":300,"cache_read":6,"cache_write":75},"provider":{"body":{"additionalModelRequestFields":{"service_tier":"ultrafast"}}}}}},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"mistral.devstral-2-123b":{"id":"mistral.devstral-2-123b","name":"Devstral 2 123B","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"status":"deprecated","cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/devstral-2512"},"google.gemma-4-31b":{"id":"google.gemma-4-31b","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"au.anthropic.claude-sonnet-4-5-20250929-v1:0":{"id":"au.anthropic.claude-sonnet-4-5-20250929-v1:0","name":"Claude Sonnet 4.5 (AU)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"google.gemma-4-26b-a4b":{"id":"google.gemma-4-26b-a4b","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":0.13,"output":0.4},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"nvidia.nemotron-nano-12b-v2":{"id":"nvidia.nemotron-nano-12b-v2","name":"NVIDIA Nemotron Nano 12B v2 VL BF16","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-28","last_updated":"2025-10-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.2,"output":0.6},"canonical_model_id":"nvidia/nemotron-nano-12b-v2-vl"},"global.anthropic.claude-sonnet-5-5":{"id":"global.anthropic.claude-sonnet-5-5","name":"Claude Sonnet 5.5 (Global)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"openai.gpt-oss-safeguard-20b":{"id":"openai.gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.07,"output":0.2},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"openai.gpt-5.6-sol":{"id":"openai.gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":4.4,"output":22,"cache_read":0.44,"cache_write":5.5,"tiers":[{"input":8.8,"output":33,"cache_read":0.88,"cache_write":11,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8.8,"output":33,"cache_read":0.88,"cache_write":11}},"canonical_model_id":"openai/gpt-5.6-sol"},"eu.anthropic.claude-opus-5":{"id":"eu.anthropic.claude-opus-5","name":"Claude Opus 5 (EU)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-5"},"jp.anthropic.claude-sonnet-4-6":{"id":"jp.anthropic.claude-sonnet-4-6","name":"Claude Sonnet 4.6 (JP)","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"xai.grok-4.3":{"id":"xai.grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-06-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"provider":{"npm":"@ai-sdk/amazon-bedrock/mantle","api":"https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1","shape":"responses"},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"eu.anthropic.claude-opus-4-7":{"id":"eu.anthropic.claude-opus-4-7","name":"Claude Opus 4.7 (EU)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-7"},"deepseek.r1-v1:0":{"id":"deepseek.r1-v1:0","name":"DeepSeek-R1","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"status":"deprecated","cost":{"input":1.35,"output":5.4},"canonical_model_id":"deepseek/deepseek-r1"},"zai.glm-5":{"id":"zai.glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1,"output":3.2},"canonical_model_id":"zhipuai/glm-5"},"qwen.qwen3-coder-480b-a35b-v1:0":{"id":"qwen.qwen3-coder-480b-a35b-v1:0","name":"Qwen3-Coder 480B-A35B Instruct","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"status":"deprecated","cost":{"input":0.45,"output":1.8},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"mistral.ministral-3-3b-instruct":{"id":"mistral.ministral-3-3b-instruct","name":"Ministral 3 3B","description":"Compact open vision-language model for edge deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.1,"output":0.1},"canonical_model_id":"mistral/ministral-3-3b-instruct-2512"},"us.anthropic.claude-sonnet-4-6":{"id":"us.anthropic.claude-sonnet-4-6","name":"Claude Sonnet 4.6 (US)","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"global.anthropic.claude-fable-5":{"id":"global.anthropic.claude-fable-5","name":"Claude Fable 5 (Global)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"nvidia.nemotron-nano-9b-v2":{"id":"nvidia.nemotron-nano-9b-v2","name":"NVIDIA Nemotron Nano 9B v2","description":"Compact Nemotron model for efficient reasoning and deployable AI agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09","release_date":"2025-08-18","last_updated":"2025-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.06,"output":0.23},"canonical_model_id":"nvidia/nemotron-nano-9b-v2"},"anthropic.claude-sonnet-4-5-20250929-v1:0":{"id":"anthropic.claude-sonnet-4-5-20250929-v1:0","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"us-gov.openai.gpt-oss-20b-1:0":{"id":"us-gov.openai.gpt-oss-20b-1:0","name":"gpt-oss-20b (GovCloud)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.084,"output":0.36},"canonical_model_id":"openai/gpt-oss-20b"},"eu.anthropic.claude-opus-4-8":{"id":"eu.anthropic.claude-opus-4-8","name":"Claude Opus 4.8 (EU)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-8"},"global.anthropic.claude-opus-4-8":{"id":"global.anthropic.claude-opus-4-8","name":"Claude Opus 4.8 (Global)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"au.anthropic.claude-opus-4-8":{"id":"au.anthropic.claude-opus-4-8","name":"Claude Opus 4.8 (AU)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-8"},"apac.amazon.nova-micro-v1:0":{"id":"apac.amazon.nova-micro-v1:0","name":"Nova Micro (APAC)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.037,"output":0.148,"cache_read":0.00925,"cache_write":0.037},"canonical_model_id":"amazon/nova-micro"},"openai.gpt-oss-20b-1:0":{"id":"openai.gpt-oss-20b-1:0","name":"gpt-oss-20b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":128000},"cost":{"input":0.07,"output":0.3},"canonical_model_id":"openai/gpt-oss-20b"},"us.zai.glm-5.3":{"id":"us.zai.glm-5.3","name":"GLM-5.3 (US)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.848,"output":5.808,"cache_read":0.3432,"cache_write":2.31},"canonical_model_id":"zhipuai/glm-5.3"},"qwen.qwen3-vl-235b-a22b":{"id":"qwen.qwen3-vl-235b-a22b","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262000},"cost":{"input":0.53,"output":2.66},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"global.anthropic.claude-sonnet-4-20250514-v1:0":{"id":"global.anthropic.claude-sonnet-4-20250514-v1:0","name":"Claude Sonnet 4 (Global)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"in.openai.gpt-5.6-terra":{"id":"in.openai.gpt-5.6-terra","name":"GPT-5.6 Terra (India)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.2,"output":13.2,"cache_read":0.22,"cache_write":2.75,"tiers":[{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5}},"canonical_model_id":"openai/gpt-5.6-terra"},"mistral.ministral-3-14b-instruct":{"id":"mistral.ministral-3-14b-instruct","name":"Ministral 14B 3.0","description":"Open vision-language model for efficient local deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.2,"output":0.2},"canonical_model_id":"mistral/ministral-3-14b-instruct-2512"},"us.anthropic.claude-fable-5-1":{"id":"us.anthropic.claude-fable-5-1","name":"Claude Fable 5.1 (US)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":0.275,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5-1"},"in.anthropic.claude-haiku-4-5-20251001-v1:0":{"id":"in.anthropic.claude-haiku-4-5-20251001-v1:0","name":"Claude Haiku 4.5 (India)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"us.meta.llama3-3-70b-instruct-v1:0":{"id":"us.meta.llama3-3-70b-instruct-v1:0","name":"Llama 3.3 70B Instruct (US)","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"writer.palmyra-x5-v1:0":{"id":"writer.palmyra-x5-v1:0","name":"Palmyra X5","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"palmyra","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1040000,"input":1040000,"output":8192},"status":"deprecated","cost":{"input":0.6,"output":6},"canonical_model_id":"writer/palmyra-x5"},"anthropic.claude-opus-5":{"id":"anthropic.claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"mistral.magistral-small-2509":{"id":"mistral.magistral-small-2509","name":"Magistral Small 1.2","description":"Open multimodal reasoning model for transparent analysis of text and images","family":"magistral","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-18","last_updated":"2025-09-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":40000},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/magistral-small-2509"},"amazon.nova-lite-v1:0":{"id":"amazon.nova-lite-v1:0","name":"Nova Lite","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.06,"output":0.24,"cache_read":0.015,"cache_write":0.06},"canonical_model_id":"amazon/nova-lite"}}},"salad-cloud":{"id":"salad-cloud","env":["SALAD_CLOUD_API_KEY"],"npm":"@saladtechnologies-oss/ai-sdk-provider","name":"SaladCloud AI Gateway","doc":"https://docs.salad.com/ai-gateway/explanation/overview","models":{"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Qwen MoE for agentic tasks, complex reasoning, code generation, and instruction following","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":262144},"cost":{"input":0.09,"output":0.6},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"}}},"io-net":{"id":"io-net","env":["IOINTELLIGENCE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.intelligence.io.solutions/api/v1","name":"IO.NET","doc":"https://io.net/docs/guides/intelligence/io-intelligence","models":{"moonshotai/Kimi-K2-Thinking":{"id":"moonshotai/Kimi-K2-Thinking","name":"Kimi K2 Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":4096},"cost":{"input":0.6,"output":2.5,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"moonshotai/Kimi-K2-Instruct-0905":{"id":"moonshotai/Kimi-K2-Instruct-0905","name":"Kimi K2 Instruct","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":4096},"cost":{"input":0.57,"output":2.3,"cache_read":0.285},"canonical_model_id":"moonshotai/kimi-k2-0905"},"mistralai/Magistral-Small-2506":{"id":"mistralai/Magistral-Small-2506","name":"Magistral Small 2506","description":"Mistral reasoning model for transparent analysis, math, and complex decisions","family":"magistral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.5,"output":1.5,"cache_read":0.25,"cache_write":1},"canonical_model_id":"mistral/magistral-small-2506"},"mistralai/Mistral-Nemo-Instruct-2407":{"id":"mistralai/Mistral-Nemo-Instruct-2407","name":"Mistral Nemo Instruct 2407","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-05","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.0514,"output":0.082,"cache_read":0.0257},"canonical_model_id":"mistral/mistral-nemo"},"mistralai/Devstral-Small-2505":{"id":"mistralai/Devstral-Small-2505","name":"Devstral Small 2505","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-05-01","last_updated":"2025-05-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.05,"output":0.22,"cache_read":0.025,"cache_write":0.1}},"mistralai/Mistral-Large-Instruct-2411":{"id":"mistralai/Mistral-Large-Instruct-2411","name":"Mistral Large Instruct 2411","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-11-18","last_updated":"2024-11-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":2,"output":6,"cache_read":1,"cache_write":4},"canonical_model_id":"mistral/mistral-large-2411"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT-OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":4096},"cost":{"input":0.055,"output":0.194,"cache_read":0.0275},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT-OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.188,"output":0.7,"cache_read":0.094},"canonical_model_id":"openai/gpt-oss-120b"},"meta-llama/Llama-3.3-70B-Instruct":{"id":"meta-llama/Llama-3.3-70B-Instruct","name":"Llama 3.3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.5466,"output":1.0226,"cache_read":0.2733},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8":{"id":"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8","name":"Llama 4 Maverick 17B 128E Instruct","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":430000,"output":4096},"cost":{"input":0.26875,"output":0.82625,"cache_read":0.134375},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"meta-llama/Llama-3.2-90B-Vision-Instruct":{"id":"meta-llama/Llama-3.2-90B-Vision-Instruct","name":"Llama 3.2 90B Vision Instruct","description":"Open Llama multimodal model for image understanding and text reasoning","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4096},"status":"deprecated","cost":{"input":0.35,"output":0.4,"cache_read":0.175,"cache_write":0.7}},"zai-org/GLM-4.6":{"id":"zai-org/GLM-4.6","name":"GLM 4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":4096},"cost":{"input":0.52,"output":2.0375,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-4.6"},"Intel/Qwen3-Coder-480B-A35B-Instruct-int4-mixed-ar":{"id":"Intel/Qwen3-Coder-480B-A35B-Instruct-int4-mixed-ar","name":"Qwen 3 Coder 480B","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":106000,"output":4096},"cost":{"input":0.46125,"output":2.29375,"cache_read":0.230625},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"deepseek-ai/DeepSeek-R1-0528":{"id":"deepseek-ai/DeepSeek-R1-0528","name":"DeepSeek R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.523667,"output":2.205333,"cache_read":0.261834},"canonical_model_id":"deepseek/deepseek-r1-0528"},"Qwen/Qwen3-235B-A22B-Thinking-2507":{"id":"Qwen/Qwen3-235B-A22B-Thinking-2507","name":"Qwen 3 235B Thinking","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":4096},"status":"deprecated","cost":{"input":0.11,"output":0.6,"cache_read":0.055,"cache_write":0.22},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"Qwen/Qwen3-Next-80B-A3B-Instruct":{"id":"Qwen/Qwen3-Next-80B-A3B-Instruct","name":"Qwen 3 Next 80B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":4096},"cost":{"input":0.11875,"output":1.2625,"cache_read":0.059375},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"Qwen/Qwen2.5-VL-32B-Instruct":{"id":"Qwen/Qwen2.5-VL-32B-Instruct","name":"Qwen 2.5 VL 32B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2024-11-01","last_updated":"2024-11-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"status":"deprecated","cost":{"input":0.05,"output":0.22,"cache_read":0.025,"cache_write":0.1}}}},"sakana":{"id":"sakana","env":["SAKANA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.sakana.ai/v1","name":"Sakana AI","doc":"https://console.sakana.ai/models","models":{"sakana-namazu":{"id":"sakana-namazu","name":"Sakana Namazu","description":"Japanese-specialized reasoning model based on Kimi K2.6 and tuned for Japanese language, culture, and business workflows","family":"sakana-namazu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.95,"output":4,"cache_read":0.15},"canonical_model_id":"sakana/sakana-namazu"},"fugu":{"id":"fugu","name":"Fugu","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"provider":{"shape":"responses"},"canonical_model_id":"sakana/fugu"},"fugu-ultra":{"id":"fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"provider":{"shape":"responses"},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra"},"fugu-ultra-20260615":{"id":"fugu-ultra-20260615","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"provider":{"shape":"responses"},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra"}}},"azure":{"id":"azure","env":["AZURE_RESOURCE_NAME","AZURE_API_KEY"],"npm":"@ai-sdk/azure","name":"Azure","doc":"https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models","models":{"grok-4-20-reasoning":{"id":"grok-4-20-reasoning","name":"Grok 4.20 (Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-04-08","last_updated":"2026-04-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":8192},"status":"beta","cost":{"input":2,"output":6}},"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-31","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"anthropic/claude-opus-4-8"},"llama-4-scout-17b-16e-instruct":{"id":"llama-4-scout-17b-16e-instruct","name":"Llama 4 Scout 17B 16E Instruct","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.2,"output":0.78}},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-14","last_updated":"2025-11-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":2.5,"output":20,"cache_read":0.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"gpt-5-pro":{"id":"gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"status":"beta","cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"cohere-command-a":{"id":"cohere-command-a","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":2.5,"output":10}},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"o3-mini":{"id":"o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":4,"output":24,"cache_read":0.4,"cache_write":5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"phi-4-reasoning-plus":{"id":"phi-4-reasoning-plus","name":"Phi-4-reasoning-plus","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"phi","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.125,"output":0.5}},"llama-4-maverick-17b-128e-instruct-fp8":{"id":"llama-4-maverick-17b-128e-instruct-fp8","name":"Llama 4 Maverick 17B 128E Instruct FP8","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":16384},"cost":{"input":0.25,"output":1}},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"gpt-chat-latest":{"id":"gpt-chat-latest","name":"GPT Chat Latest","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-05-05","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"beta","cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5-instant"},"codex-mini":{"id":"codex-mini","name":"Codex Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-04","release_date":"2025-05-16","last_updated":"2025-05-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.5,"output":6,"cache_read":0.375}},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25}},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-01-14","last_updated":"2026-01-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"phi-4-reasoning":{"id":"phi-4-reasoning","name":"Phi-4-reasoning","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"phi","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.125,"output":0.5}},"phi-4-mini-reasoning":{"id":"phi-4-mini-reasoning","name":"Phi-4-mini-reasoning","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"phi","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.075,"output":0.3}},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"text-embedding-3-large":{"id":"text-embedding-3-large","name":"text-embedding-3-large","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":3072},"cost":{"input":0.13,"output":0}},"gpt-5.4-pro":{"id":"gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"gpt-3.5-turbo-instruct":{"id":"gpt-3.5-turbo-instruct","name":"GPT-3.5 Turbo Instruct","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2021-08","release_date":"2023-09-21","last_updated":"2023-09-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4096,"output":4096},"status":"deprecated","cost":{"input":1.5,"output":2}},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek-V4-Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.19,"output":0.51},"canonical_model_id":"deepseek/deepseek-v4-flash"},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"status":"beta","canonical_model_id":"xai/grok-4.7"},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005}},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":28,"cache_read":0.35},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.95,"output":4},"canonical_model_id":"moonshotai/kimi-k2.6"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"gpt-3.5-turbo-0125":{"id":"gpt-3.5-turbo-0125","name":"GPT-3.5 Turbo 0125","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2021-08","release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":4096},"status":"deprecated","cost":{"input":0.5,"output":1.5}},"claude-opus-4-1":{"id":"claude-opus-4-1","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"deepseek-v3.2-speciale":{"id":"deepseek-v3.2-speciale","name":"DeepSeek-V3.2-Speciale","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.58,"output":1.68}},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.58,"output":1.68}},"gpt-3.5-turbo-1106":{"id":"gpt-3.5-turbo-1106","name":"GPT-3.5 Turbo 1106","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2021-08","release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":4096},"status":"deprecated","cost":{"input":1,"output":2}},"gpt-5-codex":{"id":"gpt-5-codex","name":"GPT-5-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"o3":{"id":"o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek-V4-Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":1.74,"output":3.48},"canonical_model_id":"deepseek/deepseek-v4-pro"},"grok-4-1-fast-reasoning":{"id":"grok-4-1-fast-reasoning","name":"Grok 4.1 Fast (Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-06-27","last_updated":"2025-06-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"beta","provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"deepseek-r1":{"id":"deepseek-r1","name":"DeepSeek-R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-01-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"status":"deprecated","cost":{"input":1.35,"output":5.4}},"gpt-image-2":{"id":"gpt-image-2","type":"image","name":"GPT-Image-2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":8,"output":40,"cache_read":0.8,"cache_write":10},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"text-embedding-3-small":{"id":"text-embedding-3-small","name":"text-embedding-3-small","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536},"cost":{"input":0.02,"output":0}},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-4-turbo-vision":{"id":"gpt-4-turbo-vision","name":"GPT-4 Turbo Vision","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":10,"output":30}},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.71,"output":0.71}},"gpt-image-1":{"id":"gpt-image-1","type":"image","name":"GPT-Image-1","description":"OpenAI image model for production generation, edits, and brand-safe visual workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-24","last_updated":"2025-04-24","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":40,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-1"},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"gpt-image-2.5-sunburst":{"id":"gpt-image-2.5-sunburst","type":"image","name":"GPT Image 2.5 Sunburst","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-sunburst"},"cohere-embed-v3-multilingual":{"id":"cohere-embed-v3-multilingual","name":"Embed v3 Multilingual","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"cohere-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-11-07","last_updated":"2023-11-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":1024},"cost":{"input":0.1,"output":0}},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-07-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}}},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"gpt-4o-mini":{"id":"gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"gpt-image-1.5":{"id":"gpt-image-1.5","type":"image","name":"GPT-Image-1.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":32,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-1.5"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.6,"output":3},"canonical_model_id":"moonshotai/kimi-k2.5"},"gpt-image-2.5-flare":{"id":"gpt-image-2.5-flare","type":"image","name":"GPT Image 2.5 Flare","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-flare"},"cohere-embed-v-4-0":{"id":"cohere-embed-v-4-0","name":"Embed v4","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"cohere-embed","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":1536},"cost":{"input":0.12,"output":0}},"gpt-4-turbo":{"id":"gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"ministral-3b":{"id":"ministral-3b","name":"Ministral 3B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-03","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.04,"output":0.04}},"phi-4-mini":{"id":"phi-4-mini","name":"Phi-4-mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"phi","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.075,"output":0.3}},"grok-4-20-non-reasoning":{"id":"grok-4-20-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-04-08","last_updated":"2026-04-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":8192},"status":"beta","cost":{"input":2,"output":6}},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-11-24","last_updated":"2025-08-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek-V4.1-Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"status":"beta","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"codestral-2501":{"id":"codestral-2501","name":"Codestral 25.01","description":"Mistral coding model for code completion, generation, and developer workflows","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-03","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.3,"output":0.9}},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"o1":{"id":"o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":15,"output":60,"cache_read":7.5}},"gpt-5.1-codex-max":{"id":"gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex-max"},"mistral-small-2503":{"id":"mistral-small-2503","name":"Mistral Small 3.1","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2025-03-01","last_updated":"2025-03-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":0.1,"output":0.3}},"mistral-medium-2505":{"id":"mistral-medium-2505","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.4,"output":2}},"gpt-5.1-codex-mini":{"id":"gpt-5.1-codex-mini","name":"GPT-5.1 Codex Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-14","last_updated":"2025-11-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"grok-4-1-fast-non-reasoning":{"id":"grok-4-1-fast-non-reasoning","name":"Grok 4.1 Fast (Non-Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-06-27","last_updated":"2025-06-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}},"phi-4-multimodal":{"id":"phi-4-multimodal","name":"Phi-4-multimodal","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"phi","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.08,"output":0.32,"input_audio":4}},"model-router":{"id":"model-router","name":"Model Router","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"model-router","attachment":true,"reasoning":false,"tool_call":true,"release_date":"2025-05-19","last_updated":"2025-11-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":16384},"provider":{"npm":"@ai-sdk/azure","shape":"completions"},"cost":{"input":0.14,"output":0}},"claude-mythos-5":{"id":"claude-mythos-5","name":"Claude Mythos 5","description":"Restricted Claude model for advanced cybersecurity and biology research workflows","family":"claude-mythos","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"beta","provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-mythos-5"},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-14","last_updated":"2025-11-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"phi-4":{"id":"phi-4","name":"Phi-4","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"phi","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.125,"output":0.5}},"cohere-embed-v3-english":{"id":"cohere-embed-v3-english","name":"Embed v3 English","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"cohere-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-11-07","last_updated":"2023-11-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":1024},"cost":{"input":0.1,"output":0}},"text-embedding-ada-002":{"id":"text-embedding-ada-002","name":"text-embedding-ada-002","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2022-12-15","last_updated":"2022-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536},"cost":{"input":0.1,"output":0}},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":1.5,"output":9,"cache_read":0.15},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"}}},"zeldoc":{"id":"zeldoc","env":["ZELDOC_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.zeldoc.ai/v1","name":"Zeldoc","doc":"https://docs.zeldoc.ai","models":{"zdev":{"id":"zdev","name":"ZDev","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0}}}},"bothub":{"id":"bothub","env":["BOTHUB_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://openai.bothub.ru/v1","name":"Bothub","doc":"https://bothub.ru/models","models":{"muse-spark-1.3-contributor":{"id":"muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.1,"output":0.2},"canonical_model_id":"meta/muse-spark-1.3"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.7671,"output":5.3011},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"gemma-4-31b-it:free":{"id":"gemma-4-31b-it:free","name":"Gemma 4 31B IT (free)","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-31b-it"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2677,"output":1.6064},"canonical_model_id":"openai/gpt-5.6-luna"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1566,"output":0.522},"canonical_model_id":"zhipuai/glm-5.3-flash"},"nemotron-3-ultra-550b-a55b:free":{"id":"nemotron-3-ultra-550b-a55b:free","name":"Nemotron 3 Ultra (free)","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.589,"output":1.7671},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4243,"output":4.4765},"canonical_model_id":"zhipuai/glm-5.3"}}},"vultr":{"id":"vultr","env":["VULTR_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.vultrinference.com/v1","name":"Vultr","doc":"https://api.vultrinference.com/","models":{"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.15,"output":1},"canonical_model_id":"alibaba/qwen3.8-27b"},"nemotron-3-nano-omni-30b-a3b-reasoning":{"id":"nemotron-3-nano-omni-30b-a3b-reasoning","name":"Nemotron 3 Nano Omni 30B A3B Reasoning","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.1,"output":0.25},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"laguna-s-2.1":{"id":"laguna-s-2.1","name":"Laguna S 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"status":"deprecated","cost":{"input":0.09,"output":0.18},"canonical_model_id":"poolside/laguna-s-2.1"},"muse-glimmer-30b":{"id":"muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.25,"output":1},"canonical_model_id":"meta/muse-glimmer-30b"},"qwen3.8-flash-next":{"id":"qwen3.8-flash-next","name":"Qwen3.8 Flash Next","description":"Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.1,"output":0.2},"canonical_model_id":"alibaba/qwen3.8-flash-next"},"glm-5.x-menthol":{"id":"glm-5.x-menthol","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.4,"output":1.75},"canonical_model_id":"zhipuai/glm-5"},"mimo-v2.6-pro-rl":{"id":"mimo-v2.6-pro-rl","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.4,"output":0.8},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.35},"canonical_model_id":"zhipuai/glm-5.3-flash"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":512000},"cost":{"input":0.2,"output":0.9},"canonical_model_id":"minimax/MiniMax-M3"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.75,"output":3},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"status":"deprecated","cost":{"input":0.1,"output":0.25},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"nemotron-3.5-content-safety":{"id":"nemotron-3.5-content-safety","name":"Nemotron 3.5 Content Safety","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.05,"output":0.15},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"mimo-v2.6-flash-rl":{"id":"mimo-v2.6-flash-rl","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.25},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.75,"output":3},"canonical_model_id":"zhipuai/glm-5.3"}}},"pareto":{"id":"pareto","env":["PARETO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.paretoinference.com/v1","name":"Pareto Inference","doc":"https://docs.paretoinference.com/","models":{"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.06,"output":0.2,"cache_read":0.012},"canonical_model_id":"zhipuai/glm-5.3-flash"}}},"novita-ai":{"id":"novita-ai","env":["NOVITA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.novita.ai/openai","name":"Novita AI","doc":"https://novita.ai/docs/guides/introduction","models":{"apodex/apodex-1.1-mini":{"id":"apodex/apodex-1.1-mini","name":"Apodex 1.1 Mini","description":"Open-weight 35B MoE model for long-horizon research, forecasting, and tool-assisted work","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-24","last_updated":"2026-08-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"apodex/apodex-1.1-mini"},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.6,"output":2.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.8,"output":3.4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 Instruct 0905","description":"September 2025 Kimi K2 instruction update with a 256K context and improved agentic coding","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.6,"output":2.5},"canonical_model_id":"moonshotai/kimi-k2-0905"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2-instruct":{"id":"moonshotai/kimi-k2-instruct","name":"Kimi K2 Instruct","description":"Non-thinking Kimi K2 MoE instruction model for coding and tool-using agents","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.57,"output":2.3},"canonical_model_id":"moonshotai/kimi-k2-instruct"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"gryphe/mythomax-l2-13b":{"id":"gryphe/mythomax-l2-13b","name":"MythoMax L2 13B","description":"Llama 2 merge combining MythoLogic and Huginn for creative writing and roleplay","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2023-08-10","last_updated":"2023-08-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":3200},"cost":{"input":0.09,"output":0.09},"canonical_model_id":"gryphe/mythomax-l2-13b"},"xiaomimimo/mimo-v2.5":{"id":"xiaomimimo/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.168,"output":0.336,"cache_read":0.0034},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomimimo/mimo-v2.5-pro":{"id":"xiaomimimo/mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.522,"output":1.044,"cache_read":0.0043},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomimimo/mimo-v2.6-pro":{"id":"xiaomimimo/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"xiaomimimo/mimo-v2.6-flash":{"id":"xiaomimimo/mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"inclusionai/ling-3.0-flash":{"id":"inclusionai/ling-3.0-flash","name":"Ling 3.0 Flash","description":"Hybrid-linear MoE reasoning model with 124B total and 5.1B active parameters for efficient agents","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012},"canonical_model_id":"inclusionai/ling-3.0-flash"},"inclusionai/ling-3.0-flash-vl":{"id":"inclusionai/ling-3.0-flash-vl","name":"Ling 3.0 Flash VL","description":"Native multimodal hybrid MoE model for image and video understanding, reasoning, and visual agents","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash-vl"},"inclusionai/ling-3.1-flash":{"id":"inclusionai/ling-3.1-flash","name":"Ling 3.1 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":0.9,"cache_read":0.06},"canonical_model_id":"inclusionai/ling-3.1-flash"},"inclusionai/ling-3.0-flash-sante":{"id":"inclusionai/ling-3.0-flash-sante","name":"Ling 3.0 Flash Sante","description":"Health and medicine specialization of Ling 3.0 Flash for medical reasoning and evidence-based retrieval","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash-sante"},"inclusionai/ling-3.0-flash-fin":{"id":"inclusionai/ling-3.0-flash-fin","name":"Ling 3.0 Flash Fin","description":"Finance-enhanced model for financial research, multi-step investment workflows, and long-horizon planning and execution","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-27","last_updated":"2026-09-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash-fin"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.05,"output":0.25},"canonical_model_id":"openai/gpt-oss-120b"},"deepseek/deepseek-r1-0528":{"id":"deepseek/deepseek-r1-0528","name":"DeepSeek R1 0528","description":"May 2025 DeepSeek R1 update with improved mathematics, coding, and reasoning","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.7,"output":2.5,"cache_read":0.35},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":1.6,"output":3.2,"cache_read":0.135},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-r1-0528-qwen3-8b":{"id":"deepseek/deepseek-r1-0528-qwen3-8b","name":"DeepSeek R1 0528 Qwen3 8B","description":"DeepSeek R1 0528 distilled into Qwen3 8B for mathematics, coding, and reasoning","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"release_date":"2025-05-29","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.06,"output":0.09},"canonical_model_id":"deepseek/deepseek-r1-0528-qwen3-8b"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.44,"output":1.32,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-r1-turbo":{"id":"deepseek/deepseek-r1-turbo","name":"DeepSeek R1 (Turbo)","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":16000},"cost":{"input":0.7,"output":2.5},"canonical_model_id":"deepseek/deepseek-r1"},"deepseek/deepseek_v3":{"id":"deepseek/deepseek_v3","name":"DeepSeek-V3","description":"Open DeepSeek MoE chat model for coding, math, and general reasoning","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":16000},"cost":{"input":0.89,"output":0.89},"canonical_model_id":"deepseek/deepseek-v3"},"deepseek/deepseek-ocr-2":{"id":"deepseek/deepseek-ocr-2","name":"DeepSeek OCR 2","description":"High-accuracy OCR model for extracting text from documents, screenshots, receipts, and natural scenes","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.03,"output":0.03},"canonical_model_id":"deepseek/deepseek-ocr-2"},"paddlepaddle/paddleocr-vl":{"id":"paddlepaddle/paddleocr-vl","name":"PaddleOCR VL","description":"Compact vision-language model for multilingual document parsing, OCR, tables, formulas, and charts","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-10-16","last_updated":"2025-10-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":0.02,"output":0.02},"canonical_model_id":"paddlepaddle/paddleocr-vl"},"nousresearch/hermes-2-pro-llama-3-8b":{"id":"nousresearch/hermes-2-pro-llama-3-8b","name":"Hermes 2 Pro Llama 3 8B","description":"Llama 3 fine-tune optimized for function calling, structured JSON responses, and dialogue","family":"hermes","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2024-04-30","last_updated":"2024-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.14,"output":0.14},"canonical_model_id":"nousresearch/hermes-2-pro-llama-3-8b"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":262144},"cost":{"input":0.14,"output":0.58,"cache_read":0.035},"canonical_model_id":"tencent/hy3"},"mindai/macaron-v1-venti":{"id":"mindai/macaron-v1-venti","name":"Macaron V1 Venti","description":"GLM-5.2-based Mixture-of-LoRA model with specialists for chat, personal agents, coding, and generative UI","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.825,"output":2.475,"cache_read":0.165},"canonical_model_id":"mindlab/macaron-v1-venti"},"mindai/macaron-v1-tall":{"id":"mindai/macaron-v1-tall","name":"Macaron V1 Tall","description":"Qwen3.6-based Mixture-of-LoRA model with specialists for chat, personal agents, coding, and generative UI","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2475,"output":1.43,"cache_read":0.044},"canonical_model_id":"mindlab/macaron-v1-tall"},"baidu/cobuddy":{"id":"baidu/cobuddy","name":"CoBuddy","description":"Baidu Qianfan model optimized for code generation and tool-using coding agents","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-05-06","last_updated":"2026-05-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":0.28,"output":1.13,"cache_read":0.07},"canonical_model_id":"baidu/cobuddy"},"baidu/ernie-4.5-21B-a3b":{"id":"baidu/ernie-4.5-21B-a3b","name":"ERNIE 4.5 21B A3B","description":"Open ERNIE text MoE with 21B total and 3B active parameters","family":"ernie","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-06-30","last_updated":"2025-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":120000,"output":8000},"cost":{"input":0.07,"output":0.28},"canonical_model_id":"baidu/ernie-4.5-21b-a3b"},"meta-llama/llama-4-scout-17b-16e-instruct":{"id":"meta-llama/llama-4-scout-17b-16e-instruct","name":"Llama 4 Scout 17B Instruct","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.18,"output":0.59},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"meta-llama/llama-3.1-8b-instruct":{"id":"meta-llama/llama-3.1-8b-instruct","name":"Llama-3.1-8B-Instruct","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":0.02,"output":0.05},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta-llama/llama-4-maverick-17b-128e-instruct-fp8":{"id":"meta-llama/llama-4-maverick-17b-128e-instruct-fp8","name":"Llama 4 Maverick 17B Instruct","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":8192},"cost":{"input":0.27,"output":0.85},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"meta-llama/llama-3.2-1b-instruct":{"id":"meta-llama/llama-3.2-1b-instruct","name":"Llama 3.2 1B Instruct","description":"Compact instruction-tuned Llama model for multilingual dialogue and on-device assistants","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":32000},"cost":{"input":0.02,"output":0.02},"canonical_model_id":"meta/llama-3.2-1b-instruct"},"meta-llama/llama-3.3-70b-instruct":{"id":"meta-llama/llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":12288,"output":12288},"cost":{"input":0.135,"output":0.4},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"zai-org/autoglm-phone-9b-multilingual":{"id":"zai-org/autoglm-phone-9b-multilingual","name":"AutoGLM Phone 9B Multilingual","description":"Multilingual visual model for operating smartphone interfaces and mobile agents","family":"glm","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-10","last_updated":"2025-12-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.035,"output":0.138},"canonical_model_id":"zhipuai/autoglm-phone-9b-multilingual"},"zai-org/glm-4.5-air":{"id":"zai-org/glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.13,"output":0.85,"cache_read":0.025},"canonical_model_id":"zhipuai/glm-4.5-air"},"zai-org/glm-4.5v":{"id":"zai-org/glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":0.6,"output":1.8,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5v"},"zai-org/glm-5.1":{"id":"zai-org/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1.38,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/glm-4.7":{"id":"zai-org/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"zai-org/glm-5.3-flash":{"id":"zai-org/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai-org/glm-5.2":{"id":"zai-org/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/glm-4.7-flash":{"id":"zai-org/glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.07,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flash"},"zai-org/glm-5":{"id":"zai-org/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"zai-org/glm-4.6":{"id":"zai-org/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.55,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"zai-org/glm-5.3":{"id":"zai-org/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/glm-4.6v":{"id":"zai-org/glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.055},"canonical_model_id":"zhipuai/glm-4.6v"},"stepfun/step-3.7-flash":{"id":"stepfun/step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":256000,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"stepfun/step-5-preview":{"id":"stepfun/step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":64000},"cost":{"input":1,"output":2.7,"cache_read":0.1},"canonical_model_id":"stepfun/step-5-preview"},"baai/bge-m3":{"id":"baai/bge-m3","type":"embedding","name":"BGE M3","description":"Multilingual embedding model combining dense, sparse, and multi-vector retrieval","family":"bge","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-01-30","last_updated":"2024-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":0},"cost":{"input":0.01,"output":0},"canonical_model_id":"baai/bge-m3"},"baai/bge-reranker-v2-m3":{"id":"baai/bge-reranker-v2-m3","type":"reranking","name":"BGE Reranker V2 M3","description":"Lightweight multilingual cross-encoder that scores query-document relevance","family":"bge","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-03-15","last_updated":"2024-03-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"output":0},"cost":{"input":0.01,"output":0},"canonical_model_id":"baai/bge-reranker-v2-m3"},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"tiers":[{"input":0.6,"output":2.4,"cache_read":0.12,"tier":{"type":"context","size":524288}}],"context_over_200k":{"input":0.6,"output":2.4,"cache_read":0.12}},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"microsoft/wizardlm-2-8x22b":{"id":"microsoft/wizardlm-2-8x22b","name":"WizardLM 2 8x22B","description":"Microsoft WizardLM instruction-tuned Mixtral MoE for multilingual dialogue and complex tasks","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2024-04-15","last_updated":"2024-04-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65535,"output":8000},"cost":{"input":0.62,"output":0.62},"canonical_model_id":"microsoft/wizardlm-2-8x22b"},"nvidia/nemotron-3-nano-30b-a3b":{"id":"nvidia/nemotron-3-nano-30b-a3b","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.05,"output":0.2},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"baichuan/baichuan-m2-32b":{"id":"baichuan/baichuan-m2-32b","name":"Baichuan M2 32B","description":"Open medical model trained for clinical reasoning and patient interaction with on, off, and auto thinking modes","family":"baichuan","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"temperature":true,"release_date":"2025-08-13","last_updated":"2025-08-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.07,"output":0.07},"canonical_model_id":"baichuan/baichuan-m2-32b"},"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3-vl-30b-a3b-instruct":{"id":"qwen/qwen3-vl-30b-a3b-instruct","name":"Qwen3 VL 30B A3B Instruct","description":"Open vision-language MoE model for image and video understanding, documents, and visual agents","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.2,"output":0.7},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.25,"output":3.75,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3-omni-30b-a3b-thinking":{"id":"qwen/qwen3-omni-30b-a3b-thinking","name":"Qwen3 Omni 30B A3B Thinking","description":"Multimodal reasoning model accepting text, images, video, and audio with text output","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-22","last_updated":"2025-09-22","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":0.25,"output":0.97,"input_audio":2.2},"canonical_model_id":"alibaba/qwen3-omni-30b-a3b-thinking"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.42,"output":3,"cache_read":0.085},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.6-35b-a3b":{"id":"qwen/qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.248,"output":1.485},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":3.2},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3.8-2.4t-a95b":{"id":"qwen/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"qwen/qwen3-embedding-8b":{"id":"qwen/qwen3-embedding-8b","type":"embedding","name":"Qwen3 Embedding 8B","description":"Multilingual text embedding model for retrieval, code search, classification, and clustering","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":0},"cost":{"input":0.07,"output":0},"canonical_model_id":"alibaba/qwen3-embedding-8b"},"qwen/qwen3.8-max":{"id":"qwen/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":2.4},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen-mt-plus":{"id":"qwen/qwen-mt-plus","name":"Qwen MT Plus","description":"Flagship Qwen translation model with terminology control and support for 92 languages","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-07-24","last_updated":"2025-07-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":8192},"cost":{"input":0.25,"output":0.75},"canonical_model_id":"alibaba/qwen-mt-plus"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3-reranker-8b":{"id":"qwen/qwen3-reranker-8b","type":"reranking","name":"Qwen3 Reranker 8B","description":"Multilingual query-document reranking model supporting more than 100 languages","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":0},"cost":{"input":0.05,"output":0},"canonical_model_id":"alibaba/qwen3-reranker-8b"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.6-27b"},"minimaxai/minimax-m1-80k":{"id":"minimaxai/minimax-m1-80k","name":"MiniMax M1 80K","description":"Open hybrid-attention MoE reasoning model with a million-token context and an 80K thinking budget","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":40000},"cost":{"input":0.55,"output":2.2},"canonical_model_id":"minimax/minimax-m1-80k"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":98304,"output":16384},"cost":{"input":0.119,"output":0.2},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.13,"output":0.4},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.05,"output":0.1},"canonical_model_id":"google/gemma-3-12b-it"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"}}},"unorouter":{"id":"unorouter","env":["UNOROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.unorouter.com/v1","name":"UnoRouter","doc":"https://unorouter.com/models","models":{"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"deprecated","cost":{"input":0.425,"output":2.125},"canonical_model_id":"anthropic/claude-opus-4-8"},"glm-4.5-flash:free":{"id":"glm-4.5-flash:free","name":"GLM-4.5-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":98304},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.5-flash"},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.819,"output":3.276},"canonical_model_id":"minimax/MiniMax-M2.7"},"deepseek-v4-pro:free":{"id":"deepseek-v4-pro:free","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0.0625,"output":0.125},"canonical_model_id":"deepseek/deepseek-v4-flash"},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":1.05,"output":8.4},"canonical_model_id":"openai/gpt-5.2"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.2675,"output":5.3368},"canonical_model_id":"moonshotai/kimi-k2.6"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1875,"output":1.125},"canonical_model_id":"openai/gpt-5.5"},"glm-5.2:free":{"id":"glm-5.2:free","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.2"},"step-3.7-flash:free":{"id":"step-3.7-flash:free","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"stepfun/step-3.7-flash"},"deepseek-v4-flash:free":{"id":"deepseek-v4-flash:free","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"status":"deprecated","cost":{"input":1.8,"output":10.8},"canonical_model_id":"openai/gpt-5.4"},"gemma-4-31b-it:free":{"id":"gemma-4-31b-it:free","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-31b-it"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.8999,"output":1.7999},"canonical_model_id":"deepseek/deepseek-v4-pro"},"gpt-5.4:free":{"id":"gpt-5.4:free","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.4"},"gpt-5.5:free":{"id":"gpt-5.5:free","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.5"},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.2,"output":6},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":1.6001,"output":5.0288},"canonical_model_id":"zhipuai/glm-5.2"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1857,"output":1.1142},"canonical_model_id":"google/gemini-3.5-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":1.44,"output":7.2},"canonical_model_id":"anthropic/claude-sonnet-5"},"nemotron-3-ultra-550b-a55b:free":{"id":"nemotron-3-ultra-550b-a55b:free","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"minimax-m2.7:free":{"id":"minimax-m2.7:free","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"minimax/MiniMax-M2.7"},"qwen3.5-397b-a17b:free":{"id":"qwen3.5-397b-a17b:free","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"}}},"moonshotai":{"id":"moonshotai","env":["MOONSHOT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.moonshot.ai/v1","name":"Moonshot AI","doc":"https://platform.moonshot.ai/docs/api/chat","models":{"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"kimi-k2.7-code-highspeed":{"id":"kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code HighSpeed","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3},"canonical_model_id":"moonshotai/kimi-k3"}}},"modal":{"id":"modal","env":["MODAL_PROXY_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.us-west.modal.direct/v1","name":"Modal","doc":"https://modal.com/docs/guide/endpoints","models":{"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"reasoning":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"thinkingmachines/Inkling-NVFP4":{"id":"thinkingmachines/Inkling-NVFP4","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":1.2,"output":5,"cache_read":0.27},"canonical_model_id":"thinkingmachines/inkling"},"meta-models/Muse-Spark-Open-816B-A42B":{"id":"meta-models/Muse-Spark-Open-816B-A42B","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.45,"output":1.5,"cache_read":0.09},"canonical_model_id":"zhipuai/glm-5.3-flash"},"nvidia/GLM-5.3-Flash-NVFP4":{"id":"nvidia/GLM-5.3-Flash-NVFP4","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"deepseek-ai/DeepSeek-V4.1-Flash":{"id":"deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"Qwen/Qwen3.8-2.4T-A95B":{"id":"Qwen/Qwen3.8-2.4T-A95B","name":"Qwen3.8-Max","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1010000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"}}},"crossmodel":{"id":"crossmodel","env":["CROSSMODEL_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.crossmodel.ai/v1","name":"CrossModel","doc":"https://www.crossmodel.ai/docs","models":{"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":1,"output":3.8,"cache_read":0.2,"cache_write":1,"tiers":[{"input":1.2,"output":4.4,"cache_read":0.3,"cache_write":1.2,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.47,"output":2.16,"cache_read":0.1,"cache_write":0.47,"tiers":[{"input":0.62,"output":2.47,"cache_read":0.13,"cache_write":0.62,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":0.9,"output":3.7,"cache_read":0.18,"cache_write":0.9,"tiers":[{"input":1.1,"output":4.3,"cache_read":0.27,"cache_write":1.1,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5-turbo"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.15,"output":0.5,"cache_read":0.03,"cache_write":0.15},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.2,"output":4.4,"cache_read":0.3,"cache_write":1.2},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.6,"output":3,"cache_read":0.16,"cache_write":0.6,"tiers":[{"input":0.8,"output":3.4,"cache_read":0.2,"cache_write":0.8,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.2,"output":4.4,"cache_read":0.3,"cache_write":1.2},"canonical_model_id":"zhipuai/glm-5.3"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":5,"tiers":[{"input":10,"output":45,"cache_read":1,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1,"cache_write":10}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":2.5,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5,"cache_write":5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075,"cache_write":0.15},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075,"cache_write":0.75},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02,"cache_write":0.2},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.006,"cache_write":0.3},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.006,"cache_write":0.3},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.35,"output":4.05,"cache_read":0.045,"cache_write":1.35},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006,"cache_write":0.3},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"anthropic/claude-sonnet-5-5":{"id":"anthropic/claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4-8":{"id":"anthropic/claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-haiku-4-5":{"id":"anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-haiku-5-5":{"id":"anthropic/claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-5-5":{"id":"anthropic/claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-fable-5-1":{"id":"anthropic/claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"gemini/gemini-3.8-flash":{"id":"gemini/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.75},"canonical_model_id":"google/gemini-3.8-flash"},"gemini/gemini-2.5-flash":{"id":"gemini/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.3},"canonical_model_id":"google/gemini-2.5-flash"},"gemini/gemini-4-argon":{"id":"gemini/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2},"canonical_model_id":"google/gemini-4-argon"},"gemini/gemini-3.5-flash-lite":{"id":"gemini/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.3},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gemini/gemini-3.1-pro-preview":{"id":"gemini/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"gemini/gemini-3.6-flash":{"id":"gemini/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"gemini/gemini-3.7-flash":{"id":"gemini/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.75},"canonical_model_id":"google/gemini-3.7-flash"},"gemini/gemini-3.5-flash":{"id":"gemini/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"gemini/gemini-3-flash-preview":{"id":"gemini/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.5},"canonical_model_id":"google/gemini-3-flash-preview"},"gemini/gemini-2.5-flash-lite":{"id":"gemini/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":0.1},"canonical_model_id":"google/gemini-2.5-flash-lite"},"gemini/gemini-2.5-pro":{"id":"gemini/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"cache_write":1.25,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"cache_write":2.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":2.5}},"canonical_model_id":"google/gemini-2.5-pro"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0.96,"output":2.88,"cache_read":0.048,"cache_write":0.96},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":131072},"cost":{"input":0.16,"output":0.64,"cache_read":0.04,"cache_write":0.16},"canonical_model_id":"tencent/hy3"},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.47,"output":0.94,"cache_read":0.005,"cache_write":0.47},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"xiaomi/mimo-v2.6-flash":{"id":"xiaomi/mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.16,"output":0.32,"cache_read":0.004,"cache_write":0.16},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"moonshot/kimi-k2.6":{"id":"moonshot/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":1,"output":4.16,"cache_read":0.18,"cache_write":1},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshot/kimi-k2.7-code":{"id":"moonshot/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":1,"output":4.16,"cache_read":0.18,"cache_write":1},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshot/kimi-k3":{"id":"moonshot/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3},"canonical_model_id":"moonshotai/kimi-k3"},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.33,"output":1.32,"cache_read":0.066,"cache_write":0.42},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":512000},"cost":{"input":0.33,"output":1.32,"cache_read":0.066,"cache_write":0.33,"tiers":[{"input":0.66,"output":2.63,"cache_read":0.132,"cache_write":0.66,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":0.66,"output":2.63,"cache_read":0.132,"cache_write":0.66}},"canonical_model_id":"minimax/MiniMax-M3"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.88,"output":5.63,"cache_read":0.375,"cache_write":2.35},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.32,"output":1.25,"cache_read":0.032,"cache_write":0.4,"tiers":[{"input":0.96,"output":3.75,"cache_read":0.096,"cache_write":1.2,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.96,"output":3.75,"cache_read":0.096,"cache_write":1.2}},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3.7-flash":{"id":"qwen/qwen3.7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":65536},"cost":{"input":0.04,"output":0.13,"cache_read":0.01,"cache_write":0.04,"tiers":[{"input":0.1,"output":0.37,"cache_read":0.02,"cache_write":0.12,"tier":{"type":"context","size":32000}},{"input":0.19,"output":0.74,"cache_read":0.04,"cache_write":0.24,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen/qwen3.8-max":{"id":"qwen/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.88,"output":5.63,"cache_read":0.23,"cache_write":2.35},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.19,"output":1.13,"cache_read":0.019,"cache_write":0.24,"tiers":[{"input":0.75,"output":4.5,"cache_read":0.075,"cache_write":0.94,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.75,"output":4.5,"cache_read":0.075,"cache_write":0.94}},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3.8-omni-flash":{"id":"qwen/qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.13,"output":0.43,"cache_read":0.016,"cache_write":0.13},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.13,"output":0.43,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.32,"output":1.88,"cache_read":0.032,"cache_write":0.4,"tiers":[{"input":1.25,"output":7.5,"cache_read":0.124,"cache_write":1.57,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.25,"output":7.5,"cache_read":0.124,"cache_write":1.57}},"canonical_model_id":"alibaba/qwen3.6-plus"},"x-ai/grok-4.5":{"id":"x-ai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3,"cache_write":2,"tiers":[{"input":4,"output":12,"cache_read":0.6,"cache_write":4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6,"cache_write":4}},"canonical_model_id":"xai/grok-4.5"},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"cache_write":2,"tiers":[{"input":4,"output":12,"cache_read":1,"cache_write":4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1,"cache_write":4}},"canonical_model_id":"xai/grok-4.6"},"x-ai/grok-4.7":{"id":"x-ai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"cache_write":2,"tiers":[{"input":4,"output":12,"cache_read":1,"cache_write":4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1,"cache_write":4}},"canonical_model_id":"xai/grok-4.7"},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"cache_write":1.25,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"cache_write":2.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4,"cache_write":2.5}},"canonical_model_id":"xai/grok-4.3"},"x-ai/grok-build-0.1":{"id":"x-ai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"cache_write":1,"tiers":[{"input":2,"output":4,"cache_read":0.4,"cache_write":2,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4,"cache_write":2}},"canonical_model_id":"xai/grok-build-0.1"}}},"scx-ai":{"id":"scx-ai","env":["SCX_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.scx.ai/v1","name":"SCX.ai","doc":"https://platform.scx.ai/docs","models":{"GLM-5.2":{"id":"GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.55,"output":1.784,"cache_read":0.111},"canonical_model_id":"zhipuai/glm-5.2"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.17,"output":0.55},"canonical_model_id":"openai/gpt-oss-120b"},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":196608},"cost":{"input":0.48,"output":1.79,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M2.7"},"Qwen3.8-Max":{"id":"Qwen3.8-Max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":983616,"output":131072},"cost":{"input":1.815,"output":5.4461,"cache_read":0.17,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"}}},"greenpt":{"id":"greenpt","env":["GREENPT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.greenpt.ai/v1","name":"GreenPT","doc":"https://docs.greenpt.ai","models":{"green-r":{"id":"green-r","name":"Green R","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.399,"output":1.083},"canonical_model_id":"openai/gpt-oss-120b"},"green-s":{"id":"green-s","type":"transcription","name":"Green S","description":"GreenPT speech-to-text model for pre-recorded and live transcription","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-01","last_updated":"2025-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":8192}},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.342,"output":2.052},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"green-l":{"id":"green-l","name":"Green L","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.285,"output":0.912},"canonical_model_id":"mistral/mistral-small-2506"},"glm-5.2-ponytail-lite":{"id":"glm-5.2-ponytail-lite","name":"GLM-5.2 Ponytail Lite","description":"glm-5.2 carrying a built-in ruleset that compresses generated code, preferring platform features over custom code. The gentlest tier: it cuts filler only and keeps the explanation intact. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.738325,"output":5.1589,"cache_read":0.4486},"canonical_model_id":"zhipuai/glm-5.2"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.7524,"output":4.275,"cache_read":0.2508},"canonical_model_id":"moonshotai/kimi-k2.6"},"kimi-k2.6-fast":{"id":"kimi-k2.6-fast","name":"Kimi K2.6 Fast","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":1.655,"output":8.778},"canonical_model_id":"moonshotai/kimi-k2.6"},"gemma-3-27b-it":{"id":"gemma-3-27b-it","name":"Gemma 3 27B","description":"Google Gemma 3 multimodal model for chat, reasoning, and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":40000,"output":8192},"status":"deprecated","cost":{"input":0.342,"output":0.684},"canonical_model_id":"google/gemma-3-27b-it"},"qwen3-235b-a22b-instruct-2507":{"id":"qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Qwen3 235B MoE instruct model for long-context multilingual chat and reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":1.026,"output":3.078},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"status":"deprecated","cost":{"input":1.756,"output":5.518},"canonical_model_id":"zhipuai/glm-5.1"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.228,"output":0.798},"canonical_model_id":"openai/gpt-oss-120b"},"green-s-pro":{"id":"green-s-pro","type":"transcription","name":"Green S Pro","description":"GreenPT advanced speech-to-text model with multilingual transcription support","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-02","last_updated":"2025-02","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":8192}},"voxtral-small-24b-2507":{"id":"voxtral-small-24b-2507","name":"Voxtral Small 24B","description":"Mistral Voxtral audio-understanding model for speech and transcription tasks","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"status":"deprecated","cost":{"input":0.228,"output":0.513},"canonical_model_id":"mistral/voxtral-small-24b-2507"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.863555,"output":4.317775,"cache_read":0.222057},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"gemma4":{"id":"gemma4","name":"gemma4","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.57,"output":1.71},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"glm-5.2-ponytail":{"id":"glm-5.2-ponytail","name":"GLM-5.2 Ponytail","description":"glm-5.2 carrying a built-in ruleset that compresses generated code, preferring platform features over custom code. The middle tier, and the ruleset as its authors wrote it. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.738325,"output":5.1589,"cache_read":0.4486},"canonical_model_id":"zhipuai/glm-5.2"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3.762,"output":18.81,"cache_read":0.9405},"canonical_model_id":"moonshotai/kimi-k3"},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.798,"output":4.959},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"mistral-medium-3.5-128b":{"id":"mistral-medium-3.5-128b","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":2.052,"output":10.26},"canonical_model_id":"mistral/mistral-medium-2604"},"qwen3-coder-30b-a3b-instruct":{"id":"qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"status":"deprecated","cost":{"input":0.285,"output":1.083},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"mistral-small-3.2-24b-instruct-2506":{"id":"mistral-small-3.2-24b-instruct-2506","name":"Mistral Small 3.2","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.228,"output":0.456},"canonical_model_id":"mistral/mistral-small-2506"},"devstral-2-123b-instruct-2512":{"id":"devstral-2-123b-instruct-2512","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":16384},"status":"deprecated","cost":{"input":0.57,"output":2.736},"canonical_model_id":"mistral/devstral-2512"},"green-r-raw":{"id":"green-r-raw","name":"Green R Raw","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.399,"output":1.083},"canonical_model_id":"openai/gpt-oss-120b"},"glm-5.2-caveman":{"id":"glm-5.2-caveman","name":"GLM-5.2 Caveman","description":"glm-5.2 carrying a built-in ruleset that compresses prose, keeping code and technical detail verbatim. The middle tier, and the ruleset as its authors wrote it. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.738325,"output":5.1589,"cache_read":0.4486},"canonical_model_id":"zhipuai/glm-5.2"},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":100000,"output":16384},"cost":{"input":1.254,"output":1.254},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"glm-5.2-caveman-ultra":{"id":"glm-5.2-caveman-ultra","name":"GLM-5.2 Caveman Ultra","description":"glm-5.2 carrying a built-in ruleset that compresses prose, keeping code and technical detail verbatim. The most aggressive tier, close to answer-only. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.738325,"output":5.1589,"cache_read":0.4486},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5.2-honey":{"id":"glm-5.2-honey","name":"GLM-5.2 Honey","description":"glm-5.2 carrying a built-in ruleset that compresses both generated code and prose. The middle tier, and the ruleset as its authors wrote it. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.738325,"output":5.1589,"cache_read":0.4486},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5.2-honey-ultra":{"id":"glm-5.2-honey-ultra","name":"GLM-5.2 Honey Ultra","description":"glm-5.2 carrying a built-in ruleset that compresses both generated code and prose. The most aggressive tier, close to answer-only. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.738325,"output":5.1589,"cache_read":0.4486},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.127754,"output":0.511016,"cache_read":0.0255508},"canonical_model_id":"zhipuai/glm-5.3-flash"},"green-l-raw":{"id":"green-l-raw","name":"Green L Raw","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.285,"output":0.912},"canonical_model_id":"mistral/mistral-small-2506"},"glm-5.2-honey-lite":{"id":"glm-5.2-honey-lite","name":"GLM-5.2 Honey Lite","description":"glm-5.2 carrying a built-in ruleset that compresses both generated code and prose. The gentlest tier: it cuts filler only and keeps the explanation intact. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.738325,"output":5.1589,"cache_read":0.4486},"canonical_model_id":"zhipuai/glm-5.2"},"holo2-30b-a3b":{"id":"holo2-30b-a3b","name":"Holo2 30B A3B","description":"H Company Holo2 vision model for GUI navigation and computer-use agents","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-11","last_updated":"2025-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":22016,"output":16384},"status":"deprecated","cost":{"input":0.399,"output":0.969}},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.738325,"output":5.1589,"cache_read":0.4486},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.255552,"output":1.27776,"cache_read":0.0127776},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"glm-5.2-caveman-lite":{"id":"glm-5.2-caveman-lite","name":"GLM-5.2 Caveman Lite","description":"glm-5.2 carrying a built-in ruleset that compresses prose, keeping code and technical detail verbatim. The gentlest tier: it cuts filler only and keeps the explanation intact. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.738325,"output":5.1589,"cache_read":0.4486},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5.2-ponytail-ultra":{"id":"glm-5.2-ponytail-ultra","name":"GLM-5.2 Ponytail Ultra","description":"glm-5.2 carrying a built-in ruleset that compresses generated code, preferring platform features over custom code. The most aggressive tier, close to answer-only. Same upstream model and price per token as glm-5.2, with fewer output tokens.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.738325,"output":5.1589,"cache_read":0.4486},"canonical_model_id":"zhipuai/glm-5.2"},"minimax-m2.5":{"id":"minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.49346,"output":2.4673,"cache_read":0.123365},"canonical_model_id":"minimax/MiniMax-M2.5"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.1596,"output":0.399,"cache_read":0.0456},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"pixtral-12b-2409":{"id":"pixtral-12b-2409","name":"Pixtral 12B","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2024-09-11","last_updated":"2024-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.285,"output":0.285},"canonical_model_id":"mistral/pixtral-12b"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.27754,"output":5.11016,"cache_read":0.319385},"canonical_model_id":"zhipuai/glm-5.3"}}},"gmicloud":{"id":"gmicloud","env":["GMICLOUD_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.gmi-serving.com/v1","name":"GMI Cloud","doc":"https://docs.gmicloud.ai/inference-engine/api-reference/llm-api-reference","models":{"moonshotai/kimi-k2.7-code-highspeed":{"id":"moonshotai/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.855,"output":3.6,"cache_read":0.144},"canonical_model_id":"moonshotai/kimi-k2.6"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.6,"output":2.4,"cache_read":0.12,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.24,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.24}},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMaxAI/MiniMax-M2.7":{"id":"MiniMaxAI/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":409600,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":4.5,"output":22.5,"cache_read":0.45},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":409600,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":127999}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":409600,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-8"},"zai-org/GLM-5.1-FP8":{"id":"zai-org/GLM-5.1-FP8","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.98,"output":3.08,"cache_read":0.182},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-5.2-FP8":{"id":"zai-org/GLM-5.2-FP8","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.979,"output":3.08,"cache_read":0.182},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5-FP8":{"id":"zai-org/GLM-5-FP8","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.6,"output":1.92,"cache_read":0.12},"canonical_model_id":"zhipuai/glm-5"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.392,"output":2.784,"cache_read":0.116},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/DeepSeek-V4-Flash":{"id":"deepseek-ai/DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048575,"output":384000},"cost":{"input":0.112,"output":0.224,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-flash"},"Qwen/Qwen3.7-Max":{"id":"Qwen/Qwen3.7-Max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.7-max"},"Qwen/Qwen3.8-Flash":{"id":"Qwen/Qwen3.8-Flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048575,"output":131072},"cost":{"input":0.16,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"Qwen/Qwen3.8-Max":{"id":"Qwen/Qwen3.8-Max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"}}},"togetherai":{"id":"togetherai","env":["TOGETHER_API_KEY"],"npm":"@ai-sdk/togetherai","name":"Together AI","doc":"https://docs.together.ai/docs/serverless-models","models":{"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":524288},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMaxAI/MiniMax-M2.7":{"id":"MiniMaxAI/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.5"},"thinkingmachines/Inkling":{"id":"thinkingmachines/Inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":131072},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"LiquidAI/LFM2-24B-A2B":{"id":"LiquidAI/LFM2-24B-A2B","name":"LFM2-24B-A2B","description":"Efficient open Liquid MoE instruct model for on-device chat and tool calling","family":"liquid","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.03,"output":0.12},"canonical_model_id":"liquid/lfm2-24b-a2b"},"meta-models/Muse-Spark-Open-816B-A42B":{"id":"meta-models/Muse-Spark-Open-816B-A42B","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"meta-llama/Llama-3.3-70B-Instruct-Turbo":{"id":"meta-llama/Llama-3.3-70B-Instruct-Turbo","name":"Llama 3.3 70B","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":1.04,"output":1.04},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"zai-org/GLM-5":{"id":"zai-org/GLM-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"status":"deprecated","cost":{"input":1,"output":3.2},"canonical_model_id":"zhipuai/glm-5"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048575,"output":164000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048575,"output":262144},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048575,"output":400000},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"reflection/Beam-1.0-501B-A23B":{"id":"reflection/Beam-1.0-501B-A23B","name":"Beam 1.0 501B A23B","description":"Sparse mixture-of-experts reasoning model with 23B active parameters for coding and agentic workflows","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.4,"output":1.2,"cache_read":0.04},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"essentialai/Rnj-1-Instruct":{"id":"essentialai/Rnj-1-Instruct","name":"Rnj-1 Instruct","description":"Open 8B Essential AI model for code, STEM reasoning, and agentic tool use","family":"rnj","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-12-04","last_updated":"2025-12-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.15,"output":0.15},"canonical_model_id":"essentialai/rnj-1-instruct"},"nvidia/NVIDIA-Nemotron-3.5-Super-VL":{"id":"nvidia/NVIDIA-Nemotron-3.5-Super-VL","name":"Nemotron 3.5 Super VL 120B A12B","description":"NVIDIA vision-language reasoning model for text, image, and video understanding and agentic tool use","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.15,"output":0.5,"cache_read":0.02},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/nemotron-3-ultra-550b-a55b":{"id":"nvidia/nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512300,"output":512300},"status":"deprecated","cost":{"input":0.6,"output":3.6,"cache_read":0.2},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-R1":{"id":"deepseek-ai/DeepSeek-R1","name":"DeepSeek-R1","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163839,"output":163839},"status":"deprecated","cost":{"input":3,"output":7},"canonical_model_id":"deepseek/deepseek-r1"},"deepseek-ai/DeepSeek-V3":{"id":"deepseek-ai/DeepSeek-V3","name":"DeepSeek-V3","description":"Open DeepSeek MoE chat model for coding, math, and general reasoning","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":1.25,"output":1.25},"canonical_model_id":"deepseek/deepseek-v3"},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek-ai/DeepSeek-V4.1-Flash":{"id":"deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-ai/DeepSeek-V3-1":{"id":"deepseek-ai/DeepSeek-V3-1","name":"DeepSeek V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":0.6,"output":1.7},"canonical_model_id":"deepseek/deepseek-v3.1"},"Qwen/Qwen3.7-Max":{"id":"Qwen/Qwen3.7-Max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.5},"canonical_model_id":"alibaba/qwen3.7-max"},"Qwen/Qwen3-235B-A22B-Instruct-2507-tput":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507-tput","name":"Qwen3 235B A22B Instruct 2507 FP8","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.2,"output":0.6},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"Qwen/Qwen3.5-397B-A17B":{"id":"Qwen/Qwen3.5-397B-A17B","name":"Qwen3.5 397B A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":130000},"status":"deprecated","cost":{"input":0.6,"output":3.6,"cache_read":0.35},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"Qwen/Qwen3-Coder-Next-FP8":{"id":"Qwen/Qwen3-Coder-Next-FP8","name":"Qwen3 Coder Next FP8","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.5,"output":1.2},"canonical_model_id":"alibaba/qwen3-coder-next"},"Qwen/Qwen3.5-9B":{"id":"Qwen/Qwen3.5-9B","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.17,"output":0.25},"canonical_model_id":"alibaba/qwen3.5-9b"},"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8":{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8","name":"Qwen3 Coder 480B A35B Instruct","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":2,"output":2},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"Qwen/Qwen3.6-Plus":{"id":"Qwen/Qwen3.6-Plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3},"canonical_model_id":"alibaba/qwen3.6-plus"},"Qwen/Qwen2.5-7B-Instruct-Turbo":{"id":"Qwen/Qwen2.5-7B-Instruct-Turbo","name":"Qwen 2.5 7B Instruct Turbo","description":"Small open Qwen2.5 instruction model for chat, structured output, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-09-19","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.3,"output":0.3},"canonical_model_id":"alibaba/qwen2.5-7b-instruct"}}},"pioneer":{"id":"pioneer","env":["PIONEER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.fastino.ai/v1","name":"Pioneer","doc":"https://agent.pioneer.ai/llms.txt","models":{"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"deprecated","cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65000},"status":"deprecated","cost":{"input":0.25,"output":1.5,"cache_read":0.03,"cache_write":0.25},"canonical_model_id":"google/gemini-3.1-flash-lite"},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":131072},"status":"deprecated","cost":{"input":1.25,"output":10,"cache_read":0.125,"cache_write":1.25},"canonical_model_id":"openai/gpt-5.1"},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":64000},"status":"deprecated","cost":{"input":1.25,"output":3.75,"cache_read":0.25,"cache_write":1.5625},"canonical_model_id":"alibaba/qwen3.7-max"},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":131072},"status":"deprecated","cost":{"input":2,"output":6,"cache_read":0.5,"cache_write":2},"canonical_model_id":"xai/grok-4.5"},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":0.4,"output":1.6,"cache_read":0.2,"cache_write":0.4},"canonical_model_id":"openai/gpt-4.1-mini"},"claude-3-7-sonnet-latest":{"id":"claude-3-7-sonnet-latest","name":"Claude Sonnet 3.7","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-10-31","release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-3-7-sonnet-20250219"},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","cost":{"input":0.32,"output":1.28,"cache_read":0.064,"cache_write":0.4},"canonical_model_id":"alibaba/qwen3.7-plus"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"deprecated","cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"status":"deprecated","cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":0.25,"output":2,"cache_read":0.025,"cache_write":0.25},"canonical_model_id":"openai/gpt-5-mini"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":1.75,"output":14,"cache_read":0.175,"cache_write":1.75},"canonical_model_id":"openai/gpt-5.3-codex"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":0.05,"output":0.4,"cache_read":0.005,"cache_write":0.05},"canonical_model_id":"openai/gpt-5-nano"},"gemini-3-flash":{"id":"gemini-3-flash","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65000},"status":"deprecated","cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.083333},"canonical_model_id":"google/gemini-3-flash-preview"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"status":"deprecated","cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":5},"canonical_model_id":"openai/gpt-5.5"},"claude-opus-4-1":{"id":"claude-opus-4-1","name":"Claude Opus 4.1 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"status":"deprecated","cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.4"},"qwen3.6-max-preview":{"id":"qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":240000,"output":64000},"status":"deprecated","cost":{"input":1.04,"output":6.24,"cache_read":0.208,"cache_write":1.3},"canonical_model_id":"alibaba/qwen3.6-max-preview"},"mistral-medium":{"id":"mistral-medium","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"status":"deprecated","cost":{"input":0.4,"output":2,"cache_read":0.4,"cache_write":0.4},"canonical_model_id":"mistral/mistral-medium-2505"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"deprecated","cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"status":"deprecated","cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"openai/gpt-5.6-sol"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65000},"status":"deprecated","cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.3},"canonical_model_id":"google/gemini-3.5-flash-lite"},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"status":"deprecated","cost":{"input":0.1875,"output":1.125,"cache_read":0.0375,"cache_write":0.234375},"canonical_model_id":"alibaba/qwen3.6-flash"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"status":"deprecated","cost":{"input":1,"output":6,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"openai/gpt-5.6-luna"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"deprecated","cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"claude-opus-5-fast":{"id":"claude-opus-5-fast","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"deprecated","cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":2.5,"output":10,"cache_read":1.25,"cache_write":2.5},"canonical_model_id":"openai/gpt-4o"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","cost":{"input":1.5,"output":7.5,"cache_read":0.15,"cache_write":1.5},"canonical_model_id":"google/gemini-3.6-flash"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"deprecated","cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":0.1,"output":0.4,"cache_read":0.05,"cache_write":0.1},"canonical_model_id":"openai/gpt-4.1-nano"},"gpt-4o-mini":{"id":"gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":0.15,"output":0.6,"cache_read":0.075,"cache_write":0.15},"canonical_model_id":"openai/gpt-4o-mini"},"magistral-medium":{"id":"magistral-medium","name":"Magistral Medium (latest)","description":"Mistral reasoning model for transparent analysis, math, and complex decisions","family":"magistral-medium","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-06","release_date":"2025-03-17","last_updated":"2025-03-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"status":"deprecated","cost":{"input":2,"output":5,"cache_read":2,"cache_write":2},"canonical_model_id":"mistral/magistral-medium-latest"},"devstral-small-2":{"id":"devstral-small-2","name":"Devstral Small 2","description":"Compact multimodal coding model for repository exploration, file editing, and software agents","family":"devstral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":131072},"status":"deprecated","cost":{"input":0.1,"output":0.3,"cache_read":0.1,"cache_write":0.1},"canonical_model_id":"mistral/devstral-small-2"},"ministral-3b":{"id":"ministral-3b","name":"Ministral 3B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-03","release_date":"2024-10-16","last_updated":"2024-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"status":"deprecated","cost":{"input":0.1,"output":0.1,"cache_read":0.1,"cache_write":0.1},"canonical_model_id":"mistral/ministral-3b"},"ministral-14b":{"id":"ministral-14b","name":"Ministral 14B","description":"Compact multimodal Mistral model for local assistants, edge agents, and efficient tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":131072},"status":"deprecated","cost":{"input":0.2,"output":0.2,"cache_read":0.2,"cache_write":0.2},"canonical_model_id":"mistral/ministral-14b"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"devstral-2":{"id":"devstral-2","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":131072},"status":"deprecated","cost":{"input":0.4,"output":2,"cache_read":0.4,"cache_write":0.4},"canonical_model_id":"mistral/devstral-2512"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.5-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"deprecated","cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":2,"output":8,"cache_read":1,"cache_write":2},"canonical_model_id":"openai/gpt-4.1"},"mistral-large-3":{"id":"mistral-large-3","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":131072},"status":"deprecated","cost":{"input":0.5,"output":1.5,"cache_read":0.5,"cache_write":0.5},"canonical_model_id":"mistral/mistral-large-2512"},"mistral-medium-3.5":{"id":"mistral-medium-3.5","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":131072},"status":"deprecated","cost":{"input":1.5,"output":7.5,"cache_read":1.5,"cache_write":1.5},"canonical_model_id":"mistral/mistral-medium-2604"},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","cost":{"input":0.325,"output":1.95,"cache_read":0.065,"cache_write":0.40625},"canonical_model_id":"alibaba/qwen3.6-plus"},"gemini-3.1-pro":{"id":"gemini-3.1-pro","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3.1-pro-preview"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":0.75,"output":4.5,"cache_read":0.075,"cache_write":0.75},"canonical_model_id":"openai/gpt-5.4-mini"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":128000},"status":"deprecated","cost":{"input":0.2,"output":1.25,"cache_read":0.02,"cache_write":0.2},"canonical_model_id":"openai/gpt-5.4-nano"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":1.25,"output":4.25,"cache_read":0.15,"cache_write":1.25},"canonical_model_id":"meta/muse-spark-1.1"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":5},"canonical_model_id":"sakana/fugu-ultra"},"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"status":"deprecated","cost":{"input":0.95,"output":4,"cache_read":0.19,"cache_write":0.95},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K3-Fast":{"id":"moonshotai/Kimi-K3-Fast","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":4.5,"output":22.5,"cache_read":0.45,"cache_write":4.5},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":131072},"status":"deprecated","cost":{"input":0.95,"output":4,"cache_read":0.34,"cache_write":0.95},"canonical_model_id":"moonshotai/kimi-k2.6"},"mistralai/Mistral-7B-Instruct-v0.3":{"id":"mistralai/Mistral-7B-Instruct-v0.3","name":"Mistral 7B Instruct v0.3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2023-04-30","last_updated":"2023-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.2,"output":0.2,"cache_read":0.2,"cache_write":0.2}},"mistralai/Magistral-Small-2506":{"id":"mistralai/Magistral-Small-2506","name":"Magistral Small","description":"Open Mistral reasoning model for transparent step-by-step problem solving","family":"magistral","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":64000},"status":"deprecated","cost":{"input":0.5,"output":1.5,"cache_read":0.5,"cache_write":0.5},"canonical_model_id":"mistral/magistral-small-2506"},"mistralai/Ministral-8B-Instruct-2410":{"id":"mistralai/Ministral-8B-Instruct-2410","name":"Ministral 8B Instruct","description":"Efficient open Mistral edge model for on-device chat and function calling","family":"ministral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-10-16","last_updated":"2024-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"status":"deprecated","cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15},"canonical_model_id":"mistral/ministral-8b-instruct-2410"},"mistralai/Mistral-Nemo-Instruct-2407":{"id":"mistralai/Mistral-Nemo-Instruct-2407","name":"Mistral Nemo","description":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.02,"output":0.03,"cache_read":0.02,"cache_write":0.02},"canonical_model_id":"mistral/mistral-nemo"},"mistralai/Pixtral-12B-2409":{"id":"mistralai/Pixtral-12B-2409","name":"Pixtral 12B","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2024-09-11","last_updated":"2024-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"status":"deprecated","cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15},"canonical_model_id":"mistral/pixtral-12b"},"mistralai/Mistral-Small-4-119B-2603":{"id":"mistralai/Mistral-Small-4-119B-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000},"status":"deprecated","cost":{"input":0.15,"output":0.6,"cache_read":0.015,"cache_write":0.15},"canonical_model_id":"mistral/mistral-small-2603"},"mistralai/Codestral-22B-v0.1":{"id":"mistralai/Codestral-22B-v0.1","name":"Codestral-22B-v0.1","description":"Open Mistral code model for fill-in-the-middle and 80+ programming languages","family":"codestral","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-05-29","last_updated":"2024-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"status":"deprecated","cost":{"input":0.3,"output":0.9,"cache_read":0.3,"cache_write":0.3},"canonical_model_id":"mistral/codestral-22b-v0.1"},"pioneer/auto":{"id":"pioneer/auto","name":"Pioneer Auto","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2024-01-01","last_updated":"2025-06-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":4096},"status":"deprecated"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.3},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMaxAI/MiniMax-M2.7":{"id":"MiniMaxAI/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131000},"status":"deprecated","cost":{"input":0.279,"output":1.2,"cache_read":0.279,"cache_write":0.279},"canonical_model_id":"minimax/MiniMax-M2.7"},"poolside/laguna-s-2.1":{"id":"poolside/laguna-s-2.1","name":"Laguna S 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0.1,"output":0.2,"cache_read":0.01,"cache_write":0.1},"canonical_model_id":"poolside/laguna-s-2.1"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0.5,"output":1.2,"cache_read":0.1,"cache_write":0.5},"canonical_model_id":"thinkingmachines/inkling-small"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.07,"output":0.3,"cache_read":0.035,"cache_write":0.07},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":0.15,"output":0.6,"cache_read":0.015,"cache_write":0.15},"canonical_model_id":"openai/gpt-oss-120b"},"XiaomiMiMo/MiMo-V2.5-Pro":{"id":"XiaomiMiMo/MiMo-V2.5-Pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":131000},"status":"deprecated","cost":{"input":0.435,"output":0.87,"cache_read":0.0036,"cache_write":0.435},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"XiaomiMiMo/MiMo-V2.5":{"id":"XiaomiMiMo/MiMo-V2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":131072},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.0028,"cache_write":0.14},"canonical_model_id":"xiaomi/mimo-v2.5"},"LiquidAI/LFM2-24B-A2B":{"id":"LiquidAI/LFM2-24B-A2B","name":"LFM2 24B A2B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"liquid","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-31","last_updated":"2026-02-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.03,"output":0.12,"cache_read":0.03,"cache_write":0.03}},"meta-llama/Llama-3.2-3B-Instruct":{"id":"meta-llama/Llama-3.2-3B-Instruct","name":"Llama 3.2 3B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-31","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":80000},"status":"deprecated","cost":{"input":0.1,"output":0.335,"cache_read":0.1,"cache_write":0.1}},"meta-llama/Llama-3.3-70B-Instruct":{"id":"meta-llama/Llama-3.3-70B-Instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"status":"deprecated","cost":{"input":0.9,"output":0.9,"cache_read":0.9,"cache_write":0.9},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/Llama-3.1-8B-Instruct":{"id":"meta-llama/Llama-3.1-8B-Instruct","name":"Llama 3.1 8B Instruct","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0.2,"output":0.2,"cache_read":0.2,"cache_write":0.2},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta-llama/Llama-3.2-1B-Instruct":{"id":"meta-llama/Llama-3.2-1B-Instruct","name":"Llama 3.2 1B Instruct","description":"Compact instruction-tuned Llama model for multilingual dialogue and on-device assistants","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":60000},"status":"deprecated","cost":{"input":0.1,"output":0.201,"cache_read":0.1,"cache_write":0.1},"canonical_model_id":"meta/llama-3.2-1b-instruct"},"meta-llama/Llama-3.2-1B":{"id":"meta-llama/Llama-3.2-1B","name":"Llama-3.2-1B","description":"Compact open Llama base model for lightweight and on-device use","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":0.1,"output":0.1,"cache_read":0.1,"cache_write":0.1},"canonical_model_id":"meta/llama-3.2-1b"},"meta-llama/Llama-3.2-3B":{"id":"meta-llama/Llama-3.2-3B","name":"Llama-3.2-3B","description":"Small open Llama base model for lightweight text generation and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":0.1,"output":0.1,"cache_read":0.1,"cache_write":0.1},"canonical_model_id":"meta/llama-3.2-3b"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1040000,"output":128000},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":1.4},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.2-Fast":{"id":"zai-org/GLM-5.2-Fast","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"status":"deprecated","cost":{"input":2.1,"output":6.6,"cache_read":0.21,"cache_write":2.1},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202000,"output":131072},"status":"deprecated","cost":{"input":0.98,"output":3.08,"cache_read":0.182,"cache_write":0.98},"canonical_model_id":"zhipuai/glm-5.1"},"HuggingFaceTB/SmolLM3-3B-Base":{"id":"HuggingFaceTB/SmolLM3-3B-Base","name":"SmolLM3 3B Base","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-06-30","last_updated":"2025-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}},"nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-NVFP4":{"id":"nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-NVFP4","name":"Nemotron 3.5 Super VL 120B A12B","description":"NVIDIA vision-language reasoning model for text, image, and video understanding and agentic tool use","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":4096},"cost":{"input":0.08,"output":0.45,"cache_read":0.04,"cache_write":0.08},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16":{"id":"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65000},"status":"deprecated","cost":{"input":0.5,"output":2.5,"cache_read":0.15,"cache_write":0.5},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16":{"id":"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"status":"deprecated","cost":{"input":0.05,"output":0.2,"cache_read":0.05,"cache_write":0.05},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16":{"id":"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16","name":"Nemotron 3.5 Lightning 30B A3B","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.07,"output":0.2,"cache_read":0.035,"cache_write":0.07},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8":{"id":"nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"status":"deprecated","cost":{"input":0.09,"output":0.45,"cache_read":0.09,"cache_write":0.09},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0.435,"output":0.87,"cache_read":0.003625,"cache_write":0.435},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/DeepSeek-V3":{"id":"deepseek-ai/DeepSeek-V3","name":"DeepSeek-V3","description":"Open DeepSeek MoE chat model for coding, math, and general reasoning","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":8192},"status":"deprecated","cost":{"input":0.27,"output":1.12,"cache_read":0.135,"cache_write":0.27},"canonical_model_id":"deepseek/deepseek-v3"},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek-V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":131072},"status":"deprecated","cost":{"input":0.56,"output":1.68,"cache_read":0.56,"cache_write":0.56},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek-ai/DeepSeek-V4-Flash":{"id":"deepseek-ai/DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0.1,"output":0.2,"cache_read":0.0197,"cache_write":0.1},"canonical_model_id":"deepseek/deepseek-v4-flash"},"Qwen/Qwen3.6-27B":{"id":"Qwen/Qwen3.6-27B","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.6,"output":0.6,"cache_read":0.6,"cache_write":0.6},"canonical_model_id":"alibaba/qwen3.6-27b"},"Qwen/Qwen3.6-35B-A3B":{"id":"Qwen/Qwen3.6-35B-A3B","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"status":"deprecated","cost":{"input":0.14,"output":1,"cache_read":0.028,"cache_write":0.175},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen/Qwen3-1.7B-Base":{"id":"Qwen/Qwen3-1.7B-Base","name":"Qwen3 1.7B Base","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-03-31","last_updated":"2025-03-31","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.1,"output":0.1,"cache_read":0.1,"cache_write":0.1}},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B-A22B Instruct 2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"status":"deprecated","cost":{"input":1.2,"output":1.2,"cache_read":1.2,"cache_write":1.2},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"Qwen/Qwen3-4B-Instruct-2507":{"id":"Qwen/Qwen3-4B-Instruct-2507","name":"Qwen3 4B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.2,"output":0.2,"cache_read":0.2,"cache_write":0.2}},"Qwen/Qwen2.5-Coder-0.5B":{"id":"Qwen/Qwen2.5-Coder-0.5B","name":"Qwen2.5-Coder-0.5B","description":"Tiny open Qwen code model for lightweight completion and on-device coding","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.1,"output":0.1,"cache_read":0.1,"cache_write":0.1},"canonical_model_id":"alibaba/qwen2.5-coder-0.5b"},"Qwen/Qwen3.5-9B":{"id":"Qwen/Qwen3.5-9B","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.3,"output":0.3,"cache_read":0.3,"cache_write":0.3},"canonical_model_id":"alibaba/qwen3.5-9b"},"Qwen/Qwen3-8B":{"id":"Qwen/Qwen3-8B","name":"Qwen3 8B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-03-31","release_date":"2025-03-31","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":40960},"status":"deprecated","cost":{"input":0.2,"output":0.2,"cache_read":0.2,"cache_write":0.2}},"Qwen/Qwen3-4B-Base":{"id":"Qwen/Qwen3-4B-Base","name":"Qwen3 4B Base","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-03-31","last_updated":"2025-03-31","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}},"Qwen/Qwen3-32B":{"id":"Qwen/Qwen3-32B","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.9,"output":0.9,"cache_read":0.9,"cache_write":0.9},"canonical_model_id":"alibaba/qwen3-32b"},"fastino/gliguard-PII-multi":{"id":"fastino/gliguard-PII-multi","name":"GLiNER2-Guardrails-PII-Multi","description":"A 300M-parameter multilingual model that runs LLM safety moderation and PII detection in a single forward pass.","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}},"fastino/GLiNER-2.5-Decide":{"id":"fastino/GLiNER-2.5-Decide","name":"GLiNER-2.5-Decide","description":"Classification-tuned GLiNER2 large. Entities, relations, and JSON extraction still work.","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-09-24","last_updated":"2026-09-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.03,"output":0,"cache_read":0.03,"cache_write":0.03}},"fastino/gliner2-large-v1":{"id":"fastino/gliner2-large-v1","name":"GLiNER2 Large","description":"Higher-accuracy NER; broader entity coverage.","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-06-30","last_updated":"2025-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.03,"output":0,"cache_read":0.03,"cache_write":0.03}},"fastino/gliguard-LLMGuardrails-300M":{"id":"fastino/gliguard-LLMGuardrails-300M","name":"GLiGuard LLM Guardrails 300M","description":"LLM safety, PII, and adversarial-attack guardrails; 300M-param encoder.","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}},"fastino/gliner2-multi-v1":{"id":"fastino/gliner2-multi-v1","name":"GLiNER2 Multi","description":"Multilingual NER for cross-lingual extraction.","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-11-30","last_updated":"2025-11-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.03,"output":0,"cache_read":0.03,"cache_write":0.03}},"fastino/gliner2-multi-large-v1":{"id":"fastino/gliner2-multi-large-v1","name":"GLiNER2 Multi Large","description":"High-accuracy multilingual NER and span extraction.","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-11-30","last_updated":"2025-11-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.03,"output":0,"cache_read":0.03,"cache_write":0.03}},"fastino/gliner2-base-v1":{"id":"fastino/gliner2-base-v1","name":"GLiNER2 Base","description":"Named entity recognition; zero-shot span extraction.","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-06-30","last_updated":"2025-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.03,"output":0,"cache_read":0.03,"cache_write":0.03}},"fastino/gliner2.5-base-v1":{"id":"fastino/gliner2.5-base-v1","name":"GLiNER 2.5 Base","description":"English boundary NER and structured extraction with LoRA fine-tuning.","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-08-24","last_updated":"2026-08-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":0.03,"output":0,"cache_read":0.03,"cache_write":0.03}},"fastino/gliner2.5-multi-v1":{"id":"fastino/gliner2.5-multi-v1","name":"GLiNER 2.5 Multi","description":"Multilingual boundary NER and span extraction; non-trainable encoder.","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-08-24","last_updated":"2026-08-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":0.03,"output":0,"cache_read":0.03,"cache_write":0.03}},"fastino/gliner2-privacy-filter-PII-multi":{"id":"fastino/gliner2-privacy-filter-PII-multi","name":"GLiNER2 Privacy Filter PII (Multi)","description":"Multilingual PII entity detection across 40+ languages; non-trainable encoder.","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.03,"output":0,"cache_read":0.03,"cache_write":0.03}},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.5,"output":0.5,"cache_read":0.5,"cache_write":0.5},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemma-4-E2B-it":{"id":"google/gemma-4-E2B-it","name":"Gemma 4 E2B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.1,"output":0.1,"cache_read":0.1,"cache_write":0.1},"canonical_model_id":"google/gemma-4-E2B-it"},"google/gemma-4-12B-it":{"id":"google/gemma-4-12B-it","name":"Gemma 4 12B IT","description":"Compact Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.25,"output":0.25,"cache_read":0.25,"cache_write":0.25},"canonical_model_id":"google/gemma-4-12b-it"},"google/diffusiongemma-26B-A4B-it":{"id":"google/diffusiongemma-26B-A4B-it","name":"DiffusionGemma 26B-A4B IT","description":"Gemma 4 based discrete text diffusion model for fast chat, reasoning, and tool use","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-06-09","last_updated":"2026-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"status":"deprecated","cost":{"input":0.5,"output":0.5,"cache_read":0.5,"cache_write":0.5},"canonical_model_id":"google/diffusiongemma-26b-a4b-it"},"google/gemma-4-E4B-it":{"id":"google/gemma-4-E4B-it","name":"Gemma 4 E4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.2,"output":0.2,"cache_read":0.2,"cache_write":0.2},"canonical_model_id":"google/gemma-4-E4B-it"},"google/gemma-3-4b-pt":{"id":"google/gemma-3-4b-pt","name":"Gemma 3 4B (Pretrained)","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-02-28","last_updated":"2025-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.15,"output":0.15,"cache_read":0.15,"cache_write":0.15}}}},"cloudflare-workers-ai":{"id":"cloudflare-workers-ai","env":["CLOUDFLARE_ACCOUNT_ID","CLOUDFLARE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.cloudflare.com/client/v4/accounts/${CLOUDFLARE_ACCOUNT_ID}/ai/v1","name":"Cloudflare Workers AI","doc":"https://developers.cloudflare.com/workers-ai/models/","models":{"@cf/meta/llama-4-scout-17b-16e-instruct":{"id":"@cf/meta/llama-4-scout-17b-16e-instruct","name":"Llama 4 Scout 17B 16E Instruct","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":16384},"cost":{"input":0.27,"output":0.85},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"@cf/meta/llama-3.2-3b-instruct":{"id":"@cf/meta/llama-3.2-3b-instruct","name":"Llama 3.2 3B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":80000,"output":80000},"cost":{"input":0.0509,"output":0.335},"canonical_model_id":"meta/llama-3.2-3b"},"@cf/meta/llama-3.2-11b-vision-instruct":{"id":"@cf/meta/llama-3.2-11b-vision-instruct","name":"Llama 3.2 11B Vision Instruct","description":"Open Llama multimodal model for image understanding and text reasoning","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.0485,"output":0.676},"canonical_model_id":"meta/llama-3.2-11b-vision-instruct"},"@cf/meta/llama-3.3-70b-instruct-fp8-fast":{"id":"@cf/meta/llama-3.3-70b-instruct-fp8-fast","name":"Llama 3.3 70B Instruct fp8 Fast","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":24000,"output":24000},"cost":{"input":0.293,"output":2.253},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"@cf/meta/llama-3.2-1b-instruct":{"id":"@cf/meta/llama-3.2-1b-instruct","name":"Llama 3.2 1B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":60000,"output":60000},"cost":{"input":0.027,"output":0.201},"canonical_model_id":"meta/llama-3.2-1b"},"@cf/meta/llama-guard-3-8b":{"id":"@cf/meta/llama-guard-3-8b","name":"Llama Guard 3 8B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.484,"output":0.03},"canonical_model_id":"meta/llama-guard-3-8b"},"@cf/meta/llama-3.1-8b-instruct-fp8":{"id":"@cf/meta/llama-3.1-8b-instruct-fp8","name":"Llama 3.1 8B Instruct fp8","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000},"cost":{"input":0.152,"output":0.287},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"@cf/moonshotai/kimi-k2.6":{"id":"@cf/moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"@cf/moonshotai/kimi-k2.7-code":{"id":"@cf/moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"@cf/mistralai/mistral-small-3.1-24b-instruct":{"id":"@cf/mistralai/mistral-small-3.1-24b-instruct","name":"Mistral Small 3.1 24B Instruct","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.351,"output":0.555},"canonical_model_id":"mistral/mistral-small-3-1-24b-instruct-2503"},"@cf/openai/gpt-oss-20b":{"id":"@cf/openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.2,"output":0.3},"canonical_model_id":"openai/gpt-oss-20b"},"@cf/openai/gpt-oss-120b":{"id":"@cf/openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.35,"output":0.75},"canonical_model_id":"openai/gpt-oss-120b"},"@cf/ibm-granite/granite-4.0-h-micro":{"id":"@cf/ibm-granite/granite-4.0-h-micro","name":"Granite 4.0 H Micro","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"granite","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-10-02","last_updated":"2025-10-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.017,"output":0.112},"canonical_model_id":"ibm/granite-4-h-micro"},"@cf/zai-org/glm-5.3-flash":{"id":"@cf/zai-org/glm-5.3-flash","name":"Glm 5.3 Flash","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"@cf/zai-org/glm-5.2":{"id":"@cf/zai-org/glm-5.2","name":"Glm 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"@cf/zai-org/glm-4.7-flash":{"id":"@cf/zai-org/glm-4.7-flash","name":"GLM-4.7-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.0605,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"@cf/zai-org/glm-5.3":{"id":"@cf/zai-org/glm-5.3","name":"Glm 5.3","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"@cf/aisingapore/gemma-sea-lion-v4-27b-it":{"id":"@cf/aisingapore/gemma-sea-lion-v4-27b-it","name":"Gemma Sea Lion V4 27B It","description":"Gemma 3 27B tuned by AI Singapore for Southeast Asian languages and instruction following","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.351,"output":0.555},"canonical_model_id":"aisingapore/gemma-sea-lion-v4-27b-it"},"@cf/nvidia/nemotron-3-120b-a12b":{"id":"@cf/nvidia/nemotron-3-120b-a12b","name":"Nemotron 3 Super 120B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"@cf/deepseek-ai/deepseek-v4.1-flash-1017":{"id":"@cf/deepseek-ai/deepseek-v4.1-flash-1017","name":"Deepseek V4.1 Flash 1017","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.3,"output":1.2,"cache_read":0.006}},"@cf/deepseek-ai/deepseek-v4-pro-0813":{"id":"@cf/deepseek-ai/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"@cf/deepseek-ai/deepseek-r1-distill-qwen-32b":{"id":"@cf/deepseek-ai/deepseek-r1-distill-qwen-32b","name":"Deepseek R1 Distill Qwen 32B","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":80000,"output":80000},"cost":{"input":0.497,"output":4.881},"canonical_model_id":"deepseek/deepseek-r1"},"@cf/deepseek-ai/deepseek-v4-flash-0731":{"id":"@cf/deepseek-ai/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"@cf/qwen/qwen3.8-27b":{"id":"@cf/qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.45,"output":3.2,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.8-27b"},"@cf/qwen/qwen2.5-coder-32b-instruct":{"id":"@cf/qwen/qwen2.5-coder-32b-instruct","name":"Qwen2.5 Coder 32B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.66,"output":1},"canonical_model_id":"alibaba/qwen2.5-coder-32b-instruct"},"@cf/qwen/qwq-32b":{"id":"@cf/qwen/qwq-32b","name":"Qwq 32B","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-03-05","last_updated":"2025-03-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":24000,"output":24000},"cost":{"input":0.66,"output":1},"canonical_model_id":"alibaba/qwq-32b"},"@cf/qwen/qwen3-30b-a3b-fp8":{"id":"@cf/qwen/qwen3-30b-a3b-fp8","name":"Qwen3 30B A3b fp8","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.0509,"output":0.335},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"@cf/google/gemma-4-26b-a4b-it":{"id":"@cf/google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.1,"output":0.3,"cache_read":0.05},"canonical_model_id":"google/gemma-4-26b-a4b-it"}}},"github-copilot":{"id":"github-copilot","env":["GITHUB_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"https://api.githubcopilot.com","name":"GitHub Copilot","doc":"https://docs.github.com/en/copilot","models":{"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":936000,"output":64000},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.8-flash"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"input":372000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.5"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"input":372000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"google/gemini-4-argon"},"claude-sonnet-5.5":{"id":"claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":264000,"input":128000,"output":64000},"status":"deprecated","cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"input":372000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"mai-code-1.1-flash":{"id":"mai-code-1.1-flash","name":"MAI-Code-1.1-Flash","description":"Microsoft coding model with native vision support, optimized for fast and efficient software development","family":"mai","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":128000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02},"canonical_model_id":"microsoft/mai-code-1.1-flash"},"claude-opus-4.7":{"id":"claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":168000,"output":32000},"status":"deprecated","experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"status":"deprecated","provider":{"npm":"@ai-sdk/openai"},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"claude-sonnet-4.6":{"id":"claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024,"max":32000}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":168000,"output":32000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"status":"deprecated","provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"claude-haiku-4.5":{"id":"claude-haiku-4.5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":32000}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":136000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"mai-code-1-flash-picker":{"id":"mai-code-1-flash-picker","name":"MAI-Code-1-Flash","description":"Microsoft coding model built for fast, efficient assistance in everyday developer workflows","family":"mai","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2026-06-02","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":128000,"output":128000},"status":"deprecated","provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"microsoft/mai-code-1-flash"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":224000,"output":32000},"status":"deprecated","cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":936000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"claude-opus-5.5":{"id":"claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]},{"type":"budget_tokens","min":256,"max":32000}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":936000,"output":64000},"status":"deprecated","cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.6-flash"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":936000,"output":64000},"status":"deprecated","cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"claude-opus-4.8":{"id":"claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":168000,"output":64000},"experimental":{"modes":{"fast":{"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]},{"type":"budget_tokens","min":256,"max":24000}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":128000,"output":64000},"status":"deprecated","cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"claude-haiku-5.5":{"id":"claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"claude-fable-5.1":{"id":"claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.githubcopilot.com/v1"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"}}},"fastrouter":{"id":"fastrouter","env":["FASTROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://go.fastrouter.ai/api/v1","name":"FastRouter","doc":"https://fastrouter.ai/models","models":{"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.05,"output":3.5,"cache_read":0.205},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.08,"cache_read":0.12},"canonical_model_id":"zhipuai/glm-5"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.75,"output":3.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k2":{"id":"moonshotai/kimi-k2","name":"Kimi K2","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.55,"output":2.2},"canonical_model_id":"moonshotai/kimi-k2"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.07,"output":0.25,"cache_read":0.007},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-10-01","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-10-01","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-image-2":{"id":"openai/gpt-image-2","type":"image","name":"GPT Image 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":128000,"output":0},"canonical_model_id":"openai/gpt-image-2"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.09,"output":0.36},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2024-10-01","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-realtime-1.5":{"id":"openai/gpt-realtime-1.5","type":"realtime","name":"GPT Realtime 1.5","description":"Realtime speech-to-speech model for voice agents and customer support","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","audio","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":32000,"output":4096},"cost":{"input":4,"output":16,"cache_read":0.4,"input_audio":32,"output_audio":64},"canonical_model_id":"openai/gpt-realtime-1.5"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.74,"output":3.48,"cache_read":0.145},"canonical_model_id":"deepseek/deepseek-v4-pro"},"anthropic/claude-opus-4.1":{"id":"anthropic/claude-opus-4.1","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":32000}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":32000}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-sonnet-4":{"id":"anthropic/claude-sonnet-4","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":32000}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"leonardo-ai/lucid-origin":{"id":"leonardo-ai/lucid-origin","name":"Lucid Origin","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"lucid","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-01","last_updated":"2025-06-01","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":4096,"output":0}},"leonardo-ai/lucid-realism":{"id":"leonardo-ai/lucid-realism","name":"Lucid Realism","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"lucid","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-01","last_updated":"2025-06-01","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":4096,"output":0}},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"deepseek-ai/deepseek-r1-distill-llama-70b":{"id":"deepseek-ai/deepseek-r1-distill-llama-70b","name":"DeepSeek R1 Distill Llama 70B","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-10","release_date":"2025-01-23","last_updated":"2025-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":0.03,"output":0.14}},"bytedance/seedance-2":{"id":"bytedance/seedance-2","type":"video","name":"Seedance 2","description":"ByteDance unified multimodal audio-video model supporting text, image, video and audio references, editing, extension and synchronized stereo sound.","family":"seed","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-12","last_updated":"2026-10-04","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":4096,"output":0},"canonical_model_id":"bytedance-seed/seedance-2.0"},"qwen/qwen3-coder":{"id":"qwen/qwen3-coder","name":"Qwen3 Coder","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.3,"output":1.2},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"sarvam/sarvam-105b":{"id":"sarvam/sarvam-105b","name":"Sarvam 105B","description":"Flagship Indian-language reasoning model for enterprise multilingual applications","family":"sarvam","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-18","last_updated":"2026-02-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.04,"output":0.16,"cache_read":0.025},"canonical_model_id":"sarvam/sarvam-105b"},"sarvam/sarvam-30b":{"id":"sarvam/sarvam-30b","name":"Sarvam 30B","description":"Efficient Indian-language reasoning model for chat, coding, and multilingual work","family":"sarvam","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-18","last_updated":"2026-02-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.02,"output":0.1},"canonical_model_id":"sarvam/sarvam-30b"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-3-pro-image-preview":{"id":"google/gemini-3-pro-image-preview","type":"image","name":"Nano Banana Pro Preview","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12},"canonical_model_id":"google/gemini-3-pro-image-preview"},"google/veo3.1":{"id":"google/veo3.1","type":"video","name":"Veo 3.1","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-15","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":400000,"output":0},"canonical_model_id":"google/veo-3.1-generate-preview"},"google/veo3.1-lite":{"id":"google/veo3.1-lite","type":"video","name":"Veo 3.1 Lite","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":400000,"output":0},"canonical_model_id":"google/veo-3.1-lite-generate-preview"},"google/gemini-3.1-flash-image-preview":{"id":"google/gemini-3.1-flash-image-preview","type":"image","name":"Nano Banana 2 Preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":65536},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image-preview"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/veo3.1-fast":{"id":"google/veo3.1-fast","type":"video","name":"Veo 3.1 Fast","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-15","last_updated":"2026-01-01","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":400000,"output":0},"canonical_model_id":"google/veo-3.1-fast-generate-preview"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"wanx/wan-v2-6":{"id":"wanx/wan-v2-6","type":"video","name":"Wan 2.6","description":"Video model for text- and image-to-video with audio input and synchronized sound","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-03","last_updated":"2025-12-03","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":400000,"output":0},"canonical_model_id":"alibaba/wan2.6"},"x-ai/grok-4":{"id":"x-ai/grok-4","name":"Grok 4","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.75,"cache_write":15}},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5}},"canonical_model_id":"xai/grok-4.3"},"x-ai/grok-build-0.1":{"id":"x-ai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}},"canonical_model_id":"xai/grok-build-0.1"}}},"siliconflow":{"id":"siliconflow","env":["SILICONFLOW_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.siliconflow.com/v1","name":"SiliconFlow","doc":"https://cloud.siliconflow.com/models","models":{"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262000},"cost":{"input":2.7,"output":13.5,"cache_read":0.27},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.85916,"output":3.8,"cache_read":0.17993},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K2.5":{"id":"moonshotai/Kimi-K2.5","name":"moonshotai/Kimi-K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.45,"output":2.25,"cache_read":0.07},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"moonshotai/Kimi-K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.77,"output":3.4,"cache_read":0.14},"canonical_model_id":"moonshotai/kimi-k2.6"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMaxAI/MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":197000,"output":131000},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"inclusionAI/Ling-flash-2.0":{"id":"inclusionAI/Ling-flash-2.0","name":"inclusionAI/Ling-flash-2.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-18","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.14,"output":0.57}},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"openai/gpt-oss-20b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":8000},"cost":{"input":0.04,"output":0.18},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"openai/gpt-oss-120b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":8000},"cost":{"input":0.05,"output":0.45},"canonical_model_id":"openai/gpt-oss-120b"},"tencent/Hunyuan-A13B-Instruct":{"id":"tencent/Hunyuan-A13B-Instruct","name":"tencent/Hunyuan-A13B-Instruct","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-06-30","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.14,"output":0.57}},"tencent/Hy3":{"id":"tencent/Hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":262144},"cost":{"input":0.132,"output":0.528,"cache_read":0.033},"canonical_model_id":"tencent/hy3"},"zai-org/GLM-5":{"id":"zai-org/GLM-5","name":"zai-org/GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":205000,"output":131072},"cost":{"input":0.95,"output":2.55,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1049000,"output":262000},"cost":{"input":1.302,"output":4.092,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5V-Turbo":{"id":"zai-org/GLM-5V-Turbo","name":"zai-org/GLM-5V-Turbo","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24,"cache_write":0},"canonical_model_id":"zhipuai/glm-5v-turbo"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1049000,"output":262000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-4.5-Air":{"id":"zai-org/GLM-4.5-Air","name":"zai-org/GLM-4.5-Air","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-air","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.14,"output":0.86},"canonical_model_id":"zhipuai/glm-4.5-air"},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"zai-org/GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":205000,"output":131000},"cost":{"input":1.19,"output":3.74,"cache_read":0.6,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1049000,"output":262000},"cost":{"input":0.15,"output":0.5,"cache_read":0.03,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"stepfun-ai/Step-3.5-Flash":{"id":"stepfun-ai/Step-3.5-Flash","name":"stepfun-ai/Step-3.5-Flash","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","family":"step","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":66000},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"stepfun/step-3.5-flash"},"meituan-longcat/LongCat-2.0":{"id":"meituan-longcat/LongCat-2.0","name":"LongCat-2.0","description":"Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1049000,"output":131072},"cost":{"input":0.75,"output":2.95,"cache_read":0.015},"canonical_model_id":"meituan/longcat-2.0"},"deepseek-ai/DeepSeek-V4-Flash-Vision-Exp":{"id":"deepseek-ai/DeepSeek-V4-Flash-Vision-Exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.44,"output":1.32,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.22,"output":0.66,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-V3.1-Terminus":{"id":"deepseek-ai/DeepSeek-V3.1-Terminus","name":"deepseek-ai/DeepSeek-V3.1-Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-29","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.27,"output":1}},"deepseek-ai/DeepSeek-V3.2":{"id":"deepseek-ai/DeepSeek-V3.2","name":"deepseek-ai/DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.27,"output":0.42,"cache_read":0.135},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek-ai/DeepSeek-R1":{"id":"deepseek-ai/DeepSeek-R1","name":"deepseek-ai/DeepSeek-R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.5,"output":2.18},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.50162,"output":3.135,"cache_read":0.135},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/DeepSeek-V3":{"id":"deepseek-ai/DeepSeek-V3","name":"deepseek-ai/DeepSeek-V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.25,"output":1},"canonical_model_id":"deepseek/deepseek-v3-0324"},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"deepseek-ai/DeepSeek-V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.27,"output":1},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek-ai/DeepSeek-V4-Flash":{"id":"deepseek-ai/DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.13,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-ai/DeepSeek-V3.2-Exp":{"id":"deepseek-ai/DeepSeek-V3.2-Exp","name":"deepseek-ai/DeepSeek-V3.2-Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-10","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.27,"output":0.41}},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"nex-agi/Nex-N2-Pro":{"id":"nex-agi/Nex-N2-Pro","name":"Nex-N2-Pro","description":"Open agentic MoE model (397B total, 17B active) for coding, tool use, and research workflows","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"status":"deprecated","cost":{"input":0.5,"output":2.5,"cache_read":0.25},"canonical_model_id":"nex-agi/nex-n2-pro"},"Qwen/Qwen3.6-27B":{"id":"Qwen/Qwen3.6-27B","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":3.2},"canonical_model_id":"alibaba/qwen3.6-27b"},"Qwen/Qwen3-VL-32B-Thinking":{"id":"Qwen/Qwen3-VL-32B-Thinking","name":"Qwen/Qwen3-VL-32B-Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-21","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.2,"output":1.5}},"Qwen/Qwen3.6-35B-A3B":{"id":"Qwen/Qwen3.6-35B-A3B","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.24,"output":1.8,"cache_read":0.15},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen/Qwen3.5-27B":{"id":"Qwen/Qwen3.5-27B","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.25,"output":2},"canonical_model_id":"alibaba/qwen3.5-27b"},"Qwen/Qwen3-VL-32B-Instruct":{"id":"Qwen/Qwen3-VL-32B-Instruct","name":"Qwen/Qwen3-VL-32B-Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-21","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.2,"output":0.6}},"Qwen/Qwen3-VL-8B-Instruct":{"id":"Qwen/Qwen3-VL-8B-Instruct","name":"Qwen/Qwen3-VL-8B-Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.18,"output":0.68}},"Qwen/Qwen3-VL-30B-A3B-Instruct":{"id":"Qwen/Qwen3-VL-30B-A3B-Instruct","name":"Qwen/Qwen3-VL-30B-A3B-Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.29,"output":1},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"Qwen/Qwen3.5-35B-A3B":{"id":"Qwen/Qwen3.5-35B-A3B","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.24,"output":1.8,"cache_read":0.15},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"Qwen/Qwen3-30B-A3B-Instruct-2507":{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507","name":"Qwen/Qwen3-30B-A3B-Instruct-2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.09,"output":0.3},"canonical_model_id":"alibaba/qwen3-30b-a3b-instruct-2507"},"Qwen/Qwen3-Coder-480B-A35B-Instruct":{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct","name":"Qwen/Qwen3-Coder-480B-A35B-Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.25,"output":1},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"Qwen/Qwen3.5-397B-A17B":{"id":"Qwen/Qwen3.5-397B-A17B","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.39,"output":2.34},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"Qwen/Qwen3-VL-30B-A3B-Thinking":{"id":"Qwen/Qwen3-VL-30B-A3B-Thinking","name":"Qwen/Qwen3-VL-30B-A3B-Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-11","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.29,"output":1}},"Qwen/Qwen3-Coder-30B-A3B-Instruct":{"id":"Qwen/Qwen3-Coder-30B-A3B-Instruct","name":"Qwen/Qwen3-Coder-30B-A3B-Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.07,"output":0.28},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"Qwen/Qwen3-14B":{"id":"Qwen/Qwen3-14B","name":"Qwen/Qwen3-14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-30","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.07,"output":0.28}},"Qwen/Qwen3.5-122B-A10B":{"id":"Qwen/Qwen3.5-122B-A10B","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.26,"output":2.08},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"Qwen/Qwen2.5-72B-Instruct":{"id":"Qwen/Qwen2.5-72B-Instruct","name":"Qwen/Qwen2.5-72B-Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-09-18","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":33000,"output":4000},"cost":{"input":0.59,"output":0.59}},"Qwen/Qwen3.5-9B":{"id":"Qwen/Qwen3.5-9B","name":"Qwen/Qwen3.5-9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.15},"canonical_model_id":"alibaba/qwen3.5-9b"},"Qwen/Qwen3-8B":{"id":"Qwen/Qwen3-8B","name":"Qwen/Qwen3-8B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.06,"output":0.06},"canonical_model_id":"alibaba/qwen3-8b"},"Qwen/Qwen2.5-7B-Instruct":{"id":"Qwen/Qwen2.5-7B-Instruct","name":"Qwen/Qwen2.5-7B-Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-09-19","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":33000,"output":4000},"cost":{"input":0.05,"output":0.05},"canonical_model_id":"alibaba/qwen2.5-7b-instruct"},"Qwen/Qwen3.8-2.4T-A95B":{"id":"Qwen/Qwen3.8-2.4T-A95B","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1049000,"output":131000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"Qwen/Qwen3-32B":{"id":"Qwen/Qwen3-32B","name":"Qwen/Qwen3-32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.14,"output":0.57},"canonical_model_id":"alibaba/qwen3-32b"},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.75,"output":1,"cache_read":0.25},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemma-4-26B-A4B-it":{"id":"google/gemma-4-26B-A4B-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.14,"output":0.4,"cache_read":0.05},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemma-4-12B-it":{"id":"google/gemma-4-12B-it","name":"Gemma 4 12B IT","description":"Compact Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"google/gemma-4-12b-it"},"ByteDance-Seed/Seed-OSS-36B-Instruct":{"id":"ByteDance-Seed/Seed-OSS-36B-Instruct","name":"ByteDance-Seed/Seed-OSS-36B-Instruct","description":"Tool-capable chat model for instruction following and agentic application workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-04","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.21,"output":0.57}}}},"vivgrid":{"id":"vivgrid","env":["VIVGRID_API_KEY"],"npm":"@ai-sdk/openai","api":"https://api.vivgrid.com/v1","name":"Vivgrid","doc":"https://docs.vivgrid.com/models","models":{"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.75,"output":3.75,"cache_read":0.15},"canonical_model_id":"google/gemini-3.8-flash"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"gemini-3.1-flash-lite-preview":{"id":"gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":1},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT 5.6 Terra","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.25,"output":2,"cache_read":0.03},"canonical_model_id":"openai/gpt-5-mini"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.44,"output":1.32,"reasoning":1.32,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.28,"output":0.42,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":1.35,"output":3,"reasoning":3,"cache_read":0.05},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":1.74,"output":3.48,"cache_read":0.14},"canonical_model_id":"deepseek/deepseek-v4-pro"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":10,"output":50,"cache_read":1.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT 5.6 Sol","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":10,"output":50,"cache_read":0.5,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT 5.6 Luna","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":1,"output":6,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"openai/gpt-5.6-luna"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":3,"output":12,"cache_read":0.3,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.15,"output":0.5,"cache_read":0.04},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":1.2,"output":4.2,"cache_read":0.3},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.31,"output":1.23,"cache_read":0.01},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-5.1-codex-max":{"id":"gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex-max"},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"viv-fast":{"id":"viv-fast","name":"Viv Fast","description":"Fast coding model","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-09","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":256000},"cost":{"input":0.13,"output":0.4,"cache_read":0.05},"canonical_model_id":"vivgrid/viv-fast"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":1.2,"output":4.2,"cache_read":0.3},"canonical_model_id":"zhipuai/glm-5.3"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible"},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"}}},"standardcompute":{"id":"standardcompute","env":["STANDARDCOMPUTE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.stdcmpt.com/v1","name":"Standard Compute","doc":"https://docs.standardcompute.com/models","models":{"standardcompute":{"id":"standardcompute","name":"Standard Compute","description":"Flat-rate smart-routing gateway: one model id, each request routed across the customer's enabled models from a curated ~1M-context catalog (Claude, GPT, Gemini, DeepSeek, GLM, Qwen, Kimi, MiniMax, Grok, MiMo) or pinned to a user-selected model","family":"auto","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-03-01","last_updated":"2026-10-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"}},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.08,"output":0.5,"cache_read":0.012},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.5625,"output":1.8,"cache_read":0.105},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"mistralai/mistral-large-4-0":{"id":"mistralai/mistral-large-4-0","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.0826,"output":0.1652,"cache_read":0.01652},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":1.122,"output":3.366,"cache_read":0.1122},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":1.044,"output":2.088,"cache_read":0.2088},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.14,"output":0.42,"cache_read":0.0042},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2.5-pro":{"id":"xiaomi/mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.43,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"xiaomi/mimo-v2.6-flash":{"id":"xiaomi/mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":1.475,"output":4.425,"cache_read":0.295,"cache_write":1.84375},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.32,"output":1.28,"cache_read":0.064,"cache_write":0.4,"tiers":[{"input":0.96,"output":3.84,"cache_read":0.192,"cache_write":1.2,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.96,"output":3.84,"cache_read":0.192,"cache_write":1.2}},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.425,"output":2.55,"cache_read":0.085,"cache_write":0.53125},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.8-max-0902":{"id":"qwen/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"qwen/qwen3.7-flash":{"id":"qwen/qwen3.7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.038,"tiers":[{"input":0.1,"output":0.4,"cache_read":0.02,"cache_write":0.125,"tier":{"type":"context","size":32000}},{"input":0.2,"output":0.8,"cache_read":0.04,"cache_write":0.25,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.083333},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":0.375,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":850000,"output":50000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":1.25,"output":10,"cache_read":0.125,"cache_write":0.375,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":850000,"output":30000},"provider":{"npm":"@openrouter/ai-sdk-provider"},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"}}},"mistral":{"id":"mistral","env":["MISTRAL_API_KEY"],"npm":"@ai-sdk/mistral","name":"Mistral","doc":"https://docs.mistral.ai/getting-started/models/","models":{"devstral-2512":{"id":"devstral-2512","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.4,"output":2}},"mistral-medium-latest":{"id":"mistral-medium-latest","name":"Mistral Medium (latest)","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"mistral/mistral-medium-latest"},"zai-glm-5-3":{"id":"zai-glm-5-3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.3"},"magistral-medium-latest":{"id":"magistral-medium-latest","name":"Magistral Medium (latest)","description":"Mistral reasoning model for transparent analysis, math, and complex decisions","family":"magistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-03-17","last_updated":"2025-03-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"status":"deprecated","cost":{"input":2,"output":5}},"ministral-3b-2512":{"id":"ministral-3b-2512","name":"Ministral 3 3B","description":"Compact open vision-language model for edge deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":262144},"cost":{"input":0.1,"output":0.1,"cache_read":0.01},"canonical_model_id":"mistral/ministral-3-3b-instruct-2512"},"mistral-small-latest":{"id":"mistral-small-latest","name":"Mistral Small (latest)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015}},"mistral-medium-2604":{"id":"mistral-medium-2604","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5,"cache_read":0.15}},"mistral-large-2411":{"id":"mistral-large-2411","name":"Mistral Large 2.1","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-18","last_updated":"2024-11-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":2,"output":6},"canonical_model_id":"mistral/mistral-large-2411"},"mistral-large-4":{"id":"mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":262144},"status":"beta","cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"voxtral-mini-latest":{"id":"voxtral-mini-latest","type":"transcription","name":"Voxtral Mini (latest)","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"voxtral","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-04","last_updated":"2026-02-04","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0}},"mistral-large-latest":{"id":"mistral-large-latest","name":"Mistral Large (latest)","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5,"cache_read":0.05}},"devstral-medium-2507":{"id":"devstral-medium-2507","name":"Devstral Medium","description":"Legacy model retained for compatibility with older integrations","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-07-10","last_updated":"2025-07-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.4,"output":2}},"devstral-medium-latest":{"id":"devstral-medium-latest","name":"Devstral 2 (latest)","description":"Legacy model retained for compatibility with older integrations","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.4,"output":2}},"pixtral-large-latest":{"id":"pixtral-large-latest","name":"Pixtral Large (latest)","description":"Mistral's larger vision model for document-heavy image understanding and chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2024-11-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":2,"output":6}},"pixtral-12b":{"id":"pixtral-12b","name":"Pixtral 12B","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2024-09-01","last_updated":"2024-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.15,"output":0.15}},"ministral-8b-2512":{"id":"ministral-8b-2512","name":"Ministral 3 8B","description":"Compact open vision-language model for edge deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.15,"output":0.15,"cache_read":0.015},"canonical_model_id":"mistral/ministral-3-8b-instruct-2512"},"mistral-small-2603":{"id":"mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015}},"devstral-small-2507":{"id":"devstral-small-2507","name":"Devstral Small","description":"Legacy model retained for compatibility with older integrations","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-07-10","last_updated":"2025-07-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.1,"output":0.3}},"devstral-small-2505":{"id":"devstral-small-2505","name":"Devstral Small 2505","description":"Legacy model retained for compatibility with older integrations","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.1,"output":0.3}},"codestral-2508":{"id":"codestral-2508","name":"Codestral 25.08","description":"Mistral code model for low-latency completion, fill-in-the-middle, and code generation","family":"codestral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":8192},"cost":{"input":0.3,"output":0.9,"cache_read":0.03},"canonical_model_id":"mistral/codestral-2508"},"zai-glm-5-2":{"id":"zai-glm-5-2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"labs-leanstral-1-5-1":{"id":"labs-leanstral-1-5-1","name":"Leanstral 1.5","description":"An updated Lean 4 formal proof engineering model optimised for automated theorem proving and autoformalization. 119B total parameters, 6.5B active.","family":"leanstral","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"status":"deprecated","cost":{"input":0,"output":0}},"ministral-8b-latest":{"id":"ministral-8b-latest","name":"Ministral 8B (latest)","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-10-01","last_updated":"2024-10-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"cost":{"input":0.15,"output":0.15,"cache_read":0.015}},"ministral-14b-2512":{"id":"ministral-14b-2512","name":"Ministral 3 14B","description":"Open vision-language model for efficient local deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.2,"output":0.2,"cache_read":0.02},"canonical_model_id":"mistral/ministral-3-14b-instruct-2512"},"voxtral-small-latest":{"id":"voxtral-small-latest","name":"Voxtral Small (latest)","description":"Instruct model with native audio input for speech understanding and tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32000},"cost":{"input":0.1,"output":0.4}},"ministral-3b-latest":{"id":"ministral-3b-latest","name":"Ministral 3B (latest)","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-10-01","last_updated":"2024-10-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":128000},"cost":{"input":0.1,"output":0.1,"cache_read":0.01}},"mistral-nemo":{"id":"mistral-nemo","name":"Mistral Nemo","description":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.15,"output":0.15}},"labs-devstral-small-2512":{"id":"labs-devstral-small-2512","name":"Devstral Small 2","description":"Legacy model retained for compatibility with older integrations","family":"devstral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"status":"deprecated","cost":{"input":0,"output":0}},"mistral-medium-2508":{"id":"mistral-medium-2508","name":"Mistral Medium 3.1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-08-12","last_updated":"2025-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.4,"output":2}},"mistral-embed":{"id":"mistral-embed","type":"embedding","name":"Mistral Embed","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"mistral-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-12-11","last_updated":"2023-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":3072},"cost":{"input":0.1,"output":0}},"voxtral-mini-tts-latest":{"id":"voxtral-mini-tts-latest","type":"speech","name":"Voxtral Mini TTS (latest)","description":"Multilingual text-to-speech model with zero-shot voice cloning","family":"voxtral","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-23","last_updated":"2026-03-23","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":0,"output":0}},"glm-5-2":{"id":"glm-5-2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"codestral-latest":{"id":"codestral-latest","name":"Codestral (latest)","description":"Mistral code model for completions, refactors, and developer IDE workflows","family":"codestral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-05-29","last_updated":"2025-01-04","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":4096},"cost":{"input":0.3,"output":0.9,"cache_read":0.03}},"mistral-medium-2505":{"id":"mistral-medium-2505","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":0.4,"output":2}},"open-mistral-nemo":{"id":"open-mistral-nemo","name":"Open Mistral Nemo","description":"Legacy model retained for compatibility with older integrations","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.15,"output":0.15}},"mistral-small-2506":{"id":"mistral-small-2506","name":"Mistral Small 3.2","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":0.1,"output":0.3}},"voxtral-small-2507":{"id":"voxtral-small-2507","name":"Voxtral Small","description":"Instruct model with native audio input for speech understanding and tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32000},"cost":{"input":0.1,"output":0.4}},"devstral-latest":{"id":"devstral-latest","name":"Devstral 2","description":"Legacy model retained for compatibility with older integrations","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.4,"output":2}},"mistral-large-2512":{"id":"mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5,"cache_read":0.05}}}},"melious":{"id":"melious","env":["MELIOUS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.melious.ai/v1","name":"Melious","doc":"https://melious.ai/docs/reference/models","models":{"deepseek-r1-0528":{"id":"deepseek-r1-0528","name":"DeepSeek-R1","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":32768},"status":"deprecated","cost":{"input":0.69552,"output":2.78208,"cache_read":0.185472},"canonical_model_id":"deepseek/deepseek-r1"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.78505,"output":3.3645,"cache_read":0.280375},"canonical_model_id":"moonshotai/kimi-k2.6"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":64000},"status":"deprecated","cost":{"input":0.34776,"output":0.5796,"cache_read":0.092736},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.33675,"output":1.347,"cache_read":0.033675},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131072},"status":"deprecated","cost":{"input":1.50696,"output":4.6368,"cache_read":0.370944},"canonical_model_id":"zhipuai/glm-5.1"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.728975,"output":3.644875,"cache_read":0.190655},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3.196275,"output":15.981375,"cache_read":0.841125},"canonical_model_id":"moonshotai/kimi-k3"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"status":"deprecated","cost":{"input":0.5796,"output":2.95596,"cache_read":0.139104},"canonical_model_id":"moonshotai/kimi-k2.5"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.11592,"output":0.46368,"cache_read":0.023184},"canonical_model_id":"zhipuai/glm-5.3-flash"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.23184,"output":1.1592,"cache_read":0.011592},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.11592,"output":0.2898,"cache_read":0.023184},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"glm-5":{"id":"glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131072},"status":"deprecated","cost":{"input":1.10124,"output":3.36168,"cache_read":0.266616},"canonical_model_id":"zhipuai/glm-5"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.1592,"output":3.4776,"cache_read":0.23184},"canonical_model_id":"zhipuai/glm-5.3"}}},"siliconflow-cn":{"id":"siliconflow-cn","env":["SILICONFLOW_CN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.siliconflow.cn/v1","name":"SiliconFlow (China)","doc":"https://cloud.siliconflow.com/models","models":{"inclusionAI/Ling-flash-2.0":{"id":"inclusionAI/Ling-flash-2.0","name":"inclusionAI/Ling-flash-2.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-18","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.14944748,"output":0.59778991}},"Pro/moonshotai/Kimi-K2.5":{"id":"Pro/moonshotai/Kimi-K2.5","name":"Pro/moonshotai/Kimi-K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"status":"deprecated","cost":{"input":0.45,"output":2.25,"cache_read":0.07},"canonical_model_id":"moonshotai/kimi-k2.5"},"Pro/moonshotai/Kimi-K2.6":{"id":"Pro/moonshotai/Kimi-K2.6","name":"Pro/moonshotai/Kimi-K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.97140861,"output":4.0350819,"cache_read":0.16439223},"canonical_model_id":"moonshotai/kimi-k2.6"},"Pro/MiniMaxAI/MiniMax-M2.5":{"id":"Pro/MiniMaxAI/MiniMax-M2.5","name":"Pro/MiniMaxAI/MiniMax-M2.5","description":"Frontier MiniMax model for engineering, office tasks, and agentic reasoning","family":"minimax","attachment":false,"reasoning":false,"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":192000,"output":131000},"status":"deprecated","cost":{"input":0.3,"output":1.22},"canonical_model_id":"minimax/MiniMax-M2.5"},"Pro/zai-org/GLM-5":{"id":"Pro/zai-org/GLM-5","name":"Pro/zai-org/GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":205000,"output":205000},"status":"deprecated","cost":{"input":1,"output":3.2},"canonical_model_id":"zhipuai/glm-5"},"Pro/zai-org/GLM-5.1":{"id":"Pro/zai-org/GLM-5.1","name":"Pro/zai-org/GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":205000,"output":205000},"cost":{"input":0.89668487,"output":3.58673947,"cache_read":0.19428172,"cache_write":0,"tiers":[{"input":1.19557982,"output":4.18452938,"cache_read":0.29889496,"cache_write":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5.1"},"Pro/deepseek-ai/DeepSeek-V3.1-Terminus":{"id":"Pro/deepseek-ai/DeepSeek-V3.1-Terminus","name":"Pro/deepseek-ai/DeepSeek-V3.1-Terminus","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-29","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.59778991,"output":1.79336973,"cache_read":0.05977899}},"Pro/deepseek-ai/DeepSeek-V3.2":{"id":"Pro/deepseek-ai/DeepSeek-V3.2","name":"Pro/deepseek-ai/DeepSeek-V3.2","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.59778991,"output":0.89668487,"cache_read":0.05977899},"canonical_model_id":"deepseek/deepseek-v3.2"},"Pro/deepseek-ai/DeepSeek-R1":{"id":"Pro/deepseek-ai/DeepSeek-R1","name":"Pro/deepseek-ai/DeepSeek-R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.59778991,"output":2.39115964,"cache_read":0.05977899},"canonical_model_id":"deepseek/deepseek-r1-0528"},"Pro/deepseek-ai/DeepSeek-V3":{"id":"Pro/deepseek-ai/DeepSeek-V3","name":"Pro/deepseek-ai/DeepSeek-V3","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.29889496,"output":1.19557982,"cache_read":0.0298895},"canonical_model_id":"deepseek/deepseek-v3-0324"},"tencent/Hunyuan-A13B-Instruct":{"id":"tencent/Hunyuan-A13B-Instruct","name":"tencent/Hunyuan-A13B-Instruct","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-06-30","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.14944748,"output":0.59778991}},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1049000,"output":262000},"cost":{"input":1.19557982,"output":4.18452938,"cache_read":0.29889496,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-4.5-Air":{"id":"zai-org/GLM-4.5-Air","name":"zai-org/GLM-4.5-Air","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-air","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.14944748,"output":0.89668487},"canonical_model_id":"zhipuai/glm-4.5-air"},"PaddlePaddle/PaddleOCR-VL-1.5":{"id":"PaddlePaddle/PaddleOCR-VL-1.5","name":"PaddlePaddle/PaddleOCR-VL-1.5","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-29","last_updated":"2026-01-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":0,"output":0}},"stepfun-ai/Step-3.5-Flash":{"id":"stepfun-ai/Step-3.5-Flash","name":"stepfun-ai/Step-3.5-Flash","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","family":"step","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.10461323,"output":0.3138397},"canonical_model_id":"stepfun/step-3.5-flash"},"deepseek-ai/DeepSeek-V3.1-Terminus":{"id":"deepseek-ai/DeepSeek-V3.1-Terminus","name":"deepseek-ai/DeepSeek-V3.1-Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-29","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.59778991,"output":1.79336973,"cache_read":0.05977899}},"deepseek-ai/DeepSeek-V3.2":{"id":"deepseek-ai/DeepSeek-V3.2","name":"deepseek-ai/DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.59778991,"output":0.89668487,"cache_read":0.05977899},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek-ai/DeepSeek-R1":{"id":"deepseek-ai/DeepSeek-R1","name":"deepseek-ai/DeepSeek-R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.59778991,"output":2.39115964,"cache_read":0.05977899},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"deepseek-ai/DeepSeek-V4-Pro","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1049000,"output":393000},"cost":{"input":1.79336973,"output":3.58673947,"cache_read":0.14944748},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/DeepSeek-V3":{"id":"deepseek-ai/DeepSeek-V3","name":"deepseek-ai/DeepSeek-V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.29889496,"output":1.19557982,"cache_read":0.0298895},"canonical_model_id":"deepseek/deepseek-v3-0324"},"deepseek-ai/DeepSeek-V4-Flash":{"id":"deepseek-ai/DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.44834243,"output":1.3450273,"cache_read":0.04483424},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-ai/DeepSeek-OCR":{"id":"deepseek-ai/DeepSeek-OCR","name":"deepseek-ai/DeepSeek-OCR","description":"OCR model for extracting structured text from documents and screenshots","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-10-20","last_updated":"2025-10-20","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0}},"Qwen/Qwen3-VL-32B-Thinking":{"id":"Qwen/Qwen3-VL-32B-Thinking","name":"Qwen/Qwen3-VL-32B-Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-21","last_updated":"2025-11-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.14944748,"output":1.49447478}},"Qwen/Qwen3.6-35B-A3B":{"id":"Qwen/Qwen3.6-35B-A3B","name":"Qwen/Qwen3.6-35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.26900546,"output":1.61403276},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen/Qwen3.5-27B":{"id":"Qwen/Qwen3.5-27B","name":"Qwen/Qwen3.5-27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.08966849,"output":0.71734789,"tiers":[{"input":0.26900546,"output":2.15204368,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-27b"},"Qwen/Qwen3-235B-A22B-Thinking-2507":{"id":"Qwen/Qwen3-235B-A22B-Thinking-2507","name":"Qwen/Qwen3-235B-A22B-Thinking-2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"status":"deprecated","cost":{"input":0.13,"output":0.6},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"Qwen/Qwen3-VL-32B-Instruct":{"id":"Qwen/Qwen3-VL-32B-Instruct","name":"Qwen/Qwen3-VL-32B-Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-21","last_updated":"2025-11-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.14944748,"output":0.59778991}},"Qwen/Qwen3.5-4B":{"id":"Qwen/Qwen3.5-4B","name":"Qwen/Qwen3.5-4B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-03-02","last_updated":"2026-03-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.5-4b"},"Qwen/Qwen3-VL-8B-Instruct":{"id":"Qwen/Qwen3-VL-8B-Instruct","name":"Qwen/Qwen3-VL-8B-Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-11-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.07472374,"output":0.29889496}},"Qwen/Qwen3-VL-30B-A3B-Instruct":{"id":"Qwen/Qwen3-VL-30B-A3B-Instruct","name":"Qwen/Qwen3-VL-30B-A3B-Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.10461323,"output":0.41845294},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"Qwen/Qwen3.5-35B-A3B":{"id":"Qwen/Qwen3.5-35B-A3B","name":"Qwen/Qwen3.5-35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.05977899,"output":0.47823193,"tiers":[{"input":0.23911596,"output":1.91292771,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"Qwen/Qwen3-30B-A3B-Instruct-2507":{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507","name":"Qwen/Qwen3-30B-A3B-Instruct-2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.10461323,"output":0.41845294},"canonical_model_id":"alibaba/qwen3-30b-a3b-instruct-2507"},"Qwen/Qwen3-Coder-480B-A35B-Instruct":{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct","name":"Qwen/Qwen3-Coder-480B-A35B-Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"status":"deprecated","cost":{"input":0.25,"output":1},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"Qwen/Qwen3.5-397B-A17B":{"id":"Qwen/Qwen3.5-397B-A17B","name":"Qwen/Qwen3.5-397B-A17B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.29,"output":1.74},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"Qwen/Qwen3-VL-30B-A3B-Thinking":{"id":"Qwen/Qwen3-VL-30B-A3B-Thinking","name":"Qwen/Qwen3-VL-30B-A3B-Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-11","last_updated":"2025-11-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.10461323,"output":0.41845294}},"Qwen/Qwen3-Coder-30B-A3B-Instruct":{"id":"Qwen/Qwen3-Coder-30B-A3B-Instruct","name":"Qwen/Qwen3-Coder-30B-A3B-Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.10461323,"output":0.41845294},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"Qwen/Qwen3-14B":{"id":"Qwen/Qwen3-14B","name":"Qwen/Qwen3-14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-30","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.07472374,"output":0.29889496}},"Qwen/Qwen3.5-122B-A10B":{"id":"Qwen/Qwen3.5-122B-A10B","name":"Qwen/Qwen3.5-122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.11955798,"output":0.95646386,"tiers":[{"input":0.29889496,"output":2.39115964,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"Qwen/Qwen2.5-72B-Instruct":{"id":"Qwen/Qwen2.5-72B-Instruct","name":"Qwen/Qwen2.5-72B-Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-09-18","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":33000,"output":4000},"cost":{"input":0.61721808,"output":0.61721808}},"Qwen/Qwen3.5-9B":{"id":"Qwen/Qwen3.5-9B","name":"Qwen/Qwen3.5-9B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.07472374,"output":0.59778991,"tiers":[{"input":0.22417122,"output":1.79336973,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-9b"},"Qwen/Qwen3-8B":{"id":"Qwen/Qwen3-8B","name":"Qwen/Qwen3-8B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-8b"},"Qwen/Qwen2.5-7B-Instruct":{"id":"Qwen/Qwen2.5-7B-Instruct","name":"Qwen/Qwen2.5-7B-Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-09-19","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":33000,"output":4000},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen2.5-7b-instruct"},"Qwen/Qwen3-32B":{"id":"Qwen/Qwen3-32B","name":"Qwen/Qwen3-32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.14944748,"output":0.59778991},"canonical_model_id":"alibaba/qwen3-32b"},"ByteDance-Seed/Seed-OSS-36B-Instruct":{"id":"ByteDance-Seed/Seed-OSS-36B-Instruct","name":"ByteDance-Seed/Seed-OSS-36B-Instruct","description":"Tool-capable chat model for instruction following and agentic application workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-04","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.22417122,"output":0.59778991}}}},"302ai":{"id":"302ai","env":["302AI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.302.ai/v1","name":"302.AI","doc":"https://doc.302.ai","models":{"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.8-flash"},"qwen3.5-35b-a3b":{"id":"qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.06,"output":0.46,"tiers":[{"input":0.23,"output":1.83,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.1-flash-lite"},"qwen3-30b-a3b":{"id":"qwen3-30b-a3b","name":"Qwen3-30B-A3B","description":"Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.11,"output":1.08},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"kimi-k2-thinking":{"id":"kimi-k2-thinking","name":"kimi-k2-thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.575,"output":2.3},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4}},"canonical_model_id":"openai/gpt-6-sol"},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"gemini-2.5-flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":2.5},"canonical_model_id":"google/gemini-2.5-flash"},"gpt-5.1":{"id":"gpt-5.1","name":"gpt-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10},"canonical_model_id":"openai/gpt-5.1"},"grok-4-fast-non-reasoning":{"id":"grok-4-fast-non-reasoning","name":"grok-4-fast-non-reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5},"canonical_model_id":"xai/grok-4-fast-non-reasoning"},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.8,"output":5.3},"canonical_model_id":"alibaba/qwen3.7-max"},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.5"},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"gpt-4.1-mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6},"canonical_model_id":"openai/gpt-4.1-mini"},"gemini-2.5-flash-nothink":{"id":"gemini-2.5-flash-nothink","name":"gemini-2.5-flash-nothink","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":2.5},"canonical_model_id":"google/gemini-2.5-flash"},"gpt-5-pro":{"id":"gpt-5-pro","name":"gpt-5-pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":15,"output":120,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-pro"},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.228,"output":0.92,"cache_read":0.023,"cache_write":0.286,"tiers":[{"input":0.68,"output":2.74,"cache_read":0.068,"cache_write":0.86,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.68,"output":2.74,"cache_read":0.068,"cache_write":0.86}},"canonical_model_id":"alibaba/qwen3.7-plus"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"kimi-k2-0905-preview":{"id":"kimi-k2-0905-preview","name":"kimi-k2-0905-preview","description":"September 2025 Kimi K2 instruction update with a 256K context and improved agentic coding","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.62854,"output":2.5146},"canonical_model_id":"moonshotai/kimi-k2-0905"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1},"canonical_model_id":"openai/gpt-6-astra"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":4,"output":10,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-4-argon"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"claude-opus-4-7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"anthropic/claude-opus-4-7"},"gemini-3.1-flash-lite-preview":{"id":"gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-5-mini":{"id":"gpt-5-mini","name":"gpt-5-mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"MiniMax-M2":{"id":"MiniMax-M2","name":"MiniMax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.33,"output":1.32},"canonical_model_id":"minimax/MiniMax-M2"},"qwen3.7-max-2026-06-08":{"id":"qwen3.7-max-2026-06-08","name":"qwen3.7-max-2026-06-08","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.8,"output":5.3},"canonical_model_id":"alibaba/qwen3.7-max"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"claude-haiku-4-5","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5},"canonical_model_id":"anthropic/claude-haiku-4-5"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.28,"output":1.7},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"cache_write":0.5},"canonical_model_id":"xai/grok-4.7"},"gpt-5.2":{"id":"gpt-5.2","name":"gpt-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14},"canonical_model_id":"openai/gpt-5.2"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4},"canonical_model_id":"moonshotai/kimi-k2.6"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.5"},"claude-opus-4-1-20250805":{"id":"claude-opus-4-1-20250805","name":"claude-opus-4-1-20250805","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"gemini-3-pro-image-preview":{"id":"gemini-3-pro-image-preview","type":"image","name":"gemini-3-pro-image-preview","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-06","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":64000},"cost":{"input":2,"output":120},"canonical_model_id":"google/gemini-3-pro-image-preview"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"claude-sonnet-4-6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5}},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gemini-2.5-flash-image":{"id":"gemini-2.5-flash-image","type":"image","name":"gemini-2.5-flash-image","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":30},"canonical_model_id":"google/gemini-2.5-flash-image"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"deepseek-v3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.29,"output":0.43},"canonical_model_id":"deepseek/deepseek-v3.2"},"glm-4.5v":{"id":"glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":16384},"cost":{"input":0.29,"output":0.86,"tiers":[{"input":0.58,"output":1.72,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.5v"},"o3":{"id":"o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8},"canonical_model_id":"openai/o3"},"gpt-5.4":{"id":"gpt-5.4","name":"gpt-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":0,"tiers":[{"input":5,"output":22.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5}},"canonical_model_id":"openai/gpt-5.4"},"deepseek-flash":{"id":"deepseek-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02}},"canonical_model_id":"openai/gpt-6-luna"},"qwen3-235b-a22b-instruct-2507":{"id":"qwen3-235b-a22b-instruct-2507","name":"qwen3-235b-a22b-instruct-2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":65536},"cost":{"input":0.29,"output":1.143},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"claude-opus-4-7-thinking":{"id":"claude-opus-4-7-thinking","name":"claude-opus-4-7-thinking","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-7"},"grok-4-1-fast-reasoning":{"id":"grok-4-1-fast-reasoning","name":"grok-4-1-fast-reasoning","description":"xAI's fast agentic tool-calling model with a 2M context window and built-in reasoning","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5},"canonical_model_id":"xai/grok-4.1-fast-reasoning"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"grok-4-fast-reasoning":{"id":"grok-4-fast-reasoning","name":"grok-4-fast-reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5},"canonical_model_id":"xai/grok-4-fast-reasoning"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.72,"output":2.88,"cache_read":0.144,"cache_write":0.9,"tiers":[{"input":1.44,"output":5.76,"cache_read":0.144,"cache_write":0.9,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":1.44,"output":5.76,"cache_read":0.144,"cache_write":0.9}},"canonical_model_id":"minimax/MiniMax-M3"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.6-sol"},"gemini-2.5-flash-preview-09-2025":{"id":"gemini-2.5-flash-preview-09-2025","name":"gemini-2.5-flash-preview-09-2025","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-09-26","last_updated":"2025-09-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":2.5}},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"grok-4.20-beta-0309-reasoning":{"id":"grok-4.20-beta-0309-reasoning","name":"grok-4.20-beta-0309-reasoning","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"gpt-5.3-chat-latest":{"id":"gpt-5.3-chat-latest","name":"GPT-5.3 Chat (latest)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.75,"output":14},"canonical_model_id":"openai/gpt-5.3-chat-latest"},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.7"},"gemini-2.0-flash-lite":{"id":"gemini-2.0-flash-lite","name":"gemini-2.0-flash-lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-11","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":8192},"cost":{"input":0.075,"output":0.3},"canonical_model_id":"google/gemini-2.0-flash-lite"},"qwen3-235b-a22b":{"id":"qwen3-235b-a22b","name":"Qwen3-235B-A22B","description":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.29,"output":2.86},"canonical_model_id":"alibaba/qwen3-235b-a22b"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.16,"output":6.36,"cache_read":0.168,"cache_write":2.58},"canonical_model_id":"alibaba/qwen3.8-max"},"doubao-seed-1-8-251215":{"id":"doubao-seed-1-8-251215","name":"doubao-seed-1-8-251215","description":"ByteDance Seed model for multimodal reasoning, long-context analysis, and agent workflows","family":"seed","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-12-28","last_updated":"2025-12-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":224000,"output":64000},"cost":{"input":0.1143,"output":0.286,"tiers":[{"input":0.1715,"output":2.286,"tier":{"type":"context","size":32000}},{"input":0.343,"output":3.43,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-8"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"claude-sonnet-4-6-thinking":{"id":"claude-sonnet-4-6-thinking","name":"claude-sonnet-4-6-thinking","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5}},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gpt-5-thinking":{"id":"gpt-5-thinking","name":"gpt-5-thinking","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10},"canonical_model_id":"openai/gpt-5"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2}},"canonical_model_id":"openai/gpt-6.1-sol"},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.19,"output":1.13,"cache_read":0.0171,"cache_write":0.214,"tiers":[{"input":0.76,"output":4.52,"cache_read":0.0686,"cache_write":0.857,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.76,"output":4.52,"cache_read":0.0686,"cache_write":0.857}},"canonical_model_id":"alibaba/qwen3.6-flash"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.6-luna"},"gemini-3.1-flash-image-preview":{"id":"gemini-3.1-flash-image-preview","type":"image","name":"gemini-3.1-flash-image-preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":60},"canonical_model_id":"google/gemini-3.1-flash-image-preview"},"claude-opus-4-5-20251101":{"id":"claude-opus-4-5-20251101","name":"claude-opus-4-5-20251101","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5"},"claude-opus-4-1-20250805-thinking":{"id":"claude-opus-4-1-20250805-thinking","name":"claude-opus-4-1-20250805-thinking","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"glm-4.7":{"id":"glm-4.7","name":"glm-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.286,"output":1.142,"tiers":[{"input":0.572,"output":2.29,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.7"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"tiers":[{"input":4,"output":18,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"gpt-4o":{"id":"gpt-4o","name":"gpt-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10},"canonical_model_id":"openai/gpt-4o"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.6-flash"},"gemini-2.5-flash-lite-preview-09-2025":{"id":"gemini-2.5-flash-lite-preview-09-2025","name":"gemini-2.5-flash-lite-preview-09-2025","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-09-26","last_updated":"2025-09-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1,"output":0.4}},"gpt-5.6-luna-pro":{"id":"gpt-5.6-luna-pro","name":"gpt-5.6-luna-pro","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.6-luna"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"claude-opus-5-thinking":{"id":"claude-opus-5-thinking","name":"claude-opus-5-thinking","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"gpt-4.1-nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4},"canonical_model_id":"openai/gpt-4.1-nano"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"claude-sonnet-4-5-20250929-thinking":{"id":"claude-sonnet-4-5-20250929-thinking","name":"claude-sonnet-4-5-20250929-thinking","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5}},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"glm-5-turbo":{"id":"glm-5-turbo","name":"glm-5-turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":0.72,"output":3.2,"tiers":[{"input":1.1,"output":3.8,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5-turbo"},"glm-4.5":{"id":"glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.286,"output":1.143,"tiers":[{"input":0.5714,"output":2.29,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.5"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.627,"output":3.3},"canonical_model_id":"moonshotai/kimi-k2.5"},"gpt-5.6-sol-pro":{"id":"gpt-5.6-sol-pro","name":"gpt-5.6-sol-pro","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.6-sol"},"grok-4.1":{"id":"grok-4.1","name":"grok-4.1","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":2,"output":10}},"claude-sonnet-4-5-20250929":{"id":"claude-sonnet-4-5-20250929","name":"claude-sonnet-4-5-20250929","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5}},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"claude-haiku-4-5-20251001","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"ministral-14b-2512":{"id":"ministral-14b-2512","name":"ministral-14b-2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.33,"output":0.33},"canonical_model_id":"mistral/ministral-14b-2512"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.075,"output":0.25,"cache_read":0.015},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9},"canonical_model_id":"google/gemini-3.5-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"doubao-seed-1-6-vision-250815":{"id":"doubao-seed-1-6-vision-250815","name":"doubao-seed-1-6-vision-250815","description":"ByteDance Seed multimodal model for image understanding, visual reasoning, and tool-assisted tasks","family":"seed","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-08-15","last_updated":"2025-08-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.114,"output":1.143},"canonical_model_id":"bytedance-seed/seed-1-6-vision"},"gpt-4.1":{"id":"gpt-4.1","name":"gpt-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8},"canonical_model_id":"openai/gpt-4.1"},"qwen3-coder-480b-a35b-instruct":{"id":"qwen3-coder-480b-a35b-instruct","name":"qwen3-coder-480b-a35b-instruct","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.86,"output":3.43,"tiers":[{"input":1.29,"output":5.15,"tier":{"type":"context","size":32000}},{"input":2.15,"output":8.58,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"qwen3.5-plus":{"id":"qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.12,"output":0.69,"tiers":[{"input":0.29,"output":1.72,"tier":{"type":"context","size":128000}},{"input":0.58,"output":3.44,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.5-plus"},"gpt-5":{"id":"gpt-5","name":"gpt-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10},"canonical_model_id":"openai/gpt-5"},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":1.8,"tiers":[{"input":1.2,"output":7.2,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.2,"output":7.2}},"canonical_model_id":"alibaba/qwen3.6-plus"},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5},"canonical_model_id":"xai/grok-4.3"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"gemini-3-flash-preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3-flash-preview"},"MiniMax-M2.1":{"id":"MiniMax-M2.1","name":"MiniMax-M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.1"},"glm-5":{"id":"glm-5","name":"glm-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.6,"tiers":[{"input":0.9,"output":3.2,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5"},"gemini-3.5-flash-thinking":{"id":"gemini-3.5-flash-thinking","name":"gemini-3.5-flash-thinking","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9},"canonical_model_id":"google/gemini-3.5-flash"},"deepseek-v3.2-thinking":{"id":"deepseek-v3.2-thinking","name":"DeepSeek-V3.2-Thinking","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.29,"output":0.43},"canonical_model_id":"deepseek/deepseek-v3.2"},"doubao-seed-1-6-thinking-250715":{"id":"doubao-seed-1-6-thinking-250715","name":"doubao-seed-1-6-thinking-250715","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":16000},"cost":{"input":0.121,"output":1.21}},"glm-4.6":{"id":"glm-4.6","name":"glm-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.286,"output":1.142,"tiers":[{"input":0.572,"output":2.29,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6"},"gpt-5.2-chat-latest":{"id":"gpt-5.2-chat-latest","name":"gpt-5.2-chat-latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.75,"output":14},"canonical_model_id":"openai/gpt-5.2-chat-latest"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"glm-5v-turbo":{"id":"glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":0.72,"output":3.2,"tiers":[{"input":1.1,"output":3.8,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5v-turbo"},"gpt-5.1-chat-latest":{"id":"gpt-5.1-chat-latest","name":"gpt-5.1-chat-latest","description":"Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.25,"output":10},"canonical_model_id":"openai/gpt-5.1-chat-latest"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"gemini-2.5-pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.25,"output":10,"tiers":[{"input":2.5,"output":15,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15}},"canonical_model_id":"google/gemini-2.5-pro"},"qwen3-max-2025-09-23":{"id":"qwen3-max-2025-09-23","name":"qwen3-max-2025-09-23","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":258048,"output":65536},"cost":{"input":0.86,"output":3.43,"tiers":[{"input":1.43,"output":5.72,"tier":{"type":"context","size":32000}},{"input":2.15,"output":8.58,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3-max"},"glm-4.6v":{"id":"glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.145,"output":0.43,"tiers":[{"input":0.29,"output":0.86,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6v"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"gpt-5.4-mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"gpt-5.6-terra-pro":{"id":"gpt-5.6-terra-pro","name":"gpt-5.6-terra-pro","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"openai/gpt-5.6-terra"},"gemini-3-pro-preview":{"id":"gemini-3-pro-preview","name":"gemini-3-pro-preview","description":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","family":"gemini-pro","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2,"output":12,"tiers":[{"input":4,"output":18,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18}},"canonical_model_id":"google/gemini-3-pro-preview"},"MiniMax-M1":{"id":"MiniMax-M1","name":"MiniMax-M1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-06-16","last_updated":"2025-06-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.132,"output":1.254}},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"gpt-5.4-nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.5"},"mistral-large-2512":{"id":"mistral-large-2512","name":"mistral-large-2512","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":262144},"cost":{"input":1.1,"output":3.3},"canonical_model_id":"mistral/mistral-large-2512"}}},"groq":{"id":"groq","env":["GROQ_API_KEY"],"npm":"@ai-sdk/groq","name":"Groq","doc":"https://console.groq.com/docs/models","models":{"whisper-large-v3-turbo":{"id":"whisper-large-v3-turbo","type":"transcription","name":"Whisper Large V3 Turbo","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-10-01","last_updated":"2024-10-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":448},"canonical_model_id":"openai/whisper-large-v3-turbo"},"allam-2-7b":{"id":"allam-2-7b","name":"ALLaM-2-7b","description":"ALLaM-2-7b instruction tuned model by SDAIA","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-01-23","last_updated":"2025-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"sdaia/allam-2-7b"},"whisper-large-v3":{"id":"whisper-large-v3","type":"transcription","name":"Whisper","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":448},"canonical_model_id":"openai/whisper-large-v3"},"llama-3.3-70b-versatile":{"id":"llama-3.3-70b-versatile","name":"Llama 3.3 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.59,"output":0.79},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"llama-3.1-8b-instant":{"id":"llama-3.1-8b-instant","name":"Llama 3.1 8B","description":"Compact Llama instruction model for fast chat and local deployment","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":0.05,"output":0.08},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"groq/compound":{"id":"groq/compound","name":"Compound","description":"General-purpose chat model for instruction following, writing, and analysis","family":"groq","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-09-04","last_updated":"2025-09-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated"},"groq/compound-mini":{"id":"groq/compound-mini","name":"Compound Mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"groq","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-09-04","last_updated":"2025-09-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-oss-safeguard-20b":{"id":"openai/gpt-oss-safeguard-20b","name":"Safety GPT OSS 20B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"status":"beta","cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-oss-120b"},"meta-llama/llama-prompt-guard-2-22m":{"id":"meta-llama/llama-prompt-guard-2-22m","name":"Llama Prompt Guard 2 22M","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":512},"status":"beta","cost":{"input":0.03,"output":0.03},"canonical_model_id":"meta/llama-prompt-guard-2-22m"},"meta-llama/llama-prompt-guard-2-86m":{"id":"meta-llama/llama-prompt-guard-2-86m","name":"Prompt Guard 2 86M","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":512},"status":"beta","cost":{"input":0.04,"output":0.04},"canonical_model_id":"meta/llama-prompt-guard-2-86m"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","default","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"beta","cost":{"input":0.8,"output":4},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","default"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.6,"output":3,"cache_read":0.3},"canonical_model_id":"alibaba/qwen3.6-27b"},"canopylabs/orpheus-v1-english":{"id":"canopylabs/orpheus-v1-english","type":"speech","name":"Canopy Labs Orpheus V1 English","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"canopylabs","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-19","last_updated":"2025-12-19","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":4000,"output":50000},"status":"beta","canonical_model_id":"canopylabs/orpheus-v1-english"},"canopylabs/orpheus-arabic-saudi":{"id":"canopylabs/orpheus-arabic-saudi","type":"speech","name":"Canopy Labs Orpheus Arabic Saudi","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"canopylabs","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":4000,"output":50000},"status":"beta","canonical_model_id":"canopylabs/orpheus-arabic-saudi"}}},"commandcode":{"id":"commandcode","env":["CMD_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.commandcode.ai/provider/v1","name":"CommandCode","doc":"https://commandcode.ai/docs/provider","models":{"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":2,"output":8,"cache_read":0.5,"cache_write":0},"canonical_model_id":"openai/gpt-5.3-codex"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":0},"canonical_model_id":"openai/gpt-5.5"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":0},"canonical_model_id":"openai/gpt-5.4"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":45,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075,"cache_write":0},"canonical_model_id":"openai/gpt-5.4-mini"},"meta/muse-wonder-1.0":{"id":"meta/muse-wonder-1.0","name":"Muse Wonder","description":"Muse Wonder is a frontier multimodal reasoning model from Meta for personal and professional agents, long-horizon coding, and computer use. It combines Active Memory, which carries relevant context across conversations, with Active Focus, which keeps it working through blockers toward a goal.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.3-contributor":{"id":"meta/muse-spark-1.3-contributor","name":"Muse Spark 1.3 (Contributor)","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.002},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"meta/muse-spark-1.2-contributor":{"id":"meta/muse-spark-1.2-contributor","name":"Muse Spark 1.2 (Contributor)","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.002},"canonical_model_id":"meta/muse-spark-1.2"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"sakana/fugu-ultra"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.3-flashx":{"id":"z-ai/glm-5.3-flashx","name":"GLM-5.3-FlashX","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.37,"output":1.25,"cache_read":0.075},"canonical_model_id":"zhipuai/glm-5.3-flash"},"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.7-Code-Highspeed":{"id":"moonshotai/Kimi-K2.7-Code-Highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"mistral/mistral-large-4":{"id":"mistral/mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":262144},"cost":{"input":1.36,"output":4.18,"cache_read":0.14},"canonical_model_id":"mistral/mistral-large-4"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMaxAI/MiniMax-M2.7":{"id":"MiniMaxAI/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"inclusionai/ling-3.1-flash:free":{"id":"inclusionai/ling-3.1-flash:free","name":"Ling 3.1 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"inclusionai/ling-3.1-flash"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":1000000},"cost":{"input":0.5,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4-flash-fast":{"id":"deepseek/deepseek-v4-flash-fast","name":"DeepSeek V4 Flash Fast","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.28,"output":0.56,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4.1-flash-fast":{"id":"deepseek/deepseek-v4.1-flash-fast","name":"DeepSeek V4.1 Flash Fast","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.16,"output":0.58,"cache_read":0.016},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"tencent/hy3-paid":{"id":"tencent/hy3-paid","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0.14,"output":0.58,"cache_read":0.035},"canonical_model_id":"tencent/hy3"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.2-Fast":{"id":"zai-org/GLM-5.2-Fast","name":"GLM-5.2-Fast","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3,"output":10.25,"cache_read":0.5},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"stepfun/Step-3.5-Flash":{"id":"stepfun/Step-3.5-Flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.09,"output":0.3,"cache_read":0.02},"canonical_model_id":"stepfun/step-3.5-flash"},"stepfun/Step-3.7-Flash":{"id":"stepfun/Step-3.7-Flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"stepfun/Step-5-Preview":{"id":"stepfun/Step-5-Preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":1000000},"cost":{"input":1,"output":2.7,"cache_read":0.05},"canonical_model_id":"stepfun/step-5-preview"},"xai/grok-4.5":{"id":"xai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.5"},"xai/grok-4.6":{"id":"xai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"xai/grok-4.7":{"id":"xai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"nvidia/nemotron-3-ultra-550b-a55b":{"id":"nvidia/nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"Qwen/Qwen3.8-Omni-Flash":{"id":"Qwen/Qwen3.8-Omni-Flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.4,"output":3,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3.8-Flash":{"id":"Qwen/Qwen3.8-Flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.16,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-flash"},"Qwen/Qwen3.8-Max":{"id":"Qwen/Qwen3.8-Max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"Qwen/Qwen3.8-Max-0902":{"id":"Qwen/Qwen3.8-Max-0902","name":"Qwen3.8 Max 0902","description":"2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"google/gemini-4-argon"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15,"cache_write":0.08334},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"google/gemini-3.5-flash"}}},"ainetcafe":{"id":"ainetcafe","env":["AINETCAFE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://microquickjs.com/v1","name":"ainetcafe","doc":"https://ainetcafe.com/k3/guides/","models":{"Kimi-K3":{"id":"Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":2.1,"output":10.5,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"}}},"liquid":{"id":"liquid","env":["LIQUID_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://router.inference.ai.exchange/v1","name":"Liquid Inference","doc":"https://inference.ai.exchange/docs","models":{"liquid.auto":{"id":"liquid.auto","name":"Liquid Auto","description":"Routes each request to the lowest-priced eligible offer on the Liquid Inference exchange","family":"auto","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-20","last_updated":"2026-10-20","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":10000000,"output":1048576}}}},"cerebras":{"id":"cerebras","env":["CEREBRAS_API_KEY"],"npm":"@ai-sdk/cerebras","name":"Cerebras","doc":"https://inference-docs.cerebras.ai/models/overview","models":{"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":40960},"cost":{"input":0.35,"output":0.75,"cache_read":0.35},"canonical_model_id":"openai/gpt-oss-120b"},"qwen-3.8-27b":{"id":"qwen-3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":40960},"cost":{"input":0.99,"output":1.49,"cache_read":0.99},"canonical_model_id":"alibaba/qwen3.8-27b"}}},"tempr":{"id":"tempr","env":["TEMPR_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.temprhq.io/v1","name":"Tempr Gateway","doc":"https://temprhq.io/docs/gateway-reference.html","models":{"mistral/mistral-medium-latest":{"id":"mistral/mistral-medium-latest","name":"Mistral Medium (latest)","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"mistral/mistral-medium-latest"},"mistral/zai-glm-5-3":{"id":"mistral/zai-glm-5-3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.3"},"mistral/mistral-small-latest":{"id":"mistral/mistral-small-latest","name":"Mistral Small (latest)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"mistral/mistral-small-latest"},"mistral/mistral-medium-2604":{"id":"mistral/mistral-medium-2604","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"mistral/mistral-medium-2604"},"mistral/mistral-large-latest":{"id":"mistral/mistral-large-latest","name":"Mistral Large (latest)","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"mistral/mistral-large-latest"},"mistral/mistral-small-2603":{"id":"mistral/mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"mistral/mistral-small-2603"},"mistral/zai-glm-5-2":{"id":"mistral/zai-glm-5-2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"mistral/voxtral-small-latest":{"id":"mistral/voxtral-small-latest","name":"Voxtral Small (latest)","description":"Instruct model with native audio input for speech understanding and tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000},"cost":{"input":0.1,"output":0.4},"canonical_model_id":"mistral/voxtral-small-latest"},"mistral/mistral-embed":{"id":"mistral/mistral-embed","type":"embedding","name":"Mistral Embed","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"mistral-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-12-11","last_updated":"2023-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":3072},"cost":{"input":0.1,"output":0},"canonical_model_id":"mistral/mistral-embed"},"mistral/mistral-large-2512":{"id":"mistral/mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"mistral/mistral-large-2512"},"groq/allam-2-7b":{"id":"groq/allam-2-7b","name":"ALLaM-2-7b","description":"ALLaM-2-7b instruction tuned model by SDAIA","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-01-23","last_updated":"2025-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"sdaia/allam-2-7b"},"groq/openai/gpt-oss-20b":{"id":"groq/openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-oss-20b"},"groq/openai/gpt-oss-120b":{"id":"groq/openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-oss-120b"},"groq/qwen/qwen3.8-27b":{"id":"groq/qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131042,"output":16384},"cost":{"input":0.8,"output":4,"cache_read":0.4},"canonical_model_id":"alibaba/qwen3.8-27b"},"cerebras/gpt-oss-120b":{"id":"cerebras/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":40960},"cost":{"input":0.35,"output":0.75},"canonical_model_id":"openai/gpt-oss-120b"},"cerebras/qwen-3.8-27b":{"id":"cerebras/qwen-3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":40960},"cost":{"input":0.99,"output":1.49},"canonical_model_id":"alibaba/qwen3.8-27b"},"deepseek/deepseek-flash":{"id":"deepseek/deepseek-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"reasoning":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"reasoning":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"anthropic/claude-opus-4-8":{"id":"anthropic/claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-haiku-4-5":{"id":"anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-sonnet-4-5":{"id":"anthropic/claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-fable-5-1":{"id":"anthropic/claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-opus-4-5-20251101":{"id":"anthropic/claude-opus-4-5-20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-sonnet-4-5-20250929":{"id":"anthropic/claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"anthropic/claude-haiku-4-5-20251001":{"id":"anthropic/claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"anthropic/claude-opus-4-5":{"id":"anthropic/claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"moonshot/kimi-k2.6":{"id":"moonshot/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshot/kimi-k2.7-code-highspeed":{"id":"moonshot/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshot/kimi-k2.7-code":{"id":"moonshot/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshot/kimi-k3":{"id":"moonshot/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3},"canonical_model_id":"moonshotai/kimi-k3"},"minimax/MiniMax-M3":{"id":"minimax/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"tiers":[{"input":0.6,"output":2.4,"cache_read":0.12,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":0.6,"output":2.4,"cache_read":0.12}},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/MiniMax-M2.7":{"id":"minimax/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/MiniMax-M2.7-highspeed":{"id":"minimax/MiniMax-M2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"xai/grok-4.5":{"id":"xai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"xai/grok-4.6":{"id":"xai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"xai/grok-4.20-0309-non-reasoning":{"id":"xai/grok-4.20-0309-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"xai/grok-4.7":{"id":"xai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"xai/grok-4.3":{"id":"xai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"xai/grok-4.20-0309-reasoning":{"id":"xai/grok-4.20-0309-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"xai/grok-build-0.1":{"id":"xai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}},"canonical_model_id":"xai/grok-build-0.1"},"zai/glm-4.5-air":{"id":"zai/glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.2,"output":1.1,"cache_read":0.03,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-air"},"zai/glm-4.7-flashx":{"id":"zai/glm-4.7-flashx","name":"GLM-4.7-FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.07,"output":0.4,"cache_read":0.01,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"zai/glm-4.6v-flash":{"id":"zai/glm-4.6v-flash","name":"GLM-4.6V-Flash","description":"Lightweight GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6v-flash"},"zai/glm-4.5v":{"id":"zai/glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":16384},"cost":{"input":0.6,"output":1.8,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5v"},"zai/glm-5.1":{"id":"zai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"zai/glm-4.7":{"id":"zai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7"},"zai/glm-5-turbo":{"id":"zai/glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24,"cache_write":0},"canonical_model_id":"zhipuai/glm-5-turbo"},"zai/glm-4.5":{"id":"zai/glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5"},"zai/glm-5.3-flash":{"id":"zai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai/glm-5.2":{"id":"zai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"zai/glm-4.7-flash":{"id":"zai/glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"zai/glm-4.5-flash":{"id":"zai/glm-4.5-flash","name":"GLM-4.5-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":98304},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-flash"},"zai/glm-5.3-flashx":{"id":"zai/glm-5.3-flashx","name":"GLM-5.3-FlashX","description":"High-speed GLM-5.3-Flash serving option for coding and agent workflows","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.37,"output":1.25,"cache_read":0.075,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai/glm-5":{"id":"zai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2,"cache_write":0},"canonical_model_id":"zhipuai/glm-5"},"zai/glm-4.6":{"id":"zai/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6"},"zai/glm-5.3":{"id":"zai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"zai/glm-5v-turbo":{"id":"zai/glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24,"cache_write":0},"canonical_model_id":"zhipuai/glm-5v-turbo"},"zai/glm-4.6v":{"id":"zai/glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.05},"canonical_model_id":"zhipuai/glm-4.6v"},"xiaomi-mimo/mimo-v2.6-pro-ultraspeed":{"id":"xiaomi-mimo/mimo-v2.6-pro-ultraspeed","name":"MiMo-V2.6-Pro-UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.35,"output":8.7,"cache_read":0.036},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"xiaomi-mimo/mimo-v2.6-pro":{"id":"xiaomi-mimo/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"xiaomi-mimo/mimo-v2.6-flash":{"id":"xiaomi-mimo/mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemini-embedding-001":{"id":"google/gemini-embedding-001","type":"embedding","name":"Gemini Embedding 001","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-05","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2048,"output":1},"cost":{"input":0.15,"output":0},"canonical_model_id":"google/gemini-embedding-001"},"google/gemini-flash-latest":{"id":"google/gemini-flash-latest","name":"Gemini Flash Latest","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-flash-latest"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-flash-lite-latest":{"id":"google/gemini-flash-lite-latest","name":"Gemini Flash-Lite Latest","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-flash-lite-latest"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-embedding-2":{"id":"google/gemini-embedding-2","type":"embedding","name":"Gemini Embedding 2","description":"Multimodal embedding model mapping text, images, video, audio, and PDFs into a unified embedding space","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-11","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1},"cost":{"input":0.2,"output":0},"canonical_model_id":"google/gemini-embedding-2"}}},"klokintegration":{"id":"klokintegration","env":["KLOKINTEGRATION_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api-gw.klok.ipaas.se/proxy/kloker-key/v1","name":"klokintegration.se","doc":"https://klokintegration.se/docs/ai-api","models":{"Kloker-Integration-Architect":{"id":"Kloker-Integration-Architect","name":"Kloker Integration Architect","description":"Knows the customer integration environment and Klok best practices. Opinionated about structure. The gateway runs ecosystem lookup tools server-side and appends a system-prompt injection (data contracts, CloudEvents, event-driven flows). Client system prompts and OpenAI tool calls are preserved.","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-29","last_updated":"2026-08-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":50000},"status":"beta","cost":{"input":0.23,"output":1.16}},"Kloker":{"id":"Kloker","name":"Kloker","description":"Cheap general model with a clean context. Nothing from the customer environment is packed in. It tracks the current best open source model. The Klok team verifies it and upgrades it periodically.","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-29","last_updated":"2026-08-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":50000},"status":"beta","cost":{"input":0.23,"output":1.16}},"Kloker-Integration-Developer":{"id":"Kloker-Integration-Developer","name":"Kloker Integration Developer","description":"Knows the customer integration environment and Klok best practices. Opinionated about implementation. The gateway runs ecosystem lookup tools server-side and appends a system-prompt injection. Client system prompts and OpenAI tool calls are preserved.","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-29","last_updated":"2026-08-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":50000},"status":"beta","cost":{"input":0.23,"output":1.16}}}},"tensorx":{"id":"tensorx","env":["TENSORX_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.tensorx.ai/v1","name":"TensorX","doc":"https://docs.tensorx.ai/","models":{"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":202752},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":0.35,"cache_write":1.75},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"status":"deprecated","cost":{"input":1.2,"output":4,"cache_read":0.3,"cache_write":1.5},"canonical_model_id":"zhipuai/glm-5-turbo"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.5,"output":4.5,"cache_read":0.375},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":202752},"status":"deprecated","cost":{"input":1,"output":3.2,"cache_read":0.25,"cache_write":1.25},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":1.75,"output":4.5,"cache_read":0.44},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-5v-turbo":{"id":"z-ai/glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"status":"deprecated","cost":{"input":1.2,"output":4,"cache_read":0.3,"cache_write":1.5},"canonical_model_id":"zhipuai/glm-5v-turbo"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":1,"output":4,"cache_read":0.25,"cache_write":1.25},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.25,"output":4.5,"cache_read":0.3125},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.75},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.5,"output":2.8,"cache_read":0.125,"cache_write":0.625},"canonical_model_id":"moonshotai/kimi-k2.5"},"deepseek/deepseek-r1-0528":{"id":"deepseek/deepseek-r1-0528","name":"DeepSeek R1-0528","description":"May 2025 DeepSeek R1 update with improved mathematics, coding, and reasoning","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":8192},"status":"deprecated","cost":{"input":0.66,"output":2.6,"cache_read":0.165,"cache_write":0.825},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"status":"deprecated","cost":{"input":0.3,"output":0.5,"cache_read":0.075,"cache_write":0.375},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":2,"output":4,"cache_read":0.5},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"status":"deprecated","cost":{"input":1.75,"output":3.5,"cache_read":0.4375,"cache_write":2.185},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.5,"output":1.5,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.25,"output":0.3,"cache_read":0.06},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.4,"output":2,"cache_read":0.1},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":65536},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.075,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.4,"output":2.4,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.5,"output":3.5,"cache_read":0.125,"cache_write":0.625},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3.8-2.4t-a95b":{"id":"qwen/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":64000},"cost":{"input":2.5,"output":6,"cache_read":0.63},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"qwen/qwen3.8-flash-next":{"id":"qwen/qwen3.8-flash-next","name":"Qwen3.8 Flash Next","description":"Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":64000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.8-flash-next"},"qwen/qwen3-235b-a22b-2507":{"id":"qwen/qwen3-235b-a22b-2507","name":"Qwen3 235B-A22B-2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06-30","release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":262144},"status":"deprecated","cost":{"input":0.072,"output":0.464,"cache_read":0.018,"cache_write":0.09},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"qwen/qwen3.5-9b":{"id":"qwen/qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.15,"output":0.2,"cache_read":0.0375,"cache_write":0.1875},"canonical_model_id":"alibaba/qwen3.5-9b"}}},"poolside":{"id":"poolside","env":["POOLSIDE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.poolside.ai/v1","name":"Poolside","doc":"https://platform.poolside.ai","models":{"poolside/laguna-xs-2.1":{"id":"poolside/laguna-xs-2.1","name":"Laguna XS 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-07-02","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"poolside/laguna-xs-2.1"},"poolside/laguna-s-2.1":{"id":"poolside/laguna-s-2.1","name":"Laguna S 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"poolside/laguna-s-2.1"},"poolside/laguna-m.1":{"id":"poolside/laguna-m.1","name":"Laguna M.1","description":"Poolside's open-weight model for agentic coding and long-horizon work","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"poolside/laguna-m.1"}}},"ambient":{"id":"ambient","env":["AMBIENT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.ambient.xyz/v1","name":"Ambient","doc":"https://ambient.xyz","models":{"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":202752},"cost":{"input":0.6,"output":2,"cache_read":0.15,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.95,"output":4,"cache_read":0.2,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.69,"output":3.49,"cache_read":0.14,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"ambient/small":{"id":"ambient/small","name":"Ambient Small","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":101072,"output":101072},"cost":{"input":0.001,"output":0.001,"cache_read":0,"cache_write":0}},"ambient/large":{"id":"ambient/large","name":"Ambient Large","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":202752},"cost":{"input":0.6,"output":2,"cache_read":0.15,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"unsloth/qwen3.8-27b-gguf":{"id":"unsloth/qwen3.8-27b-gguf","name":"qwen3.8-27b-4q","description":"Tool-capable chat model for instruction following and agentic application workflows","family":"unsloth","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":101072,"output":101072},"cost":{"input":0.001,"output":0.001,"cache_read":0,"cache_write":0}},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.028,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"status":"deprecated","cost":{"input":0.08,"output":0.18,"cache_read":0.016,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo-V2.5","description":"MiMo omni model for text, image, video, audio, and agents","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.4,"output":2,"cache_read":0.08,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"zai-org/GLM-5.1-FP8":{"id":"zai-org/GLM-5.1-FP8","name":"GLM 5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-5.2-FP8":{"id":"zai-org/GLM-5.2-FP8","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":202752},"status":"deprecated","cost":{"input":1.2,"output":4.2,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"stepfun/step-3.7-flash":{"id":"stepfun/step-3.7-flash","name":"Step 3.7 Flash","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.19,"output":1.14,"cache_read":0.03,"cache_write":0},"canonical_model_id":"stepfun/step-3.7-flash"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.23,"output":0.96,"cache_read":0.05,"cache_write":0},"canonical_model_id":"minimax/MiniMax-M3"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0.32,"output":3.2,"cache_read":0.16,"cache_write":0},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.6-35b-a3b":{"id":"qwen/qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.15,"output":1,"cache_read":0.15,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3-30b-a3b-instruct-2507":{"id":"qwen/qwen3-30b-a3b-instruct-2507","name":"Qwen3 30B-A3B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.09,"output":0.3,"cache_read":0.09,"cache_write":0},"canonical_model_id":"alibaba/qwen3-30b-a3b-instruct-2507"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0.32,"output":3.2,"cache_read":0.16,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-27b"}}},"thinkingmachines":{"id":"thinkingmachines","env":["TINKER_API_KEY"],"npm":"@ai-sdk/anthropic","api":"https://tinker.thinkingmachines.dev/services/tinker-prod/anthropic/api/v1","name":"Thinking Machines","doc":"https://tinker-docs.thinkingmachines.ai/tinker/compatible-apis/anthropic/","models":{"moonshotai/Kimi-K2.6:peft:131072":{"id":"moonshotai/Kimi-K2.6:peft:131072","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":2.205,"output":5.49,"cache_read":0.441},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":2.205,"output":5.49,"cache_read":0.441},"canonical_model_id":"moonshotai/kimi-k2.6"},"thinkingmachines/Inkling":{"id":"thinkingmachines/Inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":1.87,"output":4.68,"cache_read":0.374},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/Inkling:peft:262144":{"id":"thinkingmachines/Inkling:peft:262144","name":"Inkling (256K)","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.87,"output":4.68,"cache_read":0.374},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/Inkling-Small:peft:262144":{"id":"thinkingmachines/Inkling-Small:peft:262144","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.58,"output":1.44,"cache_read":0.116},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/Inkling-Small:peft:262144:sampling-nvfp4":{"id":"thinkingmachines/Inkling-Small:peft:262144:sampling-nvfp4","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/Inkling-Small":{"id":"thinkingmachines/Inkling-Small","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.58,"output":1.44,"cache_read":0.116},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/Inkling:peft:262144:sampling-nvfp4":{"id":"thinkingmachines/Inkling:peft:262144:sampling-nvfp4","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.18,"output":0.45,"cache_read":0.036},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-oss-120b:peft:131072":{"id":"openai/gpt-oss-120b:peft:131072","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.33,"output":0.84,"cache_read":0.066},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.33,"output":0.84,"cache_read":0.066},"canonical_model_id":"openai/gpt-oss-120b"},"zai-org/GLM-5.3-Flash:peft:262144":{"id":"zai-org/GLM-5.3-Flash:peft:262144","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.84,"output":2.11,"cache_read":0.168},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai-org/GLM-5.3:peft:262144":{"id":"zai-org/GLM-5.3:peft:262144","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":3.78,"output":12.15,"cache_read":0.756},"canonical_model_id":"zhipuai/glm-5.3"},"nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16":{"id":"nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16","name":"Nemotron 3.5 Super VL 120B A12B","description":"NVIDIA vision-language reasoning model for text, image, and video understanding and agentic tool use","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.57,"output":1.44,"cache_read":0.114},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16":{"id":"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":1.8,"output":6.225,"cache_read":0.36},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16":{"id":"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.195,"output":0.495,"cache_read":0.039},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16:peft:262144":{"id":"nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16:peft:262144","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.57,"output":1.44,"cache_read":0.114},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16:peft:262144":{"id":"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16:peft:262144","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.8,"output":6.225,"cache_read":0.36},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16:peft:262144":{"id":"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16:peft:262144","name":"Nemotron 3.5 Lightning 30B A3B","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.184,"output":0.495,"cache_read":0.0368},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16":{"id":"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16","name":"Nemotron 3.5 Lightning 30B A3B","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.184,"output":0.495,"cache_read":0.0368},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16:peft:262144":{"id":"nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16:peft:262144","name":"Nemotron 3.5 Super VL 120B A12B","description":"NVIDIA vision-language reasoning model for text, image, and video understanding and agentic tool use","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.57,"output":1.44,"cache_read":0.114},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16":{"id":"nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.57,"output":1.44,"cache_read":0.114},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"deepseek-ai/DeepSeek-V4.1-Flash:peft:262144":{"id":"deepseek-ai/DeepSeek-V4.1-Flash:peft:262144","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3.6,"cache_read":0.12},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-ai/DeepSeek-V4.1-Flash-1017:peft:262144":{"id":"deepseek-ai/DeepSeek-V4.1-Flash-1017:peft:262144","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3.6,"cache_read":0.12},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek-V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":1.695,"output":4.215,"cache_read":0.339},"canonical_model_id":"deepseek/deepseek-v3.1"},"Qwen/Qwen3.6-27B":{"id":"Qwen/Qwen3.6-27B","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":1.86,"output":5.595,"cache_read":0.372},"canonical_model_id":"alibaba/qwen3.6-27b"},"Qwen/Qwen3.6-35B-A3B":{"id":"Qwen/Qwen3.6-35B-A3B","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.45,"output":1.335,"cache_read":0.09},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen/Qwen3.5-4B":{"id":"Qwen/Qwen3.5-4B","name":"Qwen3.5 4B","description":"Small Qwen3.5 multimodal model for low-cost chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-03-02","last_updated":"2026-03-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.33,"output":1.005,"cache_read":0.066},"canonical_model_id":"alibaba/qwen3.5-4b"},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.6975,"output":5.595,"cache_read":0.1395},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3.5-397B-A17B":{"id":"Qwen/Qwen3.5-397B-A17B","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":1.65,"output":7.5,"cache_read":0.33},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"Qwen/Qwen3.5-397B-A17B:peft:262144":{"id":"Qwen/Qwen3.5-397B-A17B:peft:262144","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.65,"output":7.5,"cache_read":0.33},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"Qwen/Qwen3.5-35B-A3B-Base":{"id":"Qwen/Qwen3.5-35B-A3B-Base","name":"Qwen3.5 35B-A3B Base","description":"Pre-trained Qwen3.5 base checkpoint for fine-tuning and research, without instruction tuning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.54,"output":1.335,"cache_read":0.108},"canonical_model_id":"alibaba/qwen3.5-35b-a3b-base"},"Qwen/Qwen3.8-27B:peft:262144":{"id":"Qwen/Qwen3.8-27B:peft:262144","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6975,"output":5.595,"cache_read":0.1395},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3.5-9B-Base":{"id":"Qwen/Qwen3.5-9B-Base","name":"Qwen3.5 9B Base","description":"Pre-trained Qwen3.5 base checkpoint for fine-tuning and research, without instruction tuning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.3,"output":0.807,"cache_read":0.06},"canonical_model_id":"alibaba/qwen3.5-9b-base"},"Qwen/Qwen3.6-35B-A3B:peft:262144":{"id":"Qwen/Qwen3.6-35B-A3B:peft:262144","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.45,"output":1.335,"cache_read":0.09},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen/Qwen3.5-9B":{"id":"Qwen/Qwen3.5-9B","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.3,"output":0.807,"cache_read":0.06},"canonical_model_id":"alibaba/qwen3.5-9b"},"Qwen/Qwen3-8B":{"id":"Qwen/Qwen3-8B","name":"Qwen3 8B","description":"Dense 8B Qwen3 model with switchable thinking for chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.195,"output":0.6,"cache_read":0.039},"canonical_model_id":"alibaba/qwen3-8b"},"Qwen/Qwen3.5-4B:peft:262144":{"id":"Qwen/Qwen3.5-4B:peft:262144","name":"Qwen3.5 4B","description":"Small Qwen3.5 multimodal model for low-cost chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-03-02","last_updated":"2026-03-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.33,"output":1.005,"cache_read":0.066},"canonical_model_id":"alibaba/qwen3.5-4b"}}},"auriko":{"id":"auriko","env":["AURIKO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.auriko.ai/v1","name":"Auriko","doc":"https://docs.auriko.ai","models":{"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-2.5-flash"},"minimax-m2-7":{"id":"minimax-m2-7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax-m2-7-highspeed":{"id":"minimax-m2-7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.45,"output":2.25,"cache_read":0.07},"canonical_model_id":"moonshotai/kimi-k2.5"},"qwen-3.6-plus":{"id":"qwen-3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.1,"tiers":[{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5}},"canonical_model_id":"alibaba/qwen3.6-plus"},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"}}},"inceptron":{"id":"inceptron","env":["INCEPTRON_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.inceptron.io/v1","name":"Inceptron","doc":"https://docs.inceptron.io","models":{"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6712,"output":3.35,"cache_read":0.18,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.4946,"output":3.2,"cache_read":0.1509,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.69,"output":2.52,"cache_read":0.14,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.55,"output":1.96,"cache_read":0.18,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.15,"output":0.69,"cache_read":0.075,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.11,"output":0.3666,"cache_read":0.027,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"}}},"requesty":{"id":"requesty","env":["REQUESTY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://router.requesty.ai/v1","name":"Requesty","doc":"https://requesty.ai/solution/llm-routing/models","models":{"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.8-flash"},"qwen3.8-flash-next@eu":{"id":"qwen3.8-flash-next@eu","name":"Qwen3.8 Flash Next (EU)","description":"Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.8-flash-next"},"qwen3.5-35b-a3b":{"id":"qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.14,"output":1,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"mistral-medium-3-5@eu":{"id":"mistral-medium-3-5@eu","name":"Mistral Medium 3.5 (EU)","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":209715},"cost":{"input":1.65,"output":8.25,"cache_read":1.65},"canonical_model_id":"mistral/mistral-medium-3-5"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.1-flash-lite"},"seed-2.0-pro":{"id":"seed-2.0-pro","name":"Seed 2.0 Pro","description":"Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.5,"output":3,"cache_read":0.1},"canonical_model_id":"bytedance-seed/seed-2.0-pro"},"gpt-5.1@eu":{"id":"gpt-5.1@eu","name":"GPT-5.1 (EU)","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.375,"output":11,"cache_read":0.1375},"canonical_model_id":"openai/gpt-5.1"},"mistral-medium-latest":{"id":"mistral-medium-latest","name":"Mistral Medium (latest)","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.44,"output":2.2,"cache_read":0.44},"canonical_model_id":"mistral/mistral-medium-latest"},"nvidia-nemotron-3-super-120b-a12b":{"id":"nvidia-nemotron-3-super-120b-a12b","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.5},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.7-max"},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"cache_write":2,"tiers":[{"input":4,"output":12,"cache_read":1,"cache_write":4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1,"cache_write":4}},"canonical_model_id":"xai/grok-4.5"},"gpt-6-luna@eu":{"id":"gpt-6-luna@eu","name":"GPT-6 Luna (EU)","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.12,"output":0.6,"cache_read":0.012,"cache_write":0.15,"tiers":[{"input":0.24,"output":0.9,"cache_read":0.024,"cache_write":0.3,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.24,"output":0.9,"cache_read":0.024,"cache_write":0.3}},"canonical_model_id":"openai/gpt-6-luna"},"muse-wonder-1.0":{"id":"muse-wonder-1.0","name":"Muse Wonder","description":"Muse Wonder is a frontier multimodal reasoning model from Meta for personal and professional agents, long-horizon coding, and computer use. It combines Active Memory, which carries relevant context across conversations, with Active Focus, which keeps it working through blockers toward a goal.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"minimax-m3@eu":{"id":"minimax-m3@eu","name":"MiniMax-M3 (EU)","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.4,"output":2,"cache_read":0.1},"canonical_model_id":"minimax/MiniMax-M3"},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.32,"output":1.28,"cache_read":0.032,"cache_write":0.4},"canonical_model_id":"alibaba/qwen3.7-plus"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"cache_write":2,"tiers":[{"input":4,"output":12,"cache_read":1,"cache_write":4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1,"cache_write":4}},"canonical_model_id":"xai/grok-4.6"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"gpt-5.5-pro":{"id":"gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.5-pro"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"google/gemini-4-argon"},"glm-5.1@eu":{"id":"glm-5.1@eu","name":"GLM-5.1 (EU)","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":1.4},"canonical_model_id":"zhipuai/glm-5.1"},"gpt-4.1-mini@eu":{"id":"gpt-4.1-mini@eu","name":"GPT-4.1 mini (EU)","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.44,"output":1.76,"cache_read":0.11},"canonical_model_id":"openai/gpt-4.1-mini"},"o4-mini@eu":{"id":"o4-mini@eu","name":"o4-mini (EU)","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.21,"output":4.84,"cache_read":0.3025},"canonical_model_id":"openai/o4-mini"},"gpt-6-sol@eu":{"id":"gpt-6-sol@eu","name":"GPT-6 Sol (EU)","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.4,"output":12,"cache_read":0.24,"cache_write":3,"tiers":[{"input":4.8,"output":18,"cache_read":0.48,"cache_write":6,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.8,"output":18,"cache_read":0.48,"cache_write":6}},"canonical_model_id":"openai/gpt-6-sol"},"gpt-5@eu":{"id":"gpt-5@eu","name":"GPT-5 (EU)","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.375,"output":11,"cache_read":0.1375},"canonical_model_id":"openai/gpt-5"},"step-3.7-flash":{"id":"step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"gemini-2.5-flash-lite@eu":{"id":"gemini-2.5-flash-lite@eu","name":"Gemini 2.5 Flash-Lite (EU)","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":0.18333},"canonical_model_id":"google/gemini-2.5-flash-lite"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"nvidia-nemotron-3-ultra":{"id":"nvidia-nemotron-3-ultra","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.5,"output":2.5},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nemotron-3-super-120b-a12b":{"id":"nemotron-3-super-120b-a12b","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"glm-5.3-flash@eu":{"id":"glm-5.3-flash@eu","name":"GLM-5.3-Flash (EU)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":262144},"cost":{"input":0.2,"output":0.6,"cache_read":0.07},"canonical_model_id":"zhipuai/glm-5.3-flash"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"ling-2.6-flash":{"id":"ling-2.6-flash","name":"ling-2.6-flash","description":"Inclusion AI ling-2.6-flash","family":"ling","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.3}},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":1.2},"canonical_model_id":"minimax/MiniMax-M2.7"},"mistral-large-4@eu":{"id":"mistral-large-4@eu","name":"Mistral Large 4 (EU)","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":262144},"cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.5"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"muse-spark-1.3":{"id":"muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"gemini-3.8-flash@eu":{"id":"gemini-3.8-flash@eu","name":"Gemini 3.8 Flash (EU)","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.825,"output":4.125,"cache_read":0.0825},"canonical_model_id":"google/gemini-3.8-flash"},"kimi-k3@eu":{"id":"kimi-k3@eu","name":"Kimi K3 (EU)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":3,"output":15,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"laguna-xs.2":{"id":"laguna-xs.2","name":"Laguna XS.2","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-xs.2"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.13,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.63,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"leanstral-1-5":{"id":"leanstral-1-5","name":"leanstral-1-5","description":"Leanstral 1.5 is an updated Lean 4 formal proof engineering model from Mistral AI, optimized for automated theorem proving and autoformalization. It has 119B total parameters with 6.5B active and supports a 256K token context window. It supports native function calling and structured output.","family":"leanstral","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-27","last_updated":"2026-05-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0}},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"gpt-5.4-pro":{"id":"gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"cache_read":30},"canonical_model_id":"openai/gpt-5.4-pro"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.28,"output":0.56,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-flash"},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"cache_write":2},"canonical_model_id":"xai/grok-4.7"},"nemotron-3-ultra-550b-a55b":{"id":"nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"mistral-medium-latest@eu":{"id":"mistral-medium-latest@eu","name":"Mistral Medium (latest) (EU)","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.44,"output":2.2,"cache_read":0.44},"canonical_model_id":"mistral/mistral-medium-latest"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"mistral-large-4":{"id":"mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":262144},"cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"claude-opus-4-8@eu":{"id":"claude-opus-4-8@eu","name":"Claude Opus 4.8 (EU)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-8"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5.5,"output":33,"cache_read":0.55},"canonical_model_id":"openai/gpt-5.5"},"claude-opus-4-1":{"id":"claude-opus-4-1","name":"Claude Opus 4.1 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"nemotron-3-nano-omni-30b-a3b-reasoning":{"id":"nemotron-3-nano-omni-30b-a3b-reasoning","name":"Nemotron 3 Nano Omni 30B A3B Reasoning","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":20480},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":1,"output":2.7,"cache_read":0.05},"canonical_model_id":"stepfun/step-5-preview"},"deepseek-v4-pro@eu":{"id":"deepseek-v4-pro@eu","name":"DeepSeek V4 Pro (EU)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.75,"output":3.5,"cache_read":0.44},"canonical_model_id":"deepseek/deepseek-v4-pro"},"claude-sonnet-4-6@eu":{"id":"claude-sonnet-4-6@eu","name":"Claude Sonnet 4.6 (EU)","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3.3,"output":16.5,"cache_read":0.3,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.43,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0027},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"claude-haiku-4-5@eu":{"id":"claude-haiku-4-5@eu","name":"Claude Haiku 4.5 (latest) (EU)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5"},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"gpt-4.1-nano@eu":{"id":"gpt-4.1-nano@eu","name":"GPT-4.1 nano (EU)","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.11,"output":0.44,"cache_read":0.0275},"canonical_model_id":"openai/gpt-4.1-nano"},"seed-1.8":{"id":"seed-1.8","name":"Seed 1.8","description":"ByteDance Seed model for multimodal reasoning, long-context analysis, and agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-28","last_updated":"2025-12-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":0.25,"output":2,"cache_read":0.05},"canonical_model_id":"bytedance-seed/seed-1-8"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.75,"output":16.5,"cache_read":0.275},"canonical_model_id":"openai/gpt-5.4"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"nemotron-lightning-3.5-30b-a3b":{"id":"nemotron-lightning-3.5-30b-a3b","name":"nemotron-lightning-3.5-30b-a3b","description":"Nemotron-Lightning-3.5-30B-A3B is a 30B-parameter Mixture-of-Experts language model (3B active) from NVIDIA's Nemotron-H family, built on a hybrid Mamba-Transformer architecture for efficient long-context inference. Like other models in the family, it responds to queries by first generating a reasoning trace and then concluding with a final response, with reasoning behavior configurable through a flag in the chat template. It includes a multi-token prediction (MTP) speculative decoding head for low-latency serving.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-08-15","last_updated":"2026-08-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.05,"output":0.2,"cache_read":0.01}},"gpt-5.5@eu":{"id":"gpt-5.5@eu","name":"GPT-5.5 (EU)","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5.5,"output":33,"cache_read":0.55,"tiers":[{"input":11,"output":49.5,"cache_read":1.1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":11,"output":49.5,"cache_read":1.1}},"canonical_model_id":"openai/gpt-5.5"},"claude-sonnet-4@eu":{"id":"claude-sonnet-4@eu","name":"Claude Sonnet 4 (latest) (EU)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-0"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro"},"gemini-2.5-flash@eu":{"id":"gemini-2.5-flash@eu","name":"Gemini 2.5 Flash (EU)","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.3,"output":2.5,"cache_read":0.075,"cache_write":0.55},"canonical_model_id":"google/gemini-2.5-flash"},"claude-opus-5-5@eu":{"id":"claude-opus-5-5@eu","name":"Claude Opus 5.5 (EU)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.22,"cache_write":5.5},"canonical_model_id":"anthropic/claude-opus-5-5"},"nemotron-3-nano-omni":{"id":"nemotron-3-nano-omni","name":"nemotron-3-nano-omni","description":"The most open, efficient, and accurate omni modal reasoning model for agentic AI.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-20","last_updated":"2026-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":300000},"cost":{"input":0.06,"output":0.24,"cache_read":0.06}},"gemini-4-argon@eu":{"id":"gemini-4-argon@eu","name":"Gemini 4 Argon (EU)","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2.2,"output":11,"cache_read":0.11},"canonical_model_id":"google/gemini-4-argon"},"claude-fable-5.1@eu":{"id":"claude-fable-5.1@eu","name":"Claude Fable 5.1 (EU)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":0.275,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5-1"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5"},"gemini-3.1-flash-lite@eu":{"id":"gemini-3.1-flash-lite@eu","name":"Gemini 3.1 Flash Lite (EU)","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.275,"output":1.65,"cache_read":0.0275,"cache_write":0.091663},"canonical_model_id":"google/gemini-3.1-flash-lite"},"inkling-256k":{"id":"inkling-256k","name":"Inkling 256K","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":1.87,"output":4.68,"cache_read":0.374},"canonical_model_id":"thinkingmachines/inkling"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"qwen3.5-2b":{"id":"qwen3.5-2b","name":"Qwen3.5 2B","description":"Compact Qwen3.5 multimodal model for fast chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-02","last_updated":"2026-03-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":81920},"cost":{"input":0.02,"output":0.1},"canonical_model_id":"alibaba/qwen3.5-2b"},"claude-fable-5@eu":{"id":"claude-fable-5@eu","name":"Claude Fable 5 (EU)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5"},"nemotron-3.5-lightning-30b-a3b":{"id":"nemotron-3.5-lightning-30b-a3b","name":"Nemotron 3.5 Lightning 30B A3B","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"muse-glimmer-30b":{"id":"muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":20480},"cost":{"input":0,"output":0},"canonical_model_id":"meta/muse-glimmer-30b"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"mistral-small-2603":{"id":"mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.165,"output":0.66,"cache_read":0.165},"canonical_model_id":"mistral/mistral-small-2603"},"gemini-3.5-flash-lite@eu":{"id":"gemini-3.5-flash-lite@eu","name":"Gemini 3.5 Flash Lite (EU)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.33,"output":2.75,"cache_read":0.033},"canonical_model_id":"google/gemini-3.5-flash-lite"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"gpt-4.1@eu":{"id":"gpt-4.1@eu","name":"GPT-4.1 (EU)","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2.2,"output":8.8,"cache_read":0.55},"canonical_model_id":"openai/gpt-4.1"},"gpt-5.6-sol@eu":{"id":"gpt-5.6-sol@eu","name":"GPT-5.6 Sol (EU)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.44,"cache_write":5.5,"tiers":[{"input":8.8,"output":33,"cache_read":0.88,"cache_write":11,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8.8,"output":33,"cache_read":0.88,"cache_write":11}},"canonical_model_id":"openai/gpt-5.6-sol"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"nemotron-3-nano-omni@eu":{"id":"nemotron-3-nano-omni@eu","name":"nemotron-3-nano-omni@eu","description":"The most open, efficient, and accurate omni modal reasoning model for agentic AI.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-20","last_updated":"2026-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":300000},"cost":{"input":0.06,"output":0.24,"cache_read":0.06}},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"qwen3.5-27b":{"id":"qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.26,"output":2.6},"canonical_model_id":"alibaba/qwen3.5-27b"},"glm-5.2@eu":{"id":"glm-5.2@eu","name":"GLM-5.2 (EU)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":1.2,"output":4.2,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":3,"output":15,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"gpt-5-mini@eu":{"id":"gpt-5-mini@eu","name":"GPT-5 Mini (EU)","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":0.275,"output":2.2,"cache_read":0.0275},"canonical_model_id":"openai/gpt-5-mini"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-5-nano@eu":{"id":"gpt-5-nano@eu","name":"GPT-5 Nano (EU)","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":0.055,"output":0.44,"cache_read":0.0055},"canonical_model_id":"openai/gpt-5-nano"},"devstral-latest@eu":{"id":"devstral-latest@eu","name":"devstral-latest@eu","description":"An enterprise grade text model, that excels at using tools to explore codebases, editing multiple files and power software engineering agents.","family":"devstral","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-27","last_updated":"2026-05-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.44,"output":2.2,"cache_read":0.44}},"gpt-5.6-terra@eu":{"id":"gpt-5.6-terra@eu","name":"GPT-5.6 Terra (EU)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.2,"output":13.2,"cache_read":0.22,"cache_write":2.75,"tiers":[{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.4,"output":19.8,"cache_read":0.44,"cache_write":5.5}},"canonical_model_id":"openai/gpt-5.6-terra"},"qwen3.8-flash-next":{"id":"qwen3.8-flash-next","name":"Qwen3.8 Flash Next","description":"Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.8-flash-next"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":4.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":9,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":9}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"ring-2.6-1t":{"id":"ring-2.6-1t","name":"ring-2.6-1t","description":"Inclusion AI ring-2.6-1t","family":"ring","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-05-08","last_updated":"2026-05-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":2.5}},"gemma-4-26b-a4b-it":{"id":"gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.07,"output":0.34,"cache_read":0.07},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":1.5,"output":7,"cache_read":0.15},"canonical_model_id":"google/gemini-3.6-flash"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"gpt-6.1-sol@eu":{"id":"gpt-6.1-sol@eu","name":"GPT-6.1 Sol (EU)","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.4,"output":12,"cache_read":0.12,"cache_write":3,"tiers":[{"input":4.8,"output":18,"cache_read":0.24,"cache_write":6,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4.8,"output":18,"cache_read":0.24,"cache_write":6}},"canonical_model_id":"openai/gpt-6.1-sol"},"claude-opus-4-5@eu":{"id":"claude-opus-4-5@eu","name":"Claude Opus 4.5 (latest) (EU)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-5"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"gpt-4o-mini@eu":{"id":"gpt-4o-mini@eu","name":"GPT-4o mini (EU)","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000},"cost":{"input":0.165,"output":0.66,"cache_read":0.0825},"canonical_model_id":"openai/gpt-4o-mini"},"seed-2.0-mini":{"id":"seed-2.0-mini","name":"Seed 2.0 Mini","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.1,"output":0.4,"cache_read":0.02},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"gpt-5.4@eu":{"id":"gpt-5.4@eu","name":"GPT-5.4 (EU)","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.75,"output":16.5,"cache_read":0.275,"tiers":[{"input":5.5,"output":24.75,"cache_read":0.55,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5.5,"output":24.75,"cache_read":0.55}},"canonical_model_id":"openai/gpt-5.4"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.16,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"gemini-2.5-pro@eu":{"id":"gemini-2.5-pro@eu","name":"Gemini 2.5 Pro (EU)","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":1.25,"output":10,"cache_read":0.31,"cache_write":2.375,"tiers":[{"input":2.5,"output":15,"cache_read":0.62,"cache_write":4.75,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.62,"cache_write":4.75}},"canonical_model_id":"google/gemini-2.5-pro"},"qwen3.8-2.4T-A95B":{"id":"qwen3.8-2.4T-A95B","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"kat-coder-pro":{"id":"kat-coder-pro","name":"kat-coder-pro","description":"KAT-Coder-Pro V2 by KwaiKAT is a non-reasoning model optimized for agentic coding. It delivers strong performance on reasoning-style tasks while requiring significantly fewer output tokens than peer models. With the 1210 release, it achieved a score of 64 on the Artificial Analysis Intelligence Index, placing it in the global Top 10 and ranking first among all non-reasoning models.","family":"kat-coder","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-03-27","last_updated":"2026-03-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.3,"output":1.2}},"mistral-small-2603@eu":{"id":"mistral-small-2603@eu","name":"Mistral Small 4 (EU)","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.165,"output":0.66,"cache_read":0.165},"canonical_model_id":"mistral/mistral-small-2603"},"seed-2.0-code":{"id":"seed-2.0-code","name":"Seed 2.0 Code","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":131072},"cost":{"input":0.5,"output":3,"cache_read":0.1},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"deepseek-v4-flash-0731@eu":{"id":"deepseek-v4-flash-0731@eu","name":"DeepSeek V4 Flash 0731 (EU)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.28,"output":0.56,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"gemini-3.7-flash@eu":{"id":"gemini-3.7-flash@eu","name":"Gemini 3.7 Flash (EU)","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.825,"output":4.125,"cache_read":0.0825},"canonical_model_id":"google/gemini-3.7-flash"},"deepseek-v4-pro-0813@eu":{"id":"deepseek-v4-pro-0813@eu","name":"DeepSeek V4 Pro 0813 (EU)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":1.75,"output":3.5,"cache_read":0.44},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"claude-sonnet-4-5@eu":{"id":"claude-sonnet-4-5@eu","name":"Claude Sonnet 4.5 (latest) (EU)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3.3,"output":16.5,"cache_read":0.3,"cache_write":4.125,"tiers":[{"input":6.6,"output":24.75,"cache_read":0.6,"cache_write":8.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6.6,"output":24.75,"cache_read":0.6,"cache_write":8.25}},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":262144},"cost":{"input":0.2,"output":0.6,"cache_read":0.07},"canonical_model_id":"zhipuai/glm-5.3-flash"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"gemini-3.5-flash@eu":{"id":"gemini-3.5-flash@eu","name":"Gemini 3.5 Flash (EU)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":1.65,"output":9.9,"cache_read":0.165,"cache_write":1.7413},"canonical_model_id":"google/gemini-3.5-flash"},"glm-5.3@eu":{"id":"glm-5.3@eu","name":"GLM-5.3 (EU)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.2,"output":4.2,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"kimi-k2.6@eu":{"id":"kimi-k2.6@eu","name":"Kimi K2.6 (EU)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":128000},"cost":{"input":0.95,"output":4,"cache_read":0.95},"canonical_model_id":"moonshotai/kimi-k2.6"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.8,"output":2.55,"cache_read":0.16},"canonical_model_id":"zhipuai/glm-5.2"},"qwen3.8-2.4T-A95B@eu":{"id":"qwen3.8-2.4T-A95B@eu","name":"Qwen3.8 2.4T A95B (EU)","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":262144},"cost":{"input":2.5,"output":6,"cache_read":0.63},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"leanstral-1-5@eu":{"id":"leanstral-1-5@eu","name":"leanstral-1-5@eu","description":"Leanstral 1.5 is an updated Lean 4 formal proof engineering model from Mistral AI, optimized for automated theorem proving and autoformalization. It has 119B total parameters with 6.5B active and supports a 256K token context window. It supports native function calling and structured output.","family":"leanstral","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-27","last_updated":"2026-05-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0}},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":1.583},"canonical_model_id":"google/gemini-3.5-flash"},"claude-opus-4-6@eu":{"id":"claude-opus-4-6@eu","name":"Claude Opus 4.6 (EU)","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-6"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"claude-sonnet-5@eu":{"id":"claude-sonnet-5@eu","name":"Claude Sonnet 5 (EU)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5"},"deepseek-v4.1-flash@eu":{"id":"deepseek-v4.1-flash@eu","name":"DeepSeek V4.1 Flash (EU)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"claude-fable-5.1":{"id":"claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"laguna-m.1":{"id":"laguna-m.1","name":"Laguna M.1","description":"Poolside's open-weight model for agentic coding and long-horizon work","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-m.1"},"minimax-m3.1-flash":{"id":"minimax-m3.1-flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625},"canonical_model_id":"alibaba/qwen3.6-plus"},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"cache_write":1.25,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"cache_write":2.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4,"cache_write":2.5}},"canonical_model_id":"xai/grok-4.3"},"gemini-3.1-flash-image":{"id":"gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","video","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":2},"canonical_model_id":"google/gemini-3.1-flash-image"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.28,"output":0.56,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"ling-2.6-1t":{"id":"ling-2.6-1t","name":"ling-2.6-1t","description":"Inclusion AI ling-2.6-1t","family":"ling","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":2.5}},"claude-opus-4-7@eu":{"id":"claude-opus-4-7@eu","name":"Claude Opus 4.7 (EU)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-7"},"ling-3.0-tiny":{"id":"ling-3.0-tiny","name":"ling-3.0-tiny","description":"Ling-3.0-tiny is an efficient 7.9B parameter MoE model from inclusionAI with only 1.3B active parameters per token. Built for responsive agents, reliable instruction following and multi turn conversation, with a 256K context window, native function calling, prompt caching and switchable Thinking and Instant modes.","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0}},"fugu-ultra":{"id":"fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra"},"kimi-k2.7-code@eu":{"id":"kimi-k2.7-code@eu","name":"Kimi K2.7 Code (EU)","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.25,"output":4.5,"cache_read":0.31},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"nemotron-3.5-content-safety":{"id":"nemotron-3.5-content-safety","name":"Nemotron 3.5 Content Safety","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"minimax-m2.7-highspeed":{"id":"minimax-m2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":1.2},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"gemma-4-31b-it":{"id":"gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-31b-it"},"grok-4.2-beta":{"id":"grok-4.2-beta","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2,"cache_write":2,"tiers":[{"input":4,"output":12,"cache_read":0.4,"cache_write":4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.4,"cache_write":4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"hy3":{"id":"hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.14,"output":0.58,"cache_read":0.035},"canonical_model_id":"tencent/hy3"},"deepseek-v4.1-flash-1017@eu":{"id":"deepseek-v4.1-flash-1017@eu","name":"DeepSeek V4.1 Flash 1017 (EU)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"kimi-k2.5@eu":{"id":"kimi-k2.5@eu","name":"Kimi K2.5 (EU)","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16000},"cost":{"input":0.72,"output":3.6,"cache_read":0.72,"cache_write":0.72},"canonical_model_id":"moonshotai/kimi-k2.5"},"claude-opus-5@eu":{"id":"claude-opus-5@eu","name":"Claude Opus 5 (EU)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-5"},"grok-build-0.1":{"id":"grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.1},"canonical_model_id":"xai/grok-build-0.1"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.2,"output":4.2,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"mistral-medium-3-5":{"id":"mistral-medium-3-5","name":"Mistral Medium 3.5","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":209715},"cost":{"input":1.65,"output":8.25,"cache_read":1.65},"canonical_model_id":"mistral/mistral-medium-3-5"},"nemotron-3-ultra-nvfp4":{"id":"nemotron-3-ultra-nvfp4","name":"nemotron-3-ultra-nvfp4","description":"Nemotron-3-Ultra-550B-A55B-NVFP4 is a frontier-scale large language model (LLM) trained by NVIDIA, designed to deliver strong agentic, reasoning, and conversational capabilities. It is optimized for the most demanding workloads, including complex multi-step agents, long-context analysis, and high-accuracy reasoning over code, math, and science. The model employs a hybrid Latent Mixture-of-Experts (LatentMoE) architecture, utilizing interleaved Mamba-2 and MoE layers, along with select Attention layers. Like the Super model, the Ultra model incorporates Multi-Token Prediction (MTP) layers for faster text generation and improved quality, and it is trained using an NVFP4 pre-training recipe to maximize compute efficiency. The model has 55B active parameters and 550B parameters in total.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":2.4,"cache_read":0.12}},"gemini-3-pro-image":{"id":"gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":1048576,"output":32768},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":4.5},"canonical_model_id":"google/gemini-3-pro-image"},"inkling":{"id":"inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":32768},"cost":{"input":1.87,"output":4.68,"cache_read":0.374},"canonical_model_id":"thinkingmachines/inkling"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"glm-5.2-fast":{"id":"glm-5.2-fast","name":"glm-5.2-fast","description":"GLM-5.2 introduces a robust 1M-token context and advanced, multi-effort coding capabilities to significantly enhance performance on long-horizon tasks. Its new IndexShare architecture and improved MTP layer simultaneously boost efficiency by reducing per-token FLOPs and increasing speculative decoding lengths. A 743B-parameter model in Zhipu AI's GLM series, designed to plan, execute, and iterate autonomously on extended, engineering-grade tasks.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-07-13","last_updated":"2026-07-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.1,"output":6.6,"cache_read":0.21}},"devstral-latest":{"id":"devstral-latest","name":"devstral-latest","description":"An enterprise grade text model, that excels at using tools to explore codebases, editing multiple files and power software engineering agents.","family":"devstral","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-27","last_updated":"2026-05-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.44,"output":2.2,"cache_read":0.44}},"claude-sonnet-5-5@eu":{"id":"claude-sonnet-5-5@eu","name":"Claude Sonnet 5.5 (EU)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"muse-spark-open":{"id":"muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"claude-haiku-5-5@eu":{"id":"claude-haiku-5-5@eu","name":"Claude Haiku 5.5 (EU)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.11,"output":0.55,"cache_read":0.02,"cache_write":0.14,"tiers":[{"input":0.55,"output":2.75,"cache_read":0.06,"cache_write":0.69,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"gpt-5.6-luna@eu":{"id":"gpt-5.6-luna@eu","name":"GPT-5.6 Luna (EU)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.22,"output":1.32,"cache_read":0.022,"cache_write":0.275,"tiers":[{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.44,"output":1.98,"cache_read":0.044,"cache_write":0.55}},"canonical_model_id":"openai/gpt-5.6-luna"}}},"perplexity-agent":{"id":"perplexity-agent","env":["PERPLEXITY_API_KEY"],"npm":"@ai-sdk/openai","api":"https://api.perplexity.ai/v1","name":"Perplexity Agent","doc":"https://docs.perplexity.ai/docs/agent-api/models","models":{"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"status":"deprecated","cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04}},"canonical_model_id":"openai/gpt-5.6-luna"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0.13,"output":0.26,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"anthropic/claude-sonnet-5-5":{"id":"anthropic/claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-haiku-4-5":{"id":"anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-haiku-5-5":{"id":"anthropic/claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-sonnet-4-5":{"id":"anthropic/claude-sonnet-4-5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-5-5":{"id":"anthropic/claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-fable-5-1":{"id":"anthropic/claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-4-5":{"id":"anthropic/claude-opus-4-5","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"moonshot-ai/kimi-k2.7-code":{"id":"moonshot-ai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshot-ai/kimi-k3":{"id":"moonshot-ai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"perplexity/deepseek-v4-pro-0813":{"id":"perplexity/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"perplexity/nemotron-3.5-super-120b-a12b":{"id":"perplexity/nemotron-3.5-super-120b-a12b","name":"Nemotron 3.5 Super 120B A12B","description":"NVIDIA Nemotron 3.5 Super reasoning model (120B total, 12B active) served with text-only input","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.2,"output":0.5,"cache_read":0.04},"canonical_model_id":"nvidia/nemotron-3.5-super-120b-a12b"},"perplexity/kimi-k3":{"id":"perplexity/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"perplexity/glm-5.3-flash":{"id":"perplexity/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"perplexity/glm-5.3":{"id":"perplexity/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"perplexity/sonar":{"id":"perplexity/sonar","name":"Sonar","description":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","family":"sonar","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.25,"output":2.5,"cache_read":0.0625},"canonical_model_id":"perplexity/sonar"},"xai/grok-4.5":{"id":"xai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"xai/grok-4.6":{"id":"xai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"xai/grok-4.7":{"id":"xai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"cache_write":2,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"xai/grok-4-1-fast-non-reasoning":{"id":"xai/grok-4-1-fast-non-reasoning","name":"Grok 4.1 Fast (Non-Reasoning)","description":"xAI's fast agentic tool-calling model with a 2M context window; non-reasoning variant for low-latency responses","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"status":"deprecated","cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast"},"nvidia/nemotron-3-super-120b-a12b":{"id":"nvidia/nemotron-3-super-120b-a12b","name":"Nemotron 3 Super 120B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32000},"status":"deprecated","cost":{"input":0.25,"output":2.5},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"google/gemini-4-argon"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"tiers":[{"input":0.5,"output":3,"cache_read":0.05,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":0.5,"output":3,"cache_read":0.05}},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"}}},"abliteration-ai":{"id":"abliteration-ai","env":["ABLIT_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.abliteration.ai/v1","name":"abliteration.ai","doc":"https://docs.abliteration.ai/models","models":{"abliterated-model":{"id":"abliterated-model","name":"Abliterated Model","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-01-06","last_updated":"2026-07-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":150000,"input":150000,"output":8192},"cost":{"input":3,"output":3,"cache_read":0.3}},"abliterated-model-large":{"id":"abliterated-model-large","name":"Abliterated Model Large","description":"GLM-5.2 model abliterated and finetuned for cyber, ML red teaming, and agent testing","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-25","last_updated":"2026-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":999990},"cost":{"input":5,"output":5,"cache_read":0.5}},"abliterated-model-large-v2":{"id":"abliterated-model-large-v2","name":"Abliterated Model Large V2","description":"GLM-5.3 model abliterated and finetuned for cyber, ML red teaming, and agent testing","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-29","last_updated":"2026-08-31","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":999990},"cost":{"input":5,"output":5,"cache_read":0.5}}}},"xiaomi-token-plan-sgp":{"id":"xiaomi-token-plan-sgp","env":["XIAOMI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://token-plan-sgp.xiaomimimo.com/v1","name":"Xiaomi Token Plan (Singapore)","doc":"https://platform.xiaomimimo.com/#/docs","models":{"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"mimo-v2.5-tts-voiceclone":{"id":"mimo-v2.5-tts-voiceclone","type":"speech","name":"MiMo-V2.5-TTS-VoiceClone","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.5-tts-voiceclone"},"mimo-v2-pro":{"id":"mimo-v2-pro","name":"MiMo-V2-Pro","description":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2-pro"},"mimo-v2.5-tts":{"id":"mimo-v2.5-tts","type":"speech","name":"MiMo-V2.5-TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.5-tts"},"mimo-v2.5-tts-voicedesign":{"id":"mimo-v2.5-tts-voicedesign","type":"speech","name":"MiMo-V2.5-TTS-VoiceDesign","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.5-tts-voicedesign"},"mimo-v2-tts":{"id":"mimo-v2-tts","type":"speech","name":"MiMo-V2-TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":8192},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2-tts"},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5"}}},"nearai":{"id":"nearai","env":["NEARAI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://cloud-api.near.ai/v1","name":"NEAR AI Cloud","doc":"https://docs.near.ai/","models":{"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.08,"output":0.5,"cache_read":0.012},"canonical_model_id":"zhipuai/glm-5.3-flash"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.81,"output":3.85,"cache_read":0.41},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3.3,"output":16.5,"cache_read":0.33},"canonical_model_id":"moonshotai/kimi-k3"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1},"canonical_model_id":"openai/gpt-6-astra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/whisper-large-v3":{"id":"openai/whisper-large-v3","type":"transcription","name":"Whisper Large v3","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":448},"cost":{"input":0.01,"output":0.01},"canonical_model_id":"openai/whisper-large-v3"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":64000},"cost":{"input":1.1,"output":1.1,"cache_read":0.55},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"anthropic/claude-sonnet-5-5":{"id":"anthropic/claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4-8":{"id":"anthropic/claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-haiku-4-5":{"id":"anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-haiku-5-5":{"id":"anthropic/claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-sonnet-4-5":{"id":"anthropic/claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-5-5":{"id":"anthropic/claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-fable-5-1":{"id":"anthropic/claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2},"canonical_model_id":"anthropic/claude-sonnet-5"},"black-forest-labs/FLUX.2-klein-4B":{"id":"black-forest-labs/FLUX.2-klein-4B","type":"image","name":"FLUX.2 Klein 4B","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-15","last_updated":"2026-01-15","modalities":{"input":["text","image"],"output":["image"]},"open_weights":true,"limit":{"context":0,"output":0},"canonical_model_id":"black-forest-labs/flux-2-klein-4b"},"Qwen/Qwen3-Reranker-0.6B":{"id":"Qwen/Qwen3-Reranker-0.6B","type":"reranking","name":"Qwen3 Reranker 0.6B","description":"Reranking model for improving retrieval quality in search and recommendation systems","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":1024},"cost":{"input":0.01,"output":0.01},"canonical_model_id":"alibaba/qwen3-reranker-0.6b"},"Qwen/Qwen3-VL-30B-A3B-Instruct":{"id":"Qwen/Qwen3-VL-30B-A3B-Instruct","name":"Qwen3-VL 30B-A3B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":8192},"cost":{"input":0.15,"output":0.55,"cache_read":0.03},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.04,"output":1.35,"cache_read":0.018},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3-Embedding-0.6B":{"id":"Qwen/Qwen3-Embedding-0.6B","type":"embedding","name":"Qwen3 Embedding 0.6B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":1024},"cost":{"input":0.01,"output":0.01},"canonical_model_id":"alibaba/qwen3-embedding-0.6b"},"Qwen/Qwen3.6-35B-A3B-FP8":{"id":"Qwen/Qwen3.6-35B-A3B-FP8","name":"Qwen 3.6 35B A3B FP8","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.17,"output":1.1,"cache_read":0.056},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.8,"output":7.5},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.5,"output":3.3,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"x-ai/grok-4.7":{"id":"x-ai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"}}},"ai21":{"id":"ai21","env":["AI21_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.ai21.com/studio/v1","name":"AI21 Labs","doc":"https://docs.ai21.com/docs/jamba-foundation-models","models":{"jamba-large":{"id":"jamba-large","name":"Jamba Large","description":"AI21's hybrid SSM-Transformer long-context model for enterprise agents and grounded generation","family":"jamba","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08-22","release_date":"2025-07-03","last_updated":"2025-07-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":4096},"cost":{"input":2,"output":8},"canonical_model_id":"ai21/jamba-large"},"jamba-mini":{"id":"jamba-mini","name":"Jamba Mini","description":"AI21's efficient, lightweight hybrid SSM-Transformer model for a wide range of tasks","family":"jamba","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08-22","release_date":"2026-01-08","last_updated":"2026-01-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":4096},"cost":{"input":0.2,"output":0.4},"canonical_model_id":"ai21/jamba-mini"}}},"kimi-code-plan-global":{"id":"kimi-code-plan-global","env":["KIMI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.kimi.ai/coding/v1","name":"Kimi For Coding (kimi.ai)","doc":"https://www.kimi.ai/code/docs/en/kimi-code/models.html","models":{"kimi-for-coding":{"id":"kimi-for-coding","name":"kimi-for-coding","description":"Kimi coding model with more efficient thinking and up to 1M context, available through Kimi Code","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.8-preview"},"kimi-for-coding-highspeed":{"id":"kimi-for-coding-highspeed","name":"Kimi For Coding HighSpeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"k3":{"id":"k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k3"},"k3-256k":{"id":"k3-256k","name":"Kimi K3-256K","description":"256K-context version of Kimi K3, reducing token consumption for shorter coding sessions","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k3"}}},"echo":{"id":"echo","env":["ECHO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://echo.tracerml.ai/v1","name":"Echo","doc":"https://echo.tracerml.ai/docs/api","models":{"echo":{"id":"echo","name":"Echo","description":"Adaptive model for coding, reasoning, and tool-driven agent workflows through one OpenAI-compatible endpoint","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-19","last_updated":"2026-08-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"status":"beta","cost":{"input":10,"output":50}}}},"xpersona":{"id":"xpersona","env":["XPERSONA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://www.xpersona.co/v1","name":"Xpersona","doc":"https://www.xpersona.co/docs","models":{"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":1.5,"output":9.25,"reasoning":9.25,"cache_read":0.15},"canonical_model_id":"anthropic/claude-opus-4-8"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":372000,"output":128000},"cost":{"input":1.5,"output":2,"reasoning":2,"cache_read":0.15},"canonical_model_id":"openai/gpt-5.6-terra"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":0.6,"output":3.7,"reasoning":3.7,"cache_read":0.06},"canonical_model_id":"anthropic/claude-haiku-4-5"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":1.5,"output":12,"reasoning":12,"cache_read":0.15},"canonical_model_id":"openai/gpt-5.5"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":0.9,"output":5.55,"reasoning":5.55,"cache_read":0.09},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0.75,"output":6,"reasoning":6,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":18.5,"reasoning":18.5,"cache_read":0.3},"canonical_model_id":"anthropic/claude-fable-5"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":372000,"output":128000},"cost":{"input":1.5,"output":12,"reasoning":12,"cache_read":0.15},"canonical_model_id":"openai/gpt-5.6-sol"},"gpt-5.6":{"id":"gpt-5.6","name":"GPT-5.6","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":372000,"output":128000},"cost":{"input":1.5,"output":12,"reasoning":12,"cache_read":0.15},"canonical_model_id":"openai/gpt-5.6-sol"},"xpersona-gpt-5.5":{"id":"xpersona-gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-12-30","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"deprecated","cost":{"input":3,"output":18,"reasoning":18,"cache_read":0.3},"canonical_model_id":"openai/gpt-5.5"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":1.55,"output":12.2,"reasoning":12.2,"cache_read":0.155},"canonical_model_id":"google/gemini-3.5-flash"},"xpersona-frieren-coder":{"id":"xpersona-frieren-coder","name":"Xpersona Frieren 1","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-30","release_date":"2026-05-01","last_updated":"2026-05-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":384000},"cost":{"input":1.5,"output":6,"reasoning":6,"cache_read":0.15}},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":0.375,"output":4,"reasoning":4,"cache_read":0.0375},"canonical_model_id":"openai/gpt-5.4-mini"}}},"azure-cognitive-services":{"id":"azure-cognitive-services","env":["AZURE_COGNITIVE_SERVICES_RESOURCE_NAME","AZURE_COGNITIVE_SERVICES_API_KEY"],"npm":"@ai-sdk/azure","name":"Azure Cognitive Services","doc":"https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models","models":{"grok-4-20-reasoning":{"id":"grok-4-20-reasoning","name":"Grok 4.20 (Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-04-08","last_updated":"2026-04-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":8192},"status":"beta","cost":{"input":2,"output":6}},"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-31","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"anthropic/claude-opus-4-8"},"llama-4-scout-17b-16e-instruct":{"id":"llama-4-scout-17b-16e-instruct","name":"Llama 4 Scout 17B 16E Instruct","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.2,"output":0.78}},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-14","last_updated":"2025-11-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":2.5,"output":20,"cache_read":0.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"gpt-5-pro":{"id":"gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"status":"beta","cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"cohere-command-a":{"id":"cohere-command-a","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":2.5,"output":10}},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"o3-mini":{"id":"o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":4,"output":24,"cache_read":0.4,"cache_write":5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"phi-4-reasoning-plus":{"id":"phi-4-reasoning-plus","name":"Phi-4-reasoning-plus","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"phi","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.125,"output":0.5}},"llama-4-maverick-17b-128e-instruct-fp8":{"id":"llama-4-maverick-17b-128e-instruct-fp8","name":"Llama 4 Maverick 17B 128E Instruct FP8","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":16384},"cost":{"input":0.25,"output":1}},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"gpt-chat-latest":{"id":"gpt-chat-latest","name":"GPT Chat Latest","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-05-05","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"beta","cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5-instant"},"codex-mini":{"id":"codex-mini","name":"Codex Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-04","release_date":"2025-05-16","last_updated":"2025-05-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.5,"output":6,"cache_read":0.375}},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25}},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-01-14","last_updated":"2026-01-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"phi-4-reasoning":{"id":"phi-4-reasoning","name":"Phi-4-reasoning","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"phi","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.125,"output":0.5}},"phi-4-mini-reasoning":{"id":"phi-4-mini-reasoning","name":"Phi-4-mini-reasoning","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"phi","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.075,"output":0.3}},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"text-embedding-3-large":{"id":"text-embedding-3-large","name":"text-embedding-3-large","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":3072},"cost":{"input":0.13,"output":0}},"gpt-5.4-pro":{"id":"gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"gpt-3.5-turbo-instruct":{"id":"gpt-3.5-turbo-instruct","name":"GPT-3.5 Turbo Instruct","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2021-08","release_date":"2023-09-21","last_updated":"2023-09-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4096,"output":4096},"status":"deprecated","cost":{"input":1.5,"output":2}},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek-V4-Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.19,"output":0.51},"canonical_model_id":"deepseek/deepseek-v4-flash"},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"status":"beta","canonical_model_id":"xai/grok-4.7"},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005}},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":28,"cache_read":0.35},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.95,"output":4},"canonical_model_id":"moonshotai/kimi-k2.6"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"gpt-3.5-turbo-0125":{"id":"gpt-3.5-turbo-0125","name":"GPT-3.5 Turbo 0125","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2021-08","release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":4096},"status":"deprecated","cost":{"input":0.5,"output":1.5}},"claude-opus-4-1":{"id":"claude-opus-4-1","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"deepseek-v3.2-speciale":{"id":"deepseek-v3.2-speciale","name":"DeepSeek-V3.2-Speciale","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.58,"output":1.68}},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.58,"output":1.68}},"gpt-3.5-turbo-1106":{"id":"gpt-3.5-turbo-1106","name":"GPT-3.5 Turbo 1106","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2021-08","release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":4096},"status":"deprecated","cost":{"input":1,"output":2}},"gpt-5-codex":{"id":"gpt-5-codex","name":"GPT-5-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"o3":{"id":"o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek-V4-Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":1.74,"output":3.48},"canonical_model_id":"deepseek/deepseek-v4-pro"},"grok-4-1-fast-reasoning":{"id":"grok-4-1-fast-reasoning","name":"Grok 4.1 Fast (Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-06-27","last_updated":"2025-06-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"beta","provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"deepseek-r1":{"id":"deepseek-r1","name":"DeepSeek-R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-01-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"status":"deprecated","cost":{"input":1.35,"output":5.4}},"gpt-image-2":{"id":"gpt-image-2","type":"image","name":"GPT-Image-2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":8,"output":40,"cache_read":0.8,"cache_write":10},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"text-embedding-3-small":{"id":"text-embedding-3-small","name":"text-embedding-3-small","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536},"cost":{"input":0.02,"output":0}},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"gpt-4-turbo-vision":{"id":"gpt-4-turbo-vision","name":"GPT-4 Turbo Vision","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":10,"output":30}},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.71,"output":0.71}},"gpt-image-1":{"id":"gpt-image-1","type":"image","name":"GPT-Image-1","description":"OpenAI image model for production generation, edits, and brand-safe visual workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-24","last_updated":"2025-04-24","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":40,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-1"},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"gpt-image-2.5-sunburst":{"id":"gpt-image-2.5-sunburst","type":"image","name":"GPT Image 2.5 Sunburst","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-sunburst"},"cohere-embed-v3-multilingual":{"id":"cohere-embed-v3-multilingual","name":"Embed v3 Multilingual","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"cohere-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-11-07","last_updated":"2023-11-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":1024},"cost":{"input":0.1,"output":0}},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-07-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}}},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"gpt-4o-mini":{"id":"gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"gpt-image-1.5":{"id":"gpt-image-1.5","type":"image","name":"GPT-Image-1.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":32,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-1.5"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.6,"output":3},"canonical_model_id":"moonshotai/kimi-k2.5"},"gpt-image-2.5-flare":{"id":"gpt-image-2.5-flare","type":"image","name":"GPT Image 2.5 Flare","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-flare"},"cohere-embed-v-4-0":{"id":"cohere-embed-v-4-0","name":"Embed v4","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"cohere-embed","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":1536},"cost":{"input":0.12,"output":0}},"gpt-4-turbo":{"id":"gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"ministral-3b":{"id":"ministral-3b","name":"Ministral 3B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-03","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.04,"output":0.04}},"phi-4-mini":{"id":"phi-4-mini","name":"Phi-4-mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"phi","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.075,"output":0.3}},"grok-4-20-non-reasoning":{"id":"grok-4-20-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-04-08","last_updated":"2026-04-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262000,"output":8192},"status":"beta","cost":{"input":2,"output":6}},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-11-24","last_updated":"2025-08-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek-V4.1-Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"status":"beta","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models"},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"codestral-2501":{"id":"codestral-2501","name":"Codestral 25.01","description":"Mistral coding model for code completion, generation, and developer workflows","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-03","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.3,"output":0.9}},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"o1":{"id":"o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":15,"output":60,"cache_read":7.5}},"gpt-5.1-codex-max":{"id":"gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex-max"},"mistral-small-2503":{"id":"mistral-small-2503","name":"Mistral Small 3.1","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2025-03-01","last_updated":"2025-03-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":0.1,"output":0.3}},"mistral-medium-2505":{"id":"mistral-medium-2505","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.4,"output":2}},"gpt-5.1-codex-mini":{"id":"gpt-5.1-codex-mini","name":"GPT-5.1 Codex Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-14","last_updated":"2025-11-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"grok-4-1-fast-non-reasoning":{"id":"grok-4-1-fast-non-reasoning","name":"Grok 4.1 Fast (Non-Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-06-27","last_updated":"2025-06-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}},"phi-4-multimodal":{"id":"phi-4-multimodal","name":"Phi-4-multimodal","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"phi","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.08,"output":0.32,"input_audio":4}},"model-router":{"id":"model-router","name":"Model Router","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"model-router","attachment":true,"reasoning":false,"tool_call":true,"release_date":"2025-05-19","last_updated":"2025-11-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":16384},"provider":{"npm":"@ai-sdk/azure","shape":"completions"},"cost":{"input":0.14,"output":0}},"claude-mythos-5":{"id":"claude-mythos-5","name":"Claude Mythos 5","description":"Restricted Claude model for advanced cybersecurity and biology research workflows","family":"claude-mythos","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"beta","provider":{"npm":"@ai-sdk/anthropic","api":"https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-mythos-5"},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-14","last_updated":"2025-11-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"phi-4":{"id":"phi-4","name":"Phi-4","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"phi","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-10","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.125,"output":0.5}},"cohere-embed-v3-english":{"id":"cohere-embed-v3-english","name":"Embed v3 English","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"cohere-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-11-07","last_updated":"2023-11-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":1024},"cost":{"input":0.1,"output":0}},"text-embedding-ada-002":{"id":"text-embedding-ada-002","name":"text-embedding-ada-002","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2022-12-15","last_updated":"2022-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536},"cost":{"input":0.1,"output":0}},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":1.5,"output":9,"cache_read":0.15},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"}}},"morph":{"id":"morph","env":["MORPH_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.morphllm.com/v1","name":"Morph","doc":"https://docs.morphllm.com/api-reference/introduction","models":{"morph-v3-large":{"id":"morph-v3-large","name":"Morph v3 Large","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-08-15","last_updated":"2024-08-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"cost":{"input":0.9,"output":1.9}},"morph-v3-fast":{"id":"morph-v3-fast","name":"Morph v3 Fast","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-08-15","last_updated":"2024-08-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":16000},"cost":{"input":0.8,"output":1.2}},"auto":{"id":"auto","name":"Auto","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-06-01","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"cost":{"input":0.85,"output":1.55}}}},"merge-gateway":{"id":"merge-gateway","env":["MERGE_GATEWAY_API_KEY"],"npm":"merge-gateway-ai-sdk-provider","api":"https://api-gateway.merge.dev/v1/ai-sdk","name":"Merge Gateway","doc":"https://docs.merge.dev/merge-gateway","models":{"meta/llama-3.1-8b-instruct":{"id":"meta/llama-3.1-8b-instruct","name":"Llama 3.1 8B","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":2048},"cost":{"input":0.22,"output":0.22},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta/llama-3.1-70b-instruct":{"id":"meta/llama-3.1-70b-instruct","name":"Llama 3.1 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":2048},"cost":{"input":0.99,"output":0.99},"canonical_model_id":"meta/llama-3.1-70b-instruct"},"meta/llama-3.3-70b-instruct":{"id":"meta/llama-3.3-70b-instruct","name":"Llama 3.3 70B Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.22,"output":0.5,"cache_read":0.11},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":262144},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":262144},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"meta/muse-spark-open":{"id":"meta/muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"sakana/sakana-namazu":{"id":"sakana/sakana-namazu","name":"Sakana Namazu","description":"Japanese-specialized reasoning model based on Kimi K2.6 and tuned for Japanese language, culture, and business workflows","family":"sakana-namazu","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.95,"output":4,"cache_read":0.15},"canonical_model_id":"sakana/sakana-namazu"},"sakana/fugu-max":{"id":"sakana/fugu-max","name":"Fugu Max","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"sakana/fugu-max"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra"},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.6,"output":2.5},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"mistral/devstral-2512":{"id":"mistral/devstral-2512","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/devstral-2512"},"mistral/mistral-medium-latest":{"id":"mistral/mistral-medium-latest","name":"Mistral Medium (latest)","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/mistral-medium-latest"},"mistral/magistral-medium-latest":{"id":"mistral/magistral-medium-latest","name":"Magistral Medium (latest)","description":"Mistral reasoning model for transparent analysis, math, and complex decisions","family":"magistral-medium","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-03-17","last_updated":"2025-03-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2,"output":5},"canonical_model_id":"mistral/magistral-medium-latest"},"mistral/mistral-small-latest":{"id":"mistral/mistral-small-latest","name":"Mistral Small (latest)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"mistral/mistral-small-latest"},"mistral/mistral-large-2411":{"id":"mistral/mistral-large-2411","name":"Mistral Large 2.1","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-18","last_updated":"2024-11-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/mistral-large-2411"},"mistral/mistral-large-latest":{"id":"mistral/mistral-large-latest","name":"Mistral Large (latest)","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/mistral-large-latest"},"mistral/devstral-medium-2507":{"id":"mistral/devstral-medium-2507","name":"Devstral Medium","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-07-10","last_updated":"2025-07-10","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/devstral-medium-2507"},"mistral/devstral-medium-latest":{"id":"mistral/devstral-medium-latest","name":"Devstral 2 (latest)","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/devstral-medium-latest"},"mistral/pixtral-large-latest":{"id":"mistral/pixtral-large-latest","name":"Pixtral Large (latest)","description":"Mistral's larger vision model for document-heavy image understanding and chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2024-11-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/pixtral-large-latest"},"mistral/devstral-small-2507":{"id":"mistral/devstral-small-2507","name":"Devstral Small","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-07-10","last_updated":"2025-07-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/devstral-small-2507"},"mistral/codestral-latest":{"id":"mistral/codestral-latest","name":"Codestral (latest)","description":"Mistral code model for completions, refactors, and developer IDE workflows","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-05-29","last_updated":"2025-01-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":4096},"cost":{"input":0.3,"output":0.9},"canonical_model_id":"mistral/codestral-latest"},"mistral/mistral-medium-2505":{"id":"mistral/mistral-medium-2505","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/mistral-medium-2505"},"mistral/mistral-large-2512":{"id":"mistral/mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"mistral/mistral-large-2512"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048000,"output":32000},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"openai/gpt-5-search-api":{"id":"openai/gpt-5-search-api","type":"chat","name":"GPT-5 Search API","description":"GPT-5 model trained for web search in the Chat Completions API","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-10-14","last_updated":"2025-10-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":1.25,"output":10},"canonical_model_id":"openai/gpt-5-search-api"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 Mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-4o-2024-05-13":{"id":"openai/gpt-4o-2024-05-13","name":"GPT-4o (2024-05-13)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-05-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":5,"output":15},"canonical_model_id":"openai/gpt-4o-2024-05-13"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3 Mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-oss-safeguard-120b":{"id":"openai/gpt-oss-safeguard-120b","name":"GPT OSS Safeguard 120B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-safeguard-120b"},"openai/gpt-audio-mini":{"id":"openai/gpt-audio-mini","type":"chat","name":"GPT Audio Mini","description":"Cost-efficient audio model for speech input and output with the Chat Completions API","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-01","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.6,"output":2.4},"canonical_model_id":"openai/gpt-audio-mini"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT-OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.04,"output":0.15,"cache_read":0.02},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-5-chat-latest":{"id":"openai/gpt-5-chat-latest","name":"GPT-5 Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5-chat-latest"},"openai/gpt-4":{"id":"openai/gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":30,"output":60},"canonical_model_id":"openai/gpt-4"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-oss-safeguard-20b":{"id":"openai/gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":0.07,"output":0.2,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-audio":{"id":"openai/gpt-audio","type":"chat","name":"GPT Audio","description":"Audio model for speech input and output with the Chat Completions API","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-01","release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10},"canonical_model_id":"openai/gpt-audio"},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT-OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.05,"output":0.25},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.3-chat-latest":{"id":"openai/gpt-5.3-chat-latest","name":"GPT-5.3 Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-chat-latest"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-live-1":{"id":"openai/gpt-live-1","type":"realtime","name":"GPT-Live 1","description":"Full-duplex voice model for natural conversations that delegates reasoning and tools to a backend agent","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-07-31","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":128000,"output":4096},"provider":{"api":"wss://api-gateway.merge.dev/v1"},"canonical_model_id":"openai/gpt-live-1"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/whisper-1":{"id":"openai/whisper-1","type":"transcription","name":"Whisper","description":"General-purpose speech recognition model for transcription and translation","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2023-03-01","last_updated":"2023-03-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":448000,"output":4096},"provider":{"npm":"@ai-sdk/openai","api":"https://api-gateway.merge.dev/v1"},"canonical_model_id":"openai/whisper-1"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 Nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o Mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4o-2024-11-20":{"id":"openai/gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"openai/gpt-4o-2024-08-06":{"id":"openai/gpt-4o-2024-08-06","name":"GPT-4o (2024-08-06)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-08-06"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4 Mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.2-chat-latest":{"id":"openai/gpt-5.2-chat-latest","name":"GPT-5.2 Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-chat-latest"},"openai/gpt-5.1-chat-latest":{"id":"openai/gpt-5.1-chat-latest","name":"GPT-5.1 Chat Latest","description":"Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-chat-latest"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":1.5,"output":9,"cache_read":0.15},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-v4-flash-0731-fast":{"id":"deepseek/deepseek-v4-flash-0731-fast","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.28,"output":0.56,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-v3.1":{"id":"deepseek/deepseek-v3.1","name":"DeepSeek V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.56,"output":1.68,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek/deepseek-r1-0528":{"id":"deepseek/deepseek-r1-0528","name":"DeepSeek R1 (0528)","description":"May 2025 DeepSeek R1 update with improved mathematics, coding, and reasoning","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":32768},"cost":{"input":1.35,"output":5.4,"cache_read":0.945},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006,"cache_write":0.1125},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v3":{"id":"deepseek/deepseek-v3","name":"DeepSeek V3","description":"Open DeepSeek MoE chat model for coding, math, and general reasoning","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":81920},"cost":{"input":0.27,"output":1.1},"canonical_model_id":"deepseek/deepseek-v3"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.28,"output":0.4,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.74,"output":3.48,"cache_read":0.003625},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-r1":{"id":"deepseek/deepseek-r1","name":"DeepSeek R1","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":40960},"cost":{"input":1.35,"output":5.4},"canonical_model_id":"deepseek/deepseek-r1"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4-pro-0423":{"id":"deepseek/deepseek-v4-pro-0423","name":"DeepSeek V4 Pro 0423","description":"DeepSeek V4 Pro initial snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":1.65,"output":3.3},"canonical_model_id":"deepseek/deepseek-v4-pro-0423"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.035,"output":0.07,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-v4-flash-0423":{"id":"deepseek/deepseek-v4-flash-0423","name":"DeepSeek V4 Flash 0423","description":"Initial DeepSeek V4 Flash snapshot for economical reasoning, coding, and million-token agent workloads","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.139,"output":0.278},"canonical_model_id":"deepseek/deepseek-v4-flash-0423"},"anthropic/claude-sonnet-5-5":{"id":"anthropic/claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4-8":{"id":"anthropic/claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":128000}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-haiku-5-5":{"id":"anthropic/claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-4-1-20250805":{"id":"anthropic/claude-opus-4-1-20250805","name":"Claude Opus 4.1 (20250805)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-5-5":{"id":"anthropic/claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4-20250514":{"id":"anthropic/claude-opus-4-20250514","name":"Claude Opus 4 (20250514)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-20250514"},"anthropic/claude-fable-5-1":{"id":"anthropic/claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-opus-4-5-20251101":{"id":"anthropic/claude-opus-4-5-20251101","name":"Claude Opus 4.5 (20251101)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024,"max":127999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-sonnet-4-5-20250929":{"id":"anthropic/claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5 (20250929)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"anthropic/claude-haiku-4-5-20251001":{"id":"anthropic/claude-haiku-4-5-20251001","name":"Claude Haiku 4.5 (20251001)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-sonnet-4-20250514":{"id":"anthropic/claude-sonnet-4-20250514","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"anthropic/claude-3-7-sonnet-20250219":{"id":"anthropic/claude-3-7-sonnet-20250219","name":"Claude 3.7 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10-31","release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-3-7-sonnet-20250219"},"moonshot/kimi-k2.6":{"id":"moonshot/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshot/kimi-k2.7-code-highspeed":{"id":"moonshot/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshot/kimi-k2.7-code":{"id":"moonshot/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshot/kimi-k3":{"id":"moonshot/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":2.9,"output":14,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshot/kimi-k2.5":{"id":"moonshot/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"writer/palmyra-x5":{"id":"writer/palmyra-x5","name":"Palmyra X5","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"palmyra","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.6,"output":6},"canonical_model_id":"writer/palmyra-x5"},"writer/palmyra-x4":{"id":"writer/palmyra-x4","name":"Palmyra X4","description":"Enterprise language model for workflow automation, coding, data analysis, and tool use","family":"palmyra","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2024-10-09","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000},"cost":{"input":2.5,"output":10},"canonical_model_id":"writer/palmyra-x4"},"minimax/minimax-m2.5-highspeed":{"id":"minimax/minimax-m2.5-highspeed","name":"MiniMax M2.5 Highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":8192},"cost":{"input":0.6,"output":2.4,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5-highspeed"},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":128000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5"},"minimax/minimax-m3.1-flash":{"id":"minimax/minimax-m3.1-flash","name":"MiniMax M3.1 Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax M2.7 Highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":8192},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"xai/grok-4.5":{"id":"xai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.5"},"xai/grok-4.6":{"id":"xai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"xai/grok-4.20-0309-non-reasoning":{"id":"xai/grok-4.20-0309-non-reasoning","name":"Grok 4.20 Non-Reasoning","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"xai/grok-4.7":{"id":"xai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":2,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"xai/grok-4.3":{"id":"xai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"xai/grok-4.20-0309-reasoning":{"id":"xai/grok-4.20-0309-reasoning","name":"Grok 4.20","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"xai/grok-build-0.1":{"id":"xai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2},"canonical_model_id":"xai/grok-build-0.1"},"nvidia/nemotron-nano-9b-v2":{"id":"nvidia/nemotron-nano-9b-v2","name":"Nemotron Nano 9B","description":"Compact Nemotron model for efficient reasoning and deployable AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-09","release_date":"2025-08-18","last_updated":"2025-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.06,"output":0.23},"canonical_model_id":"nvidia/nemotron-nano-9b-v2"},"nvidia/nemotron-3.5-lightning-30b-a3b":{"id":"nvidia/nemotron-3.5-lightning-30b-a3b","name":"Nemotron 3.5 Lightning 30B A3B","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-3-nano-30b-a3b":{"id":"nvidia/nemotron-3-nano-30b-a3b","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"cost":{"input":0.05,"output":0.2},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"bytedance/dola-seed-2.0-mini":{"id":"bytedance/dola-seed-2.0-mini","name":"Seed 2.0 Mini","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.1,"output":0.4},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"bytedance/dola-seed-2.0-code-preview":{"id":"bytedance/dola-seed-2.0-code-preview","name":"Dola Seed 2.0 Code (preview)","description":"Preview coding model for repository understanding, refactors, and engineering tasks","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-28","last_updated":"2026-03-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3}},"bytedance/dola-seed-2.0-code":{"id":"bytedance/dola-seed-2.0-code","name":"Seed 2.0 Code","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"seed","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.4,"output":2.4},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"bytedance/dola-seed-2.0-pro":{"id":"bytedance/dola-seed-2.0-pro","name":"Seed 2.0 Pro","description":"Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows","family":"seed","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3},"canonical_model_id":"bytedance-seed/seed-2.0-pro"},"bytedance/dola-seed-2.0-lite":{"id":"bytedance/dola-seed-2.0-lite","name":"Seed 2.0 Lite","description":"Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.25,"output":2},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"zai/glm-4.5-air":{"id":"zai/glm-4.5-air","name":"GLM-4.5 Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":98304},"cost":{"input":0.2,"output":1.1,"cache_read":0.03,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-air"},"zai/glm-4.7-flashx":{"id":"zai/glm-4.7-flashx","name":"GLM-4.7 FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.07,"output":0.4,"cache_read":0.01,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"zai/glm-4.5v":{"id":"zai/glm-4.5v","name":"Glm 4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.6,"output":1.8,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5v"},"zai/glm-5.1":{"id":"zai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"zai/glm-4.7":{"id":"zai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7"},"zai/glm-5-turbo":{"id":"zai/glm-5-turbo","name":"GLM-5 Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24,"cache_write":0},"canonical_model_id":"zhipuai/glm-5-turbo"},"zai/glm-4.5":{"id":"zai/glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5"},"zai/glm-5.3-flash":{"id":"zai/glm-5.3-flash","name":"GLM-5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai/glm-5.2":{"id":"zai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":50000}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.05,"output":3.3,"cache_read":0.195},"canonical_model_id":"zhipuai/glm-5.2"},"zai/glm-4.7-flash":{"id":"zai/glm-4.7-flash","name":"GLM 4.7 Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.07,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"zai/glm-5":{"id":"zai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2,"cache_write":0},"canonical_model_id":"zhipuai/glm-5"},"zai/glm-4.6":{"id":"zai/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6"},"zai/glm-5.3":{"id":"zai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.7,"output":2.2,"cache_read":0.13},"canonical_model_id":"zhipuai/glm-5.3"},"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.057,"output":0.459,"cache_read":0.020357},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3-vl-30b-a3b-instruct":{"id":"qwen/qwen3-vl-30b-a3b-instruct","name":"Qwen3-VL 30B-A3B Instruct","description":"Open vision-language MoE model for image and video understanding, documents, and visual agents","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.0216},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"qwen/qwen3-30b-a3b":{"id":"qwen/qwen3-30b-a3b","name":"Qwen3 30B A3B","description":"Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.108,"output":1.076,"cache_read":0.0216},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.825,"output":2.4755,"cache_read":0.165,"cache_write":1.03125},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.4,"output":1.6},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3-max":{"id":"qwen/qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.359,"output":1.434,"cache_read":0.0718,"cache_write":0.44875},"canonical_model_id":"alibaba/qwen3-max"},"qwen/qwen-flash":{"id":"qwen/qwen-flash","name":"Qwen Flash","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.022,"output":0.216,"cache_read":0.0044,"cache_write":0.0275},"canonical_model_id":"alibaba/qwen-flash"},"qwen/qwen3.5-4b":{"id":"qwen/qwen3.5-4b","name":"Qwen3.5 4B","description":"Small Qwen3.5 multimodal model for low-cost chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-02","last_updated":"2026-03-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.04,"output":0.07},"canonical_model_id":"alibaba/qwen3.5-4b"},"qwen/qwen3-coder-plus":{"id":"qwen/qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Hosted Qwen coder for software agents, repo edits, and long-context code","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.574,"output":2.294,"cache_read":0.1148,"cache_write":0.7175},"canonical_model_id":"alibaba/qwen3-coder-plus"},"qwen/qwen3.6-35b-a3b":{"id":"qwen/qwen3.6-35b-a3b","name":"Qwen3.6 35B A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.248,"output":1.485,"cache_read":0.0496},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.115,"output":0.917,"cache_read":0.023},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3.8-2.4t-a95b":{"id":"qwen/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":1010000},"cost":{"input":2.5,"output":6.25,"cache_read":0.5},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"qwen/qwen3-coder-next":{"id":"qwen/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.15,"output":0.8,"cache_read":0.075},"canonical_model_id":"alibaba/qwen3-coder-next"},"qwen/qwen3.5-flash":{"id":"qwen/qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.029,"output":0.287,"cache_read":0.0029,"cache_write":0.03625},"canonical_model_id":"alibaba/qwen3.5-flash"},"qwen/qwen3-vl-flash":{"id":"qwen/qwen3-vl-flash","name":"Qwen3-VL Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.022,"output":0.215,"cache_read":0.0044,"cache_write":0.0275},"canonical_model_id":"alibaba/qwen3-vl-flash"},"qwen/qwen3-vl-plus":{"id":"qwen/qwen3-vl-plus","name":"Qwen3-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.143,"output":1.434,"cache_read":0.0286,"cache_write":0.17875},"canonical_model_id":"alibaba/qwen3-vl-plus"},"qwen/qwen3.6-max-preview":{"id":"qwen/qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"cost":{"input":1.31,"output":7.88},"canonical_model_id":"alibaba/qwen3.6-max-preview"},"qwen/qwen3-235b-a22b-instruct-2507":{"id":"qwen/qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.0574},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"qwen/qwen3-vl-235b-a22b-thinking":{"id":"qwen/qwen3-vl-235b-a22b-thinking","name":"Qwen3-VL 235B A22B Thinking","description":"Qwen vision-language thinking model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.287,"output":2.867,"cache_read":0.0574},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-thinking"},"qwen/qwen3-235b-a22b":{"id":"qwen/qwen3-235b-a22b","name":"Qwen3 235B A22B","description":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.287,"output":1.147,"cache_read":0.0574},"canonical_model_id":"alibaba/qwen3-235b-a22b"},"qwen/qwen3.8-max":{"id":"qwen/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.086,"output":0.688,"cache_read":0.0172},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.165,"output":0.99,"cache_read":0.0165,"cache_write":0.20625},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.172,"output":1.032,"cache_read":0.0344},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3-next-80b-a3b-thinking":{"id":"qwen/qwen3-next-80b-a3b-thinking","name":"Qwen3-Next 80B-A3B (Thinking)","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":977000,"output":128000},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3-max-preview":{"id":"qwen/qwen3-max-preview","name":"Qwen3 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-05","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"cost":{"input":1.08,"output":4.8},"canonical_model_id":"alibaba/qwen3-max-preview"},"qwen/qwen3-next-80b-a3b-instruct":{"id":"qwen/qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.1,"output":1.1,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"qwen/qwen3-32b":{"id":"qwen/qwen3-32b","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-32b"},"qwen/qwen3-coder-480b-a35b-instruct":{"id":"qwen/qwen3-coder-480b-a35b-instruct","name":"Qwen3 Coder 480B A35B Instruct","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.22,"output":1.8},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"qwen/qwen3.5-9b":{"id":"qwen/qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.09,"output":0.13},"canonical_model_id":"alibaba/qwen3.5-9b"},"qwen/qwen3.5-plus":{"id":"qwen/qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.115,"output":0.688,"cache_read":0.0115,"cache_write":0.14375},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.276,"output":1.651,"cache_read":0.0276,"cache_write":0.345},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen/qwen3-vl-235b-a22b-instruct":{"id":"qwen/qwen3-vl-235b-a22b-instruct","name":"Qwen3-VL 235B A22B Instruct","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.21,"output":1.9,"cache_read":0.105},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"qwen/qwen-plus":{"id":"qwen/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.115,"output":0.287,"cache_read":0.023,"cache_write":0.14375},"canonical_model_id":"alibaba/qwen-plus"},"qwen/qwen3-coder-flash":{"id":"qwen/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":250000},"cost":{"input":0.144,"output":0.574,"cache_read":0.0288,"cache_write":0.18},"canonical_model_id":"alibaba/qwen3-coder-flash"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.4126,"output":2.4754},"canonical_model_id":"alibaba/qwen3.6-27b"},"cohere/command-r7b-12-2024":{"id":"cohere/command-r7b-12-2024","name":"Command R7B 12-2024","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-12-13","last_updated":"2024-12-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.0375,"output":0.15},"canonical_model_id":"cohere/command-r7b-12-2024"},"cohere/command-r-plus-08-2024":{"id":"cohere/command-r-plus-08-2024","name":"Command R+ 08-2024","description":"Cohere's RAG workhorse for long-context enterprise search and tool use","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-r-plus-08-2024"},"cohere/command-r-08-2024":{"id":"cohere/command-r-08-2024","name":"Command R 08-2024","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"cohere/command-r-08-2024"},"cohere/command-a-03-2025":{"id":"cohere/command-a-03-2025","name":"Command A 03-2025","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a-03-2025"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash-Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-2.5-computer-use-preview-10-2025":{"id":"google/gemini-2.5-computer-use-preview-10-2025","name":"Gemini 2.5 Computer Use Preview (10-2025)","description":"Specialized Gemini 2.5 model for browser-control agents that automate UI tasks","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-10-07","last_updated":"2025-10-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"cost":{"input":1.25,"output":10},"canonical_model_id":"google/gemini-2.5-computer-use-preview-10-2025"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"google/gemini-4-argon"},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemini-3.1-flash-lite-preview":{"id":"google/gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.08,"output":0.45,"cache_read":0.04},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemma-3-4b-it":{"id":"google/gemma-3-4b-it","name":"Gemma 3 4B","description":"Open multimodal Gemma instruction model for efficient text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.04,"output":0.08},"canonical_model_id":"google/gemma-3-4b-it"},"google/gemini-2.5-flash-image":{"id":"google/gemini-2.5-flash-image","type":"image","name":"Gemini 2.5 Flash Image","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":2.5},"canonical_model_id":"google/gemini-2.5-flash-image"},"google/gemini-embedding-001":{"id":"google/gemini-embedding-001","type":"embedding","name":"Gemini Embedding 001","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-05","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2048,"output":4096},"cost":{"input":0.15,"output":0},"canonical_model_id":"google/gemini-embedding-001"},"google/gemini-flash-latest":{"id":"google/gemini-flash-latest","name":"Gemini Flash Latest","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-flash-latest"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash-Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.13,"output":0.4},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.09,"output":0.29},"canonical_model_id":"google/gemma-3-12b-it"},"google/gemini-flash-lite-latest":{"id":"google/gemini-flash-lite-latest","name":"Gemini Flash-Lite Latest","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-flash-lite-latest"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-3.1-flash-image":{"id":"google/gemini-3.1-flash-image","type":"image","name":"Gemini 3.1 Flash Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B It","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-3-pro-image":{"id":"google/gemini-3-pro-image","type":"image","name":"Gemini 3 Pro Image","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":65536},"cost":{"input":2,"output":12},"canonical_model_id":"google/gemini-3-pro-image"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"google/gemini-3-pro-preview":{"id":"google/gemini-3-pro-preview","name":"Gemini 3 Pro Preview","description":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3-pro-preview"},"google/gemini-embedding-2":{"id":"google/gemini-embedding-2","type":"embedding","name":"Gemini Embedding 2","description":"Multimodal embedding model mapping text, images, video, audio, and PDFs into a unified embedding space","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-11","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0.2,"output":0},"canonical_model_id":"google/gemini-embedding-2"}}},"poe":{"id":"poe","env":["POE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.poe.com/v1","name":"Poe","doc":"https://creator.poe.com/docs/external-applications/openai-compatible-api","models":{"ideogramai/ideogram-v2a-turbo":{"id":"ideogramai/ideogram-v2a-turbo","type":"image","name":"Ideogram-v2a-Turbo","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"ideogram","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-02-27","last_updated":"2025-02-27","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":150,"output":0}},"ideogramai/ideogram-v2":{"id":"ideogramai/ideogram-v2","type":"image","name":"Ideogram-v2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"ideogram","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-08-21","last_updated":"2024-08-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":150,"output":0}},"ideogramai/ideogram":{"id":"ideogramai/ideogram","type":"image","name":"Ideogram","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"ideogram","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-04-03","last_updated":"2024-04-03","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":150,"output":0}},"ideogramai/ideogram-v2a":{"id":"ideogramai/ideogram-v2a","type":"image","name":"Ideogram-v2a","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"ideogram","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-02-27","last_updated":"2025-02-27","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":150,"output":0}},"trytako/tako":{"id":"trytako/tako","name":"Tako","description":"Tool-capable chat model for instruction following and agentic application workflows","family":"tako","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-08-15","last_updated":"2024-08-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2048,"output":0}},"lumalabs/ray2":{"id":"lumalabs/ray2","type":"video","name":"Ray2","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"ray","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-02-20","last_updated":"2025-02-20","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":5000,"output":0}},"cerebras/llama-3.3-70b-cs":{"id":"cerebras/llama-3.3-70b-cs","name":"llama-3.3-70b-cs","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":0,"output":0},"status":"deprecated","canonical_model_id":"meta/llama-3.3-70b-instruct"},"cerebras/gpt-oss-120b-cs":{"id":"cerebras/gpt-oss-120b-cs","name":"GPT-OSS-120B-CS","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":0},"cost":{"input":0.3535,"output":0.7576},"canonical_model_id":"openai/gpt-oss-120b"},"cerebras/qwen3-32b-cs":{"id":"cerebras/qwen3-32b-cs","name":"qwen3-32b-cs","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":0,"output":0},"status":"deprecated","canonical_model_id":"alibaba/qwen3-32b"},"cerebras/qwen3-235b-2507-cs":{"id":"cerebras/qwen3-235b-2507-cs","name":"qwen3-235b-2507-cs","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":0,"output":0},"status":"deprecated","canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"cerebras/llama-3.1-8b-cs":{"id":"cerebras/llama-3.1-8b-cs","name":"Llama-3.1-8B-CS","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":0},"cost":{"input":0.1,"output":0.1},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"runwayml/runway":{"id":"runwayml/runway","type":"video","name":"Runway","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"runway","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-10-11","last_updated":"2024-10-11","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":256,"output":0},"status":"deprecated"},"runwayml/runway-gen-4-turbo":{"id":"runwayml/runway-gen-4-turbo","type":"video","name":"Runway-Gen-4-Turbo","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"runway","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-05-09","last_updated":"2025-05-09","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":256,"output":0},"status":"deprecated"},"elevenlabs/elevenlabs-v2.5-turbo":{"id":"elevenlabs/elevenlabs-v2.5-turbo","type":"speech","name":"ElevenLabs-v2.5-Turbo","description":"Low-latency multilingual text-to-speech model, superseded by Flash v2.5","family":"elevenlabs","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-07-19","last_updated":"2024-07-19","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":128000,"output":0},"canonical_model_id":"elevenlabs/eleven_turbo_v2_5"},"elevenlabs/elevenlabs-v3":{"id":"elevenlabs/elevenlabs-v3","type":"speech","name":"ElevenLabs-v3","description":"Expressive text-to-speech model with inline audio tags across 70+ languages","family":"elevenlabs","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-08-20","last_updated":"2025-08-20","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":128000,"output":0},"canonical_model_id":"elevenlabs/eleven_v3"},"elevenlabs/elevenlabs-music":{"id":"elevenlabs/elevenlabs-music","name":"ElevenLabs-Music","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"elevenlabs","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-08-29","last_updated":"2025-08-29","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":2000,"output":0}},"openai/chatgpt-4o-latest":{"id":"openai/chatgpt-4o-latest","name":"ChatGPT-4o-Latest","description":"Legacy model retained for compatibility with older integrations","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-08-14","last_updated":"2024-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":4.5,"output":14}},"openai/gpt-4o-mini-search":{"id":"openai/gpt-4o-mini-search","name":"GPT-4o-mini-Search","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-03-11","last_updated":"2025-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0.14,"output":0.54}},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.1364,"output":9.0909,"cache_read":0.1136},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1-mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.3636,"output":1.4545,"cache_read":0.0909},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5-Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":13.6364,"output":109.0909},"canonical_model_id":"openai/gpt-5-pro"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":0.99,"output":4},"canonical_model_id":"openai/o3-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5-Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"input":922000,"output":128000},"cost":{"input":27.2727,"output":163.6364},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-4o-search":{"id":"openai/gpt-4o-search","name":"GPT-4o-Search","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-03-11","last_updated":"2025-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":2.2727,"output":9.0909}},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5-mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2273,"output":1.8182,"cache_read":0.0227},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.5909,"output":12.7273,"cache_read":0.1591},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/o1-pro":{"id":"openai/o1-pro","name":"o1-pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2023-09","release_date":"2025-03-19","last_updated":"2025-03-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":140,"output":540},"canonical_model_id":"openai/o1-pro"},"openai/gpt-5.3-instant":{"id":"openai/gpt-5.3-instant","name":"GPT-5.3-Instant","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":111616,"output":16384},"status":"deprecated","cost":{"input":1.6,"output":13,"cache_read":0.16},"canonical_model_id":"openai/gpt-5.3-chat-latest"},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2-Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":1.6,"output":13,"cache_read":0.16},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/gpt-3.5-turbo-raw":{"id":"openai/gpt-3.5-turbo-raw","name":"GPT-3.5-Turbo-Raw","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":4524,"output":2048},"status":"deprecated","cost":{"input":0.45,"output":1.4},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4-Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":27.2727,"output":163.6364},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/o3-deep-research":{"id":"openai/o3-deep-research","name":"o3-deep-research","description":"Research model for long-horizon investigation, synthesis, and analytical reports","family":"o","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-26","last_updated":"2025-06-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":9,"output":36,"cache_read":2.2},"canonical_model_id":"openai/o3-deep-research"},"openai/gpt-3.5-turbo-instruct":{"id":"openai/gpt-3.5-turbo-instruct","name":"GPT-3.5-Turbo-Instruct","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2023-09-20","last_updated":"2023-09-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":3500,"output":1024},"status":"deprecated","cost":{"input":1.4,"output":1.8}},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5-nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.0455,"output":0.3636,"cache_read":0.0045},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.5909,"output":12.7273,"cache_read":0.1591},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"input":922000,"output":128000},"cost":{"input":4.5455,"output":27.2727,"cache_read":0.4545},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-5-codex":{"id":"openai/gpt-5-codex","name":"GPT-5-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":1.1,"output":9},"canonical_model_id":"openai/gpt-5-codex"},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.8182,"output":7.2727,"cache_read":0.4545},"canonical_model_id":"openai/o3"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.2727,"output":13.6364,"cache_read":0.2273},"canonical_model_id":"openai/gpt-5.4"},"openai/o4-mini-deep-research":{"id":"openai/o4-mini-deep-research","name":"o4-mini-deep-research","description":"Research model for long-horizon investigation, synthesis, and analytical reports","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-26","last_updated":"2025-06-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.8,"output":7.2,"cache_read":0.45},"canonical_model_id":"openai/o4-mini-deep-research"},"openai/gpt-5.3-codex-spark":{"id":"openai/gpt-5.3-codex-spark","name":"GPT-5.3-Codex-Spark","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex-spark","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":100000,"output":16384},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.3-codex-spark"},"openai/gpt-image-2":{"id":"openai/gpt-image-2","type":"image","name":"GPT-Image-2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5.0505,"output":30.303,"cache_read":1.2626},"canonical_model_id":"openai/gpt-image-2"},"openai/gpt-5.1-instant":{"id":"openai/gpt-5.1-instant","name":"GPT-5.1-Instant","description":"Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":1.1,"output":9,"cache_read":0.11},"canonical_model_id":"openai/gpt-5.1-chat-latest"},"openai/gpt-4-classic":{"id":"openai/gpt-4-classic","name":"GPT-4-Classic","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":4096},"status":"deprecated","cost":{"input":27,"output":54},"canonical_model_id":"openai/gpt-4"},"openai/gpt-4o-aug":{"id":"openai/gpt-4o-aug","name":"GPT-4o-Aug","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":2.2727,"output":9.0909,"cache_read":1.1364},"canonical_model_id":"openai/gpt-4o-2024-08-06"},"openai/gpt-image-1-mini":{"id":"openai/gpt-image-1-mini","type":"image","name":"GPT-Image-1-Mini","description":"Cost-efficient GPT Image 1 variant for image generation and editing","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"openai/gpt-image-1-mini"},"openai/sora-2-pro":{"id":"openai/sora-2-pro","type":"video","name":"Sora-2-Pro","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"sora","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"status":"deprecated"},"openai/gpt-image-1":{"id":"openai/gpt-image-1","type":"image","name":"GPT-Image-1","description":"OpenAI image model for production generation, edits, and brand-safe visual workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-04-24","last_updated":"2025-04-24","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":128000,"output":0},"status":"deprecated","canonical_model_id":"openai/gpt-image-1"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1-nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":0.09,"output":0.36,"cache_read":0.022},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o-mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":124096,"output":4096},"cost":{"input":0.1364,"output":0.5455,"cache_read":0.0682},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-image-1.5":{"id":"openai/gpt-image-1.5","type":"image","name":"gpt-image-1.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":128000,"output":0},"canonical_model_id":"openai/gpt-image-1.5"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4-Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":9,"output":27},"canonical_model_id":"openai/gpt-4-turbo"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":0.99,"output":4,"cache_read":0.25},"canonical_model_id":"openai/o4-mini"},"openai/gpt-5.2-instant":{"id":"openai/gpt-5.2-instant","name":"GPT-5.2-Instant","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":1.6,"output":13,"cache_read":0.16},"canonical_model_id":"openai/gpt-5.2-chat-latest"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5-Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":2048},"status":"deprecated","cost":{"input":0.45,"output":1.4},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/dall-e-3":{"id":"openai/dall-e-3","type":"image","name":"DALL-E-3","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"dall-e","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":800,"output":0},"status":"deprecated"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":1.8182,"output":7.2727,"cache_read":0.4545},"canonical_model_id":"openai/gpt-4.1"},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":14,"output":54},"canonical_model_id":"openai/o1"},"openai/gpt-4-classic-0314":{"id":"openai/gpt-4-classic-0314","name":"GPT-4-Classic-0314","description":"Legacy model retained for compatibility with older integrations","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-08-26","last_updated":"2024-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":4096},"status":"deprecated","cost":{"input":27,"output":54}},"openai/gpt-5.1-codex-max":{"id":"openai/gpt-5.1-codex-max","name":"GPT-5.1-Codex-Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":1.1,"output":9,"cache_read":0.11},"canonical_model_id":"openai/gpt-5.1-codex-max"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1-Codex-Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":0.22,"output":1.8,"cache_read":0.022},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.1364,"output":9.0909,"cache_read":0.1136},"canonical_model_id":"openai/gpt-5"},"openai/o3-pro":{"id":"openai/o3-pro","name":"o3-pro","description":"High-effort o3 tier for difficult technical reasoning and careful answers","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":18.1818,"output":72.7273},"canonical_model_id":"openai/o3-pro"},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT-5.1-Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":1.1,"output":9,"cache_read":0.11},"canonical_model_id":"openai/gpt-5.1-codex"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2-Pro","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":19.0909,"output":152.7273},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/gpt-5-chat":{"id":"openai/gpt-5-chat","name":"GPT-5-Chat","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":272000,"output":16384},"status":"deprecated","cost":{"input":1.1,"output":9,"cache_read":0.11},"canonical_model_id":"openai/gpt-5-chat-latest"},"openai/o3-mini-high":{"id":"openai/o3-mini-high","name":"o3-mini-high","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":0.99,"output":4},"canonical_model_id":"openai/o3-mini"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4-Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.6818,"output":4.0909,"cache_read":0.0682},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/sora-2":{"id":"openai/sora-2","type":"video","name":"Sora-2","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"sora","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"status":"deprecated"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4-Nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.1818,"output":1.1364,"cache_read":0.0182},"canonical_model_id":"openai/gpt-5.4-nano"},"anthropic/claude-opus-4.1":{"id":"anthropic/claude-opus-4.1","name":"Claude-Opus-4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":31999}],"tool_call":true,"temperature":false,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":196608,"output":32000},"status":"deprecated","cost":{"input":13,"output":64,"cache_read":1.3,"cache_write":16},"canonical_model_id":"anthropic/claude-opus-4-1"},"anthropic/claude-sonnet-3.5":{"id":"anthropic/claude-sonnet-3.5","name":"Claude-Sonnet-3.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2024-04-30","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":189096,"output":8192},"status":"deprecated","cost":{"input":2.6,"output":13,"cache_read":0.26,"cache_write":3.2},"canonical_model_id":"anthropic/claude-3-5-sonnet-20241022"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude-Opus-4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":128000},"cost":{"input":4.2929,"output":21.4646,"cache_read":0.4293,"cache_write":5.3662},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude-Sonnet-4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":983040,"output":128000},"cost":{"input":2.5758,"output":12.8788,"cache_read":0.2576,"cache_write":3.2197},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude-Haiku-4.5","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":63999}],"tool_call":true,"temperature":false,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":192000,"output":64000},"cost":{"input":0.8586,"output":4.2929,"cache_read":0.0859,"cache_write":1.0732},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude-Opus-4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"temperature":false,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":983040,"output":128000},"cost":{"input":4.2929,"output":21.4646,"cache_read":0.4293,"cache_write":5.3662},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-4":{"id":"anthropic/claude-opus-4","name":"Claude-Opus-4","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":192512,"output":28672},"status":"deprecated","cost":{"input":13,"output":64,"cache_read":1.3,"cache_write":16},"canonical_model_id":"anthropic/claude-opus-4-0"},"anthropic/claude-sonnet-4":{"id":"anthropic/claude-sonnet-4","name":"Claude-Sonnet-4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":983040,"output":64000},"status":"deprecated","cost":{"input":2.6,"output":13,"cache_read":0.26,"cache_write":3.2},"canonical_model_id":"anthropic/claude-sonnet-4-0"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude-Opus-4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":0,"max":63999}],"tool_call":true,"temperature":false,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":196608,"output":64000},"cost":{"input":4.2929,"output":21.4646,"cache_read":0.4293,"cache_write":5.3662},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude-Sonnet-4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":31999}],"tool_call":true,"temperature":false,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":983040,"output":32768},"cost":{"input":2.5758,"output":12.8788,"cache_read":0.2576,"cache_write":3.2197},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude-Opus-4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":128000},"cost":{"input":4.2929,"output":21.4646,"cache_read":0.4293,"cache_write":5.3662},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-sonnet-3.7":{"id":"anthropic/claude-sonnet-3.7","name":"Claude-Sonnet-3.7","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2024-10-31","release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":196608,"output":128000},"status":"deprecated","cost":{"input":2.6,"output":13,"cache_read":0.26,"cache_write":3.2},"canonical_model_id":"anthropic/claude-3-7-sonnet-20250219"},"anthropic/claude-sonnet-3.5-june":{"id":"anthropic/claude-sonnet-3.5-june","name":"Claude-Sonnet-3.5-June","description":"Legacy model retained for compatibility with older integrations","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-11-18","last_updated":"2024-11-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":189096,"output":8192},"status":"deprecated","cost":{"input":2.6,"output":13,"cache_read":0.26,"cache_write":3.2}},"anthropic/claude-haiku-3":{"id":"anthropic/claude-haiku-3","name":"Claude-Haiku-3","description":"Legacy model retained for compatibility with older integrations","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2023-08-31","release_date":"2024-03-13","last_updated":"2024-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":189096,"output":8192},"status":"deprecated","cost":{"input":0.21,"output":1.1,"cache_read":0.021,"cache_write":0.26},"canonical_model_id":"anthropic/claude-3-haiku-20240307"},"anthropic/claude-haiku-3.5":{"id":"anthropic/claude-haiku-3.5","name":"Claude-Haiku-3.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2024-07-31","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":189096,"output":8192},"status":"deprecated","cost":{"input":0.68,"output":3.4,"cache_read":0.068,"cache_write":0.85},"canonical_model_id":"anthropic/claude-3-5-haiku-20241022"},"fireworks-ai/kimi-k2.5-fw":{"id":"fireworks-ai/kimi-k2.5-fw","name":"Kimi-K2.5-FW","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":245760,"output":16384},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"moonshotai/kimi-k2.5"},"empiriolabs/deepseek-v4-pro-el":{"id":"empiriolabs/deepseek-v4-pro-el","name":"DeepSeek-V4-Pro-EL","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":384000},"cost":{"input":1.67,"output":3.33},"canonical_model_id":"deepseek/deepseek-v4-pro"},"empiriolabs/deepseek-v4-flash-el":{"id":"empiriolabs/deepseek-v4-flash-el","name":"DeepSeek-V4-Flash-EL","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":384000},"cost":{"input":0.14,"output":0.28},"canonical_model_id":"deepseek/deepseek-v4-flash"},"novita/kimi-k2-thinking":{"id":"novita/kimi-k2-thinking","name":"kimi-k2-thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":false,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":0},"status":"deprecated","canonical_model_id":"moonshotai/kimi-k2-thinking"},"novita/glm-4.7-n":{"id":"novita/glm-4.7-n","name":"glm-4.7-n","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":false,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":205000,"output":131072},"canonical_model_id":"zhipuai/glm-4.7"},"novita/kimi-k2.6":{"id":"novita/kimi-k2.6","name":"Kimi-K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":262144},"cost":{"input":0.9596,"output":4.0404,"cache_read":0.1616},"canonical_model_id":"moonshotai/kimi-k2.6"},"novita/deepseek-v3.2":{"id":"novita/deepseek-v3.2","name":"DeepSeek-V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":0},"cost":{"input":0.2626,"output":0.3838,"cache_read":0.1313},"canonical_model_id":"deepseek/deepseek-v3.2"},"novita/minimax-m2.1":{"id":"novita/minimax-m2.1","name":"minimax-m2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":false,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":205000,"output":131072},"cost":{"input":0.303,"output":1.2121,"cache_read":0.1515},"canonical_model_id":"minimax/MiniMax-M2.1"},"novita/glm-4.7":{"id":"novita/glm-4.7","name":"glm-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":205000,"output":131072},"status":"deprecated","canonical_model_id":"zhipuai/glm-4.7"},"novita/kimi-k2.5":{"id":"novita/kimi-k2.5","name":"Kimi-K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":262144},"status":"deprecated","cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"novita/glm-4.7-flash":{"id":"novita/glm-4.7-flash","name":"glm-4.7-flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":false,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":65500},"cost":{"input":0.0707,"output":0.404},"canonical_model_id":"zhipuai/glm-4.7-flash"},"novita/glm-5":{"id":"novita/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":205000,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"novita/glm-4.6":{"id":"novita/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":false,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":0,"output":0},"cost":{"input":0.6061,"output":2.2222,"cache_read":0.1111},"canonical_model_id":"zhipuai/glm-4.6"},"novita/glm-4.6v":{"id":"novita/glm-4.6v","name":"glm-4.6v","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":false,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":32768},"canonical_model_id":"zhipuai/glm-4.6v"},"topazlabs-co/topazlabs":{"id":"topazlabs-co/topazlabs","type":"image","name":"TopazLabs","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"topazlabs","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":204,"output":0}},"stabilityai/stablediffusionxl":{"id":"stabilityai/stablediffusionxl","type":"image","name":"StableDiffusionXL","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"stable-diffusion","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2023-07-09","last_updated":"2023-07-09","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":200,"output":0}},"xai/grok-4-fast-non-reasoning":{"id":"xai/grok-4-fast-non-reasoning","name":"Grok-4-Fast-Non-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":128000},"status":"deprecated","cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4-fast-non-reasoning"},"xai/grok-4.1-fast-reasoning":{"id":"xai/grok-4.1-fast-reasoning","name":"Grok-4.1-Fast-Reasoning","description":"xAI's fast agentic tool-calling model with a 2M context window and built-in reasoning","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"status":"deprecated","canonical_model_id":"xai/grok-4.1-fast-reasoning"},"xai/grok-4.20-multi-agent":{"id":"xai/grok-4.20-multi-agent","name":"Grok-4.20-Multi-Agent","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2026-03-13","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":0},"cost":{"input":1.2626,"output":2.5253,"cache_read":0.202}},"xai/grok-4-fast-reasoning":{"id":"xai/grok-4-fast-reasoning","name":"Grok-4-Fast-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":128000},"status":"deprecated","cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4-fast-reasoning"},"xai/grok-code-fast-1":{"id":"xai/grok-code-fast-1","name":"Grok Code Fast 1","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.2,"output":1.5,"cache_read":0.02},"canonical_model_id":"xai/grok-code-fast-1"},"xai/grok-4.1-fast-non-reasoning":{"id":"xai/grok-4.1-fast-non-reasoning","name":"Grok-4.1-Fast-Non-Reasoning","description":"xAI's fast agentic tool-calling model with a 2M context window; non-reasoning variant for low-latency responses","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"status":"deprecated","canonical_model_id":"xai/grok-4.1-fast"},"xai/grok-3-mini":{"id":"xai/grok-3-mini","name":"Grok 3 Mini","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2025-04-11","last_updated":"2025-04-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.3,"output":0.5,"cache_read":0.075}},"xai/grok-4":{"id":"xai/grok-4","name":"Grok-4","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-07-10","last_updated":"2025-07-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.75}},"xai/grok-3":{"id":"xai/grok-3","name":"Grok 3","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-04-11","last_updated":"2025-04-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":3,"output":15,"cache_read":0.75}},"poetools/claude-code":{"id":"poetools/claude-code","name":"claude-code","description":"Claude model for careful reasoning, writing, coding, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":false,"release_date":"2025-11-27","last_updated":"2025-11-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0}},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini-3.8-Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.7576,"output":3.7879,"cache_read":0.0758},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini-3.1-Flash-Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.2525,"output":1.5152},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini-2.5-Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1065535,"output":65535},"status":"deprecated","cost":{"input":0.21,"output":1.8,"cache_read":0.021},"canonical_model_id":"google/gemini-2.5-flash"},"google/imagen-4":{"id":"google/imagen-4","type":"image","name":"Imagen-4","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"imagen","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":480,"output":0}},"google/veo-2":{"id":"google/veo-2","type":"video","name":"Veo-2","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-12-02","last_updated":"2024-12-02","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":0},"status":"deprecated"},"google/veo-3.1":{"id":"google/veo-3.1","type":"video","name":"Veo-3.1","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":0}},"google/gemini-3-flash":{"id":"google/gemini-3-flash","name":"Gemini-3-Flash","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.404,"output":2.4242,"cache_read":0.0404},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-2.0-flash":{"id":"google/gemini-2.0-flash","name":"Gemini-2.0-Flash","description":"Earlier Gemini Flash workhorse for responsive multimodal apps and tool use","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2024-06","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":990000,"output":8192},"status":"deprecated","cost":{"input":0.1,"output":0.42},"canonical_model_id":"google/gemini-2.0-flash"},"google/veo-3.1-fast":{"id":"google/veo-3.1-fast","type":"video","name":"Veo-3.1-Fast","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":0}},"google/imagen-3-fast":{"id":"google/imagen-3-fast","type":"image","name":"Imagen-3-Fast","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"imagen","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-10-17","last_updated":"2024-10-17","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":480,"output":0},"status":"deprecated"},"google/imagen-4-ultra":{"id":"google/imagen-4-ultra","type":"image","name":"Imagen-4-Ultra","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"imagen","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-05-24","last_updated":"2025-05-24","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":480,"output":0},"status":"deprecated"},"google/gemini-2.0-flash-lite":{"id":"google/gemini-2.0-flash-lite","name":"Gemini-2.0-Flash-Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2024-06","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":990000,"output":8192},"status":"deprecated","cost":{"input":0.052,"output":0.21},"canonical_model_id":"google/gemini-2.0-flash-lite"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini-3.5-Flash-Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.303,"output":2.5253,"cache_read":0.0303},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/imagen-4-fast":{"id":"google/imagen-4-fast","type":"image","name":"Imagen-4-Fast","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"imagen","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-06-25","last_updated":"2025-06-25","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":480,"output":0}},"google/gemini-3-pro":{"id":"google/gemini-3-pro","name":"Gemini-3-Pro","description":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":1.6,"output":9.6,"cache_read":0.16},"canonical_model_id":"google/gemini-3-pro-preview"},"google/gemma-4-31b":{"id":"google/gemma-4-31b","name":"Gemma-4-31B","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini-3.6-Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.7576,"output":3.7879,"cache_read":0.0758},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini-3.7-Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.7576,"output":3.7879,"cache_read":0.0758},"canonical_model_id":"google/gemini-3.7-flash"},"google/veo-3-fast":{"id":"google/veo-3-fast","type":"video","name":"Veo-3-Fast","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-10-13","last_updated":"2025-10-13","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":0},"status":"deprecated"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini-3.5-Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5152,"output":9.0909,"cache_read":0.1515},"canonical_model_id":"google/gemini-3.5-flash"},"google/nano-banana":{"id":"google/nano-banana","type":"image","name":"Nano-Banana","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":0},"status":"deprecated","cost":{"input":0.21,"output":1.8,"cache_read":0.021},"canonical_model_id":"google/gemini-2.5-flash-image"},"google/gemini-omni-1.1-flash":{"id":"google/gemini-omni-1.1-flash","type":"video","name":"Gemini-Omni-1.1-Flash","description":"Video generation and editing model for scene extension, frame interpolation, and reference-guided multimodal control","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":false,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":131072,"output":57920},"cost":{"input":1.5152,"output":9.0909},"canonical_model_id":"google/gemini-omni-1.1-flash"},"google/veo-3":{"id":"google/veo-3","type":"video","name":"Veo-3","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-05-21","last_updated":"2025-05-21","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":0},"status":"deprecated"},"google/gemini-deep-research":{"id":"google/gemini-deep-research","name":"gemini-deep-research","description":"Legacy model retained for compatibility with older integrations","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":0},"status":"deprecated","cost":{"input":1.6,"output":9.6}},"google/lyria":{"id":"google/lyria","name":"Lyria","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"lyria","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2025-06-04","last_updated":"2025-06-04","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"google/nano-banana-pro":{"id":"google/nano-banana-pro","type":"image","name":"Nano-Banana-Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":65536,"output":0},"cost":{"input":2.0202,"output":12.1212,"cache_read":0.202},"canonical_model_id":"google/gemini-3-pro-image-preview"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini-2.5-Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1024000,"output":64000},"status":"deprecated","cost":{"input":0.07,"output":0.28},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini-2.5-Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":32768}],"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1065535,"output":65535},"status":"deprecated","cost":{"input":0.87,"output":7,"cache_read":0.087},"canonical_model_id":"google/gemini-2.5-pro"},"google/imagen-3":{"id":"google/imagen-3","type":"image","name":"Imagen-3","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"imagen","attachment":false,"reasoning":false,"tool_call":true,"temperature":false,"release_date":"2024-10-15","last_updated":"2024-10-15","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":480,"output":0},"status":"deprecated"},"google/gemini-3.1-pro":{"id":"google/gemini-3.1-pro","name":"Gemini-3.1-Pro","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2.0202,"output":12.1212,"cache_read":0.202},"canonical_model_id":"google/gemini-3.1-pro-preview"}}},"berget":{"id":"berget","env":["BERGET_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.berget.ai/v1","name":"Berget.AI","doc":"https://api.berget.ai","models":{"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":327680,"output":32768},"cost":{"input":3.3645,"output":16.8225,"cache_read":1.00935},"canonical_model_id":"moonshotai/kimi-k3"},"mistralai/Mistral-Small-3.2-24B-Instruct-2506":{"id":"mistralai/Mistral-Small-3.2-24B-Instruct-2506","name":"Mistral Small 3.2 24B Instruct 2506","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":8192},"cost":{"input":0.33645,"output":0.33645},"canonical_model_id":"mistral/mistral-small-3.2-24b-instruct"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":32768},"status":"deprecated","cost":{"input":1.54,"output":4.84},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":16384},"cost":{"input":0.29,"output":0.58},"canonical_model_id":"zhipuai/glm-5.3-flash"},"Qwen/Qwen3.8-27B-FP8":{"id":"Qwen/Qwen3.8-27B-FP8","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.46,"output":3.48},"canonical_model_id":"alibaba/qwen3.8-27b"},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B Instruct","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["audio","image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.280375,"output":0.56075},"canonical_model_id":"google/gemma-4-31b-it"}}},"tinfoil":{"id":"tinfoil","env":["TINFOIL_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.tinfoil.sh/v1","name":"Tinfoil","doc":"https://docs.tinfoil.sh","models":{"gpt-oss-safeguard-120b":{"id":"gpt-oss-safeguard-120b","name":"gpt-oss-safeguard-120b","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-safeguard-120b"},"llama3-3-70b":{"id":"llama3-3-70b","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":1.75,"output":2.75},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"gpt-oss-120b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":4,"output":20,"cache_read":0.8},"canonical_model_id":"moonshotai/kimi-k3"},"deepseek-v4-1-flash":{"id":"deepseek-v4-1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"status":"beta","cost":{"input":0.65,"output":1.45,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"glm-5-3":{"id":"glm-5-3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.8,"output":5.75,"cache_read":0.45},"canonical_model_id":"zhipuai/glm-5.3"},"gemma4-31b":{"id":"gemma4-31b","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.4,"output":1},"canonical_model_id":"google/gemma-4-31b-it"},"nomic-embed-text":{"id":"nomic-embed-text","name":"Nomic Embed Text v1.5","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2024-02","last_updated":"2024-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":768},"cost":{"input":0.05,"output":0}}}},"inception":{"id":"inception","env":["INCEPTION_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.inceptionlabs.ai/v1/","name":"Inception","doc":"https://docs.inceptionlabs.ai/get-started/models","models":{"mercury-2.5":{"id":"mercury-2.5","name":"Mercury 2.5","description":"Mercury 2.5 is the fastest reasoning LLM, and the latest diffusion LLM (dLLM) from Inception","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-11-01","release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":260000,"output":65536},"cost":{"input":0.2,"output":0.75,"cache_read":0.02},"canonical_model_id":"inception/mercury-2.5"},"mercury-3":{"id":"mercury-3","name":"Mercury 3","description":"Inception's diffusion reasoning model for chat, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-16","last_updated":"2026-10-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1,"output":0.35,"cache_read":0.01},"canonical_model_id":"inception/mercury-3"},"mercury-2":{"id":"mercury-2","name":"Mercury 2","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":50000},"cost":{"input":0.25,"output":0.75,"cache_read":0.025},"canonical_model_id":"inception/mercury-2"},"mercury-edit-2":{"id":"mercury-edit-2","name":"Mercury Edit 2","description":"Code editing dLLM for autocomplete (FIM) and next-edit suggestions","family":"mercury","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-30","last_updated":"2026-03-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":8192},"cost":{"input":0.25,"output":0.75,"cache_read":0.025},"canonical_model_id":"inception/mercury-edit-2"}}},"kilo":{"id":"kilo","env":["KILO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.kilo.ai/api/gateway","name":"Kilo Gateway","doc":"https://kilo.ai","models":{"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It is designed to keep track of information across extended tasks, work through...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.3-contributor":{"id":"meta/muse-spark-1.3-contributor","name":"Meta: Muse Spark 1.3 Contributor","description":"Muse Spark 1.3 Contributor is the cost-efficient contributor tier of Meta’s multimodal reasoning model for experimentation, learning, and early-stage agentic, multi-agent, and coding workflows. It is designed to track information...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/muse-glimmer-30b":{"id":"meta/muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer 30B is a dense, open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark and optimized for autonomous agents on consumer hardware. It is suited for long-horizon...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.3,"output":1.2,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark 1.1 is a multimodal reasoning model from Meta, built for agentic tasks. It accepts text, images, video, audio, and PDF documents and returns text, with a 1M-token context...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a reasoning model from Meta, designed for complex agentic tasks. It accepts text, images, video, audio, and PDF documents, returns text, and offers a 1M-token context...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"meta/muse-wonder":{"id":"meta/muse-wonder","name":"Meta: Muse Wonder","description":"Muse Wonder is a frontier reasoning model from Meta, designed for complex personal and professional agentic tasks. It accepts text, images, video, audio, and PDF documents and returns text, with...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":4,"output":20,"cache_read":0.2}},"meta/muse-spark-1.2-contributor":{"id":"meta/muse-spark-1.2-contributor","name":"Meta: Muse Spark 1.2 Contributor","description":"Muse Spark 1.2 contributor tier is a reasoning model from Meta designed for developers who want to start building at an even lower cost. It’s meaningfully cheaper than Muse Spark...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-08-21","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/muse-spark-open":{"id":"meta/muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is a coding-optimized open-weight model. Trained specifically for agentic coding workflows, it delivers higher first-attempt accuracy, reliable tool calling, and deep codebase understanding across the repos developers...","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"~unbiased/pareto-latest":{"id":"~unbiased/pareto-latest","name":"Pareto Latest","description":"This model always redirects to the latest model in the Pareto family.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.8,"output":3.2,"cache_read":0.03}},"sakana/sakana-namazu":{"id":"sakana/sakana-namazu","name":"Sakana Namazu","description":"Sakana Namazu is a Japanese-specialized reasoning model from Sakana AI, based on Kimi K2.6 with additional training for Japanese language and business contexts. It is suited for Japanese instruction following,...","family":"sakana-namazu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.95,"output":4,"cache_read":0.15},"canonical_model_id":"sakana/sakana-namazu"},"sakana/fugu-ultra-v2":{"id":"sakana/fugu-ultra-v2","name":"Fugu Ultra v2","description":"Fugu Ultra v2 is the higher-performance model in Sakana AI's Fugu family. Rather than a single monolithic model, Fugu is a learned multi-agent orchestration system: a language model trained to...","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-08-28","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"sakana/fugu-ultra-v2"},"sakana/fugu-max":{"id":"sakana/fugu-max","name":"Fugu Max","description":"Fugu Max is the cost-performance model in Sakana AI's Fugu family. Rather than a single monolithic model, Fugu is a learned multi-agent orchestration system: a language model trained to route...","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"sakana/fugu-max"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Fugu Ultra is the higher-performance model in Sakana AI's Fugu family. Rather than a single monolithic model, Fugu is a learned multi-agent orchestration system: a language model trained to route...","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"sakana/fugu-ultra"},"z-ai/glm-4.5-air":{"id":"z-ai/glm-4.5-air","name":"GLM-4.5-Air","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.13,"output":0.85,"cache_read":0.025},"canonical_model_id":"zhipuai/glm-4.5-air"},"z-ai/glm-4.5v":{"id":"z-ai/glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":0.6,"output":1.8,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5v"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-5.3-prime":{"id":"z-ai/glm-5.3-prime","name":"Z.ai: GLM 5.3 Prime","description":"GLM-5.3-Prime is the high-speed variant of Z.ai's GLM-5.3, inheriting its full capabilities while delivering 1.5–2× the output throughput through inference acceleration. It supports text input and output with a 1M-token...","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.8,"output":8.8,"cache_read":0.56}},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.54,"output":1.98,"cache_read":0.099},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM-5-Turbo","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"z-ai/glm-4.5":{"id":"z-ai/glm-4.5","name":"GLM-4.5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks. Its hybrid sparse and linear attention architecture maintains accurate long-context behavior while...","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048575,"output":943717},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"GLM 5.2 is a large-scale reasoning model from Z.ai. It supports text input and output with a 1M-token context window, and is suited for long-horizon agent workflows, project-level software engineering,...","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-4.7-flash":{"id":"z-ai/glm-4.7-flash","name":"GLM-4.7-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.0605,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"z-ai/glm-5.3-flashx":{"id":"z-ai/glm-5.3-flashx","name":"Z.ai: GLM 5.3 FlashX","description":"GLM-5.3-FlashX is the high-speed variant of Z.ai's GLM-5.3-Flash, a native multimodal model delivering inference speeds of up to 200 tokens/s. Built on the same hybrid sparse and linear attention architecture...","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.37,"output":1.25,"cache_read":0.09}},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-4.6":{"id":"z-ai/glm-4.6","name":"GLM-4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.55,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"GLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks. It supports text input and output with a 1M-token context window, and improves...","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-5v-turbo":{"id":"z-ai/glm-5v-turbo","name":"GLM-5V-Turbo","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"z-ai/glm-4.6v":{"id":"z-ai/glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.055},"canonical_model_id":"zhipuai/glm-4.6v"},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.6,"output":2.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.65,"output":3.41,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 Instruct 0905","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.6,"output":2.5},"canonical_model_id":"moonshotai/kimi-k2-0905"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"MoonshotAI: Kimi K2.7 Code is a coding-focused model in Moonshot AI's Kimi K2 family, built to complete end-to-end programming tasks reliably over long contexts. It uses a native multimodal mixture-of-experts...","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.6712,"output":3.35,"cache_read":0.18},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI. It is suited for complex coding, knowledge work, and long-horizon agentic workflows, and is particularly strong at...","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.6,"output":3},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/kimi-k2":{"id":"moonshotai/kimi-k2","name":"Kimi K2 0711","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-12-31","release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.57,"output":2.3},"canonical_model_id":"moonshotai/kimi-k2"},"mistralai/devstral-2512":{"id":"mistralai/devstral-2512","name":"Devstral 2","description":"Devstral 2 is a state-of-the-art open-source model by Mistral AI specializing in agentic coding. It is a 123B-parameter dense transformer model supporting a 256K context window. Devstral 2 supports exploring...","family":"devstral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/devstral-2512"},"mistralai/mistral-saba":{"id":"mistralai/mistral-saba","name":"Saba","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-02-17","last_updated":"2025-02-17","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":26214},"cost":{"input":0.2,"output":0.6,"cache_read":0.02},"canonical_model_id":"mistral/mistral-saba"},"mistralai/ministral-3b-2512":{"id":"mistralai/ministral-3b-2512","name":"Ministral 3 3B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":104857},"cost":{"input":0.1,"output":0.1,"cache_read":0.01},"canonical_model_id":"mistral/ministral-3b-2512"},"mistralai/mixtral-8x22b-instruct":{"id":"mistralai/mixtral-8x22b-instruct","name":"Mixtral 8x22B Instruct","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-01-31","release_date":"2024-04-17","last_updated":"2024-04-17","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":52428},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mixtral-8x22b-instruct"},"mistralai/mistral-medium-3":{"id":"mistralai/mistral-medium-3","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/mistral-medium-3"},"mistralai/mistral-small-3.1-24b-instruct":{"id":"mistralai/mistral-small-3.1-24b-instruct","name":"Mistral Small 3.1 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":102400},"cost":{"input":0.351,"output":0.555},"canonical_model_id":"mistral/mistral-small-3.1-24b-instruct"},"mistralai/ministral-8b-2512":{"id":"mistralai/ministral-8b-2512","name":"Ministral 3 8B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.15,"output":0.15,"cache_read":0.015},"canonical_model_id":"mistral/ministral-8b-2512"},"mistralai/mistral-small-2603":{"id":"mistralai/mistral-small-2603","name":"Mistral Small 4","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"mistral/mistral-small-2603"},"mistralai/voxtral-small-24b-2507":{"id":"mistralai/voxtral-small-24b-2507","name":"Voxtral Small 24B 2507","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":26214},"cost":{"input":0.1,"output":0.3,"cache_read":0.01},"canonical_model_id":"mistral/voxtral-small-24b-2507"},"mistralai/mistral-large-4-0":{"id":"mistralai/mistral-large-4-0","name":"Mistral: Mistral Large 4","description":"Mistral Large 4 is a frontier multimodal (text and image input) model from Mistral AI built for reasoning, coding, and agentic workloads. It offers a 512K-token context window with up...","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":262144},"cost":{"input":1.36,"output":4.18,"cache_read":0.14},"canonical_model_id":"mistral/mistral-large-4"},"mistralai/codestral-2508":{"id":"mistralai/codestral-2508","name":"Codestral 25.08","description":"Mistral coding model for code completion, generation, and developer workflows","family":"codestral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":204800},"cost":{"input":0.3,"output":0.9,"cache_read":0.03},"canonical_model_id":"mistral/codestral-2508"},"mistralai/mistral-medium-3.1":{"id":"mistralai/mistral-medium-3.1","name":"Mistral Medium 3.1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-30","release_date":"2025-08-12","last_updated":"2025-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/mistral-medium-3.1"},"mistralai/ministral-14b-2512":{"id":"mistralai/ministral-14b-2512","name":"Ministral 3 14B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.2,"output":0.2,"cache_read":0.02},"canonical_model_id":"mistral/ministral-14b-2512"},"mistralai/mistral-nemo":{"id":"mistralai/mistral-nemo","name":"Mistral Nemo","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.15,"output":0.15,"cache_read":0.015},"canonical_model_id":"mistral/mistral-nemo"},"mistralai/mistral-large-2407":{"id":"mistralai/mistral-large-2407","name":"Mistral Large 2407","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-03-31","release_date":"2024-07-24","last_updated":"2024-07-24","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":104857},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mistral-large-2407"},"mistralai/mistral-large":{"id":"mistralai/mistral-large","name":"Mistral Large","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11-30","release_date":"2024-02-26","last_updated":"2024-02-26","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":102400},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mistral-large"},"mistralai/mistral-small-3.2-24b-instruct":{"id":"mistralai/mistral-small-3.2-24b-instruct","name":"Mistral Small 3.2 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.1,"output":0.3,"cache_read":0.01},"canonical_model_id":"mistral/mistral-small-3.2-24b-instruct"},"mistralai/mistral-medium-3-5":{"id":"mistralai/mistral-medium-3-5","name":"Mistral Medium 3.5","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":209715},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"mistral/mistral-medium-3-5"},"mistralai/mistral-small-24b-instruct-2501":{"id":"mistralai/mistral-small-24b-instruct-2501","name":"Mistral Small 3","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-01-30","last_updated":"2025-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.05,"output":0.08},"canonical_model_id":"mistral/mistral-small-24b-instruct-2501"},"mistralai/mistral-large-2512":{"id":"mistralai/mistral-large-2512","name":"Mistral Large 3","description":"Mistral Large 3 2512 is Mistral’s most capable model to date, featuring a sparse mixture-of-experts architecture with 41B active parameters (675B total), and released under the Apache 2.0 license.","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"mistral/mistral-large-2512"},"bytedance-seed/seed-2-1-turbo":{"id":"bytedance-seed/seed-2-1-turbo","name":"ByteDance Seed: Seed 2.1 Turbo","description":"Seed 2.1 Turbo is a multimodal model from ByteDance Seed for coding and long-horizon agent workflows. It is suited for end-to-end software delivery, multi-step task execution, and understanding visual and...","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0.5,"output":2.5},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"bytedance-seed/seed-1.6-flash":{"id":"bytedance-seed/seed-1.6-flash","name":"ByteDance Seed: Seed 1.6 Flash (retires Nov 11)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.075,"output":0.3},"canonical_model_id":"bytedance-seed/seed-1-6-flash"},"bytedance-seed/seed-1.6":{"id":"bytedance-seed/seed-1.6","name":"ByteDance Seed: Seed 1.6 (retires Nov 11)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.25,"output":2},"canonical_model_id":"bytedance-seed/seed-1-6"},"bytedance-seed/seed-2.0-mini":{"id":"bytedance-seed/seed-2.0-mini","name":"Seed 2.0 Mini","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.1,"output":0.4},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"bytedance-seed/seed-2.0-code":{"id":"bytedance-seed/seed-2.0-code","name":"Seed 2.0 Code","description":"Seed 2.0 Code is a model from ByteDance Seed optimized for agentic coding. It is suited for frontend development, multilingual programming tasks, and coding-agent workflows in tools such as Claude...","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"status":"deprecated","cost":{"input":0.5,"output":3},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"bytedance-seed/seed-2.0-lite":{"id":"bytedance-seed/seed-2.0-lite","name":"Seed 2.0 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.25,"output":2},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"gryphe/mythomax-l2-13b":{"id":"gryphe/mythomax-l2-13b","name":"MythoMax 13B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2023-08-10","last_updated":"2023-08-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":3686},"cost":{"input":0.08,"output":0.11},"canonical_model_id":"gryphe/mythomax-l2-13b"},"undi95/remm-slerp-l2-13b":{"id":"undi95/remm-slerp-l2-13b","name":"ReMM SLERP 13B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-06-30","release_date":"2023-07-22","last_updated":"2023-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":6144,"output":5529},"cost":{"input":0.35,"output":0.65},"canonical_model_id":"undi95/remm-slerp-l2-13b"},"kilo-auto/small":{"id":"kilo-auto/small","name":"Auto Small","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"1970-01-01","last_updated":"1970-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.05,"output":0.4,"reasoning":0,"cache_read":0.005}},"kilo-auto/free":{"id":"kilo-auto/free","name":"Auto Free","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"1970-01-01","last_updated":"1970-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32768},"cost":{"input":0,"output":0,"reasoning":0,"cache_read":0,"cache_write":0}},"kilo-auto/efficient":{"id":"kilo-auto/efficient","name":"Auto Efficient","description":"Routes each request to the cheapest model that gets the job done, based on continuously benchmarked accuracy and cost.","family":"auto","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"1970-01-01","last_updated":"1970-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.325,"output":1.95,"reasoning":0,"cache_read":0.0325,"cache_write":0.40625}},"kilo-auto/frontier":{"id":"kilo-auto/frontier","name":"Auto Frontier","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"1970-01-01","last_updated":"1970-01-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"reasoning":0,"cache_read":0.5,"cache_write":6.25}},"kilo-auto/balanced":{"id":"kilo-auto/balanced","name":"Auto Balanced","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"1970-01-01","last_updated":"1970-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.325,"output":1.95,"reasoning":0,"cache_read":0.0325,"cache_write":0.40625}},"inclusionai/ling-3.0-flash":{"id":"inclusionai/ling-3.0-flash","name":"inclusionAI: Ling 3.0 Flash","description":"*Ling-3.0-flash* is a *124B-parameter Mixture-of-Experts (MoE) model*, with approximately *5.1B parameters activated per token*. The model is designed with *token efficiency and production-scale agentic inference* as key priorities, enabling developers...","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012},"canonical_model_id":"inclusionai/ling-3.0-flash"},"inclusionai/ling-3.0-flash-vl":{"id":"inclusionai/ling-3.0-flash-vl","name":"inclusionAI: Ling 3.0 Flash VL","description":"Ling 3.0 Flash VL builds on Ling 3.0 Flash (124B total / 5.5B active MoE from InclusionAI), further strengthening its language capabilities while adding native visual perception and advanced visual...","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012},"canonical_model_id":"inclusionai/ling-3.0-flash-vl"},"inclusionai/ling-3.1-flash":{"id":"inclusionai/ling-3.1-flash","name":"inclusionAI: Ling 3.1 Flash","description":"Ling 3.1 Flash is a hybrid reasoning mixture-of-experts model from inclusionAI, with 25B active parameters out of 560B total.","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.3,"output":0.9,"cache_read":0.06},"canonical_model_id":"inclusionai/ling-3.1-flash"},"liquid/lfm-2.5-2.6b:free":{"id":"liquid/lfm-2.5-2.6b:free","name":"LiquidAI: LFM2.5-2.6B (free)","description":"LFM2.5-2.6B is a compact reasoning model from Liquid AI. It is suited for agent workflows, data extraction, RAG, and long-context processing. Liquid advises against using it for agentic coding or...","family":"liquid","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":8192},"cost":{"input":0,"output":0}},"liquid/lfm-2.5-24b-a2b":{"id":"liquid/lfm-2.5-24b-a2b","name":"LiquidAI: LFM2.5-24B-A2B","description":"LFM2.5-24B-A2B is a mixture-of-experts reasoning model from Liquid AI with 24B total and 2.45B active parameters. It always reasons before answering and is built for agents that call tools. Liquid...","family":"liquid","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.05,"output":0.7,"cache_read":0.025}},"poolside/laguna-xs-2.1:free":{"id":"poolside/laguna-xs-2.1:free","name":"Poolside: Laguna XS 2.1 (free)","description":"Laguna XS 2.1 is the latest coding agent model in the 33B-A3B category from [Poolside](https://poolside.ai/) and a step forward from their Laguna XS.2 model (released in April 2026). It combines...","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-02","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-xs-2.1"},"poolside/laguna-xs-2.1":{"id":"poolside/laguna-xs-2.1","name":"Poolside: Laguna XS 2.1 (retires Oct 31)","description":"Laguna XS 2.1 is the latest coding agent model in the 33B-A3B category from [Poolside](https://poolside.ai/) and a step forward from their Laguna XS.2 model (released in April 2026). It combines...","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-02","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.1,"output":0.2,"cache_read":0.05},"canonical_model_id":"poolside/laguna-xs-2.1"},"poolside/laguna-s-2.1":{"id":"poolside/laguna-s-2.1","name":"Poolside: Laguna S 2.1 (retires Oct 31)","description":"Laguna S 2.1 is the latest coding agent model from [Poolside](). Laguna S 2.1 is a 118B total parameter model with 8B active parameters, scoring 70.2% on Terminal-Bench 2.1 and...","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.1,"output":0.2,"cache_read":0.01},"canonical_model_id":"poolside/laguna-s-2.1"},"poolside/laguna-s-2.1:free":{"id":"poolside/laguna-s-2.1:free","name":"Poolside: Laguna S 2.1 (free)","description":"Laguna S 2.1 is the latest coding agent model from [Poolside](). Laguna S 2.1 is a 118B total parameter model with 8B active parameters, scoring 70.2% on Terminal-Bench 2.1 and...","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-s-2.1"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Inkling Small is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 12B active parameters out of 276B total. It is positioned as the smaller, more efficient member of...","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":262144},"cost":{"input":0.45,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Inkling is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 41B active parameters out of 975B total. It is designed for general-purpose reasoning, coding, agentic and tool-use systems,...","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":471859},"cost":{"input":0.95,"output":4.05,"cache_read":0.16},"canonical_model_id":"thinkingmachines/inkling"},"cognitivecomputations/dolphin-mistral-24b-venice-edition":{"id":"cognitivecomputations/dolphin-mistral-24b-venice-edition","name":"Venice: Uncensored","description":"Venice Uncensored Dolphin Mistral 24B Venice Edition is a fine-tuned variant of Mistral-Small-24B-Instruct-2501, developed by dphn.ai in collaboration with Venice.ai. This model is designed as an “uncensored” instruct-tuned LLM, preserving...","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04-30","release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.2,"output":0.9},"canonical_model_id":"cognitivecomputations/dolphin-mistral-24b-venice-edition"},"thedrummer/unslopnemo-12b":{"id":"thedrummer/unslopnemo-12b","name":"TheDrummer: UnslopNemo 12B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04-30","release_date":"2024-11-08","last_updated":"2024-11-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":819200},"cost":{"input":0.4,"output":0.4},"canonical_model_id":"thedrummer/unslopnemo-12b"},"thedrummer/skyfall-36b-v2":{"id":"thedrummer/skyfall-36b-v2","name":"TheDrummer: Skyfall 36B V2","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2025-03-10","last_updated":"2025-03-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":29491},"cost":{"input":0.55,"output":0.8,"cache_read":0.25},"canonical_model_id":"thedrummer/skyfall-36b-v2"},"thedrummer/cydonia-24b-v4.1":{"id":"thedrummer/cydonia-24b-v4.1","name":"TheDrummer: Cydonia 24B V4.1","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04-30","release_date":"2025-09-27","last_updated":"2025-09-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.3,"output":0.5,"cache_read":0.15},"canonical_model_id":"thedrummer/cydonia-24b-v4.1"},"morph/morph-v3-large":{"id":"morph/morph-v3-large","name":"Morph: Morph V3 Large","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-07","last_updated":"2025-07-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.9,"output":1.9},"canonical_model_id":"morph/morph-v3-large"},"morph/morph-v3-fast":{"id":"morph/morph-v3-fast","name":"Morph: Morph V3 Fast","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-07","last_updated":"2025-07-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.8,"output":1.2},"canonical_model_id":"morph/morph-v3-fast"},"aion-labs/aion-3.5":{"id":"aion-labs/aion-3.5","name":"AionLabs: Aion 3.5","description":"Aion 3.5 is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models. It uses a collaborative generation process in which multiple specialized models each...","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":3,"output":6,"cache_read":0.75}},"aion-labs/aion-3.0-mini":{"id":"aion-labs/aion-3.0-mini","name":"AionLabs: Aion-3.0-Mini","description":"Aion-3.0 Mini is a multi-model roleplaying and storytelling system from AionLabs, built on the DeepSeek family of models. It uses a collaborative generation process in which multiple specialized models each...","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-14","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.7,"output":1.4,"cache_read":0.18},"canonical_model_id":"aion-labs/aion-3.0-mini"},"aion-labs/aion-2.0":{"id":"aion-labs/aion-2.0","name":"AionLabs: Aion-2.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-21","last_updated":"2025-12-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.8,"output":1.6,"cache_read":0.2},"canonical_model_id":"aion-labs/aion-2.0"},"aion-labs/aion-rp-llama-3.1-8b":{"id":"aion-labs/aion-rp-llama-3.1-8b","name":"AionLabs: Aion-RP 1.0 (8B)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12-31","release_date":"2025-02-04","last_updated":"2025-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":29491},"cost":{"input":0.8,"output":1.6},"canonical_model_id":"aion-labs/aion-rp-llama-3.1-8b"},"aion-labs/aion-3.5-mini":{"id":"aion-labs/aion-3.5-mini","name":"AionLabs: Aion 3.5 Mini","description":"Aion 3.5 Mini is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models. It is the smaller, lower-cost sibling of Aion 3.5 and uses...","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.7,"output":1.4,"cache_read":0.18}},"aion-labs/aion-3.0":{"id":"aion-labs/aion-3.0","name":"AionLabs: Aion-3.0","description":"Aion-3.0 is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models. It uses a collaborative generation process in which multiple specialized models each contribute...","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-05","last_updated":"2026-05-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":3,"output":6,"cache_read":0.75},"canonical_model_id":"aion-labs/aion-3.0"},"inception/mercury-2.5":{"id":"inception/mercury-2.5","name":"Inception: Mercury 2.5","description":"Mercury 2.5 is the fastest reasoning LLM, and the latest diffusion LLM (dLLM) from Inception. Instead of generating tokens sequentially, Mercury 2.5 produces and refines multiple tokens in parallel, achieving...","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":260000,"output":65536},"cost":{"input":0.2,"output":0.75,"cache_read":0.02},"canonical_model_id":"inception/mercury-2.5"},"inception/mercury-3":{"id":"inception/mercury-3","name":"Inception: Mercury 3","description":"Mercury 3 is a diffusion-based language model from Inception, and the lab's most capable model to date. It is suited for agentic workflows and real-time applications where latency compounds, such...","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.2,"output":0.7,"cache_read":0.02}},"inception/mercury-2":{"id":"inception/mercury-2","name":"Inception: Mercury 2","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":50000},"cost":{"input":0.25,"output":0.75,"cache_read":0.025},"canonical_model_id":"inception/mercury-2"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"GPT-6 Sol is the cost-efficient high-end model in OpenAI's GPT-6 series, positioned below the flagship GPT-6 Astra and above the fast GPT-6 Luna tier. It is suited for demanding professional...","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-4o-2024-05-13":{"id":"openai/gpt-4o-2024-05-13","name":"GPT-4o (2024-05-13)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-05-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":5,"output":15},"canonical_model_id":"openai/gpt-4o-2024-05-13"},"openai/gpt-6-sol-pro":{"id":"openai/gpt-6-sol-pro","name":"OpenAI: GPT-6 Sol Pro","description":"GPT-6 Sol Pro is the same underlying model as [GPT-6 Sol](https://openrouter.ai/openai/gpt-6-sol), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5}},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's flagship model for demanding end-to-end work. It is suited for advanced analysis, software engineering, deep research, scientific work, and document creation, with particular strengths in long-horizon...","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"openai/gpt-6-astra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-audio-mini":{"id":"openai/gpt-audio-mini","type":"chat","name":"GPT-Audio Mini","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-01","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","audio","pdf"],"output":["text","audio"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.6,"output":2.4},"canonical_model_id":"openai/gpt-audio-mini"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.018,"output":0.09,"cache_read":0.009},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-6.1-sol-pro":{"id":"openai/gpt-6.1-sol-pro","name":"OpenAI: GPT-6.1 Sol Pro","description":"GPT-6.1 Sol Pro is the same underlying model as [GPT-6.1 Sol](https://openrouter.ai/openai/gpt-6.1-sol), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. **Cost note:** pro mode spends far more...","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5}},"openai/gpt-4":{"id":"openai/gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":8191,"output":4096},"cost":{"input":30,"output":60},"canonical_model_id":"openai/gpt-4"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"GPT-5.6 Terra is a balanced model in OpenAI's GPT-5.6 series, positioned between the flagship Sol tier and the cost-efficient Luna tier. It is suited for everyday coding, reasoning, and agentic...","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-6-luna-pro":{"id":"openai/gpt-6-luna-pro","name":"OpenAI: GPT-6 Luna Pro","description":"GPT-6 Luna Pro is the same underlying model as [GPT-6 Luna](https://openrouter.ai/openai/gpt-6-luna), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125}},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/o1-pro":{"id":"openai/o1-pro","name":"o1-pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2025-03-19","last_updated":"2025-03-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":150,"output":600},"canonical_model_id":"openai/o1-pro"},"openai/gpt-chat-latest":{"id":"openai/gpt-chat-latest","name":"OpenAI: GPT Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-05-05","last_updated":"2026-05-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5}},"openai/gpt-5.2-chat":{"id":"openai/gpt-5.2-chat","name":"OpenAI: GPT-5.2 Chat","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-chat-latest"},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/gpt-5.4-image-2":{"id":"openai/gpt-5.4-image-2","type":"image","name":"GPT-5.4 Image 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":false,"structured_output":true,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["image","text","pdf"],"output":["image","text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":8,"output":15,"cache_read":2},"canonical_model_id":"openai/gpt-5.4-image-2"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-3.5-turbo-instruct":{"id":"openai/gpt-3.5-turbo-instruct","name":"OpenAI: GPT-3.5 Turbo Instruct","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2023-09-28","last_updated":"2023-09-28","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":4095,"output":3685},"cost":{"input":1.5,"output":2}},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-oss-safeguard-20b":{"id":"openai/gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-audio":{"id":"openai/gpt-audio","type":"chat","name":"GPT-Audio","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-01","release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text","audio","pdf"],"output":["text","audio"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10},"canonical_model_id":"openai/gpt-audio"},"openai/o3":{"id":"openai/o3","name":"o3","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"GPT-6 Luna is the fast, cost-efficient model in OpenAI's GPT-6 series, positioned below GPT-6 Sol. It is suited for high-volume and latency-sensitive workloads such as chat, classification, and lightweight agentic...","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-3.5-turbo-0613":{"id":"openai/gpt-3.5-turbo-0613","name":"OpenAI: GPT-3.5 Turbo (older v0613)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":4095,"output":3685},"cost":{"input":1,"output":2}},"openai/gpt-6-astra-pro":{"id":"openai/gpt-6-astra-pro","name":"OpenAI: GPT-6 Astra Pro ($$$$)","description":"GPT-6 Astra Pro is the same underlying model as [GPT-6 Astra](https://openrouter.ai/openai/gpt-6-astra), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5}},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.03,"output":0.15,"cache_read":0.015},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"GPT-5.6 Sol is the flagship model in OpenAI's GPT-5.6 series. It is suited for complex reasoning, coding, and agentic workflows, and is particularly strong at command-line and multi-step coding tasks...","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"GPT-6.1 Sol is an upgrade to GPT-6 Sol from OpenAI, positioned below the flagship GPT-6 Astra in the GPT-6 series. It is suited for agentic coding, computer use, document-heavy professional...","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"GPT-5.6 Luna is a fast, cost-efficient model in OpenAI's GPT-5.6 series. It is suited for high-volume, latency-sensitive tasks such as chat, classification, and lightweight agentic workflows, providing capable reasoning for...","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-5.6-luna-pro":{"id":"openai/gpt-5.6-luna-pro","name":"GPT-5.6 Luna","description":"GPT-5.6 Luna Pro is the same underlying model as [GPT-5.6 Luna](https://openrouter.ai/openai/gpt-5.6-luna), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5-image-mini":{"id":"openai/gpt-5-image-mini","type":"image","name":"GPT-5 Image Mini","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-10-16","last_updated":"2025-10-16","modalities":{"input":["pdf","image","text"],"output":["image","text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":2.5,"output":2,"cache_read":0.25},"canonical_model_id":"openai/gpt-5-image-mini"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4o-2024-11-20":{"id":"openai/gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"openai/gpt-5.6-sol-pro":{"id":"openai/gpt-5.6-sol-pro","name":"GPT-5.6 Sol","description":"GPT-5.6 Sol Pro is the same underlying model as [GPT-5.6 Sol](https://openrouter.ai/openai/gpt-5.6-sol), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-4o-2024-08-06":{"id":"openai/gpt-4o-2024-08-06","name":"GPT-4o (2024-08-06)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-08-06"},"openai/gpt-3.5-turbo-16k":{"id":"openai/gpt-3.5-turbo-16k","name":"OpenAI: GPT-3.5 Turbo 16k","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2023-08-28","last_updated":"2023-08-28","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":3,"output":4}},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/gpt-5-image":{"id":"openai/gpt-5-image","type":"image","name":"GPT-5 Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-10-01","release_date":"2025-10-14","last_updated":"2025-10-14","modalities":{"input":["image","text","pdf"],"output":["image","text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":10,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-5-image"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5-turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"openai/gpt-5.1-codex-max":{"id":"openai/gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex-max"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.03},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/o4-mini-high":{"id":"openai/o4-mini-high","name":"OpenAI: o4 Mini High","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275}},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/o3-pro":{"id":"openai/o3-pro","name":"o3-pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","pdf","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":20,"output":80},"canonical_model_id":"openai/o3-pro"},"openai/gpt-4o-mini-2024-07-18":{"id":"openai/gpt-4o-mini-2024-07-18","name":"OpenAI: GPT-4o-mini (2024-07-18)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5.1-codex"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/o3-mini-high":{"id":"openai/o3-mini-high","name":"OpenAI: o3 Mini High","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-02-12","last_updated":"2025-02-12","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55}},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.6-terra-pro":{"id":"openai/gpt-5.6-terra-pro","name":"GPT-5.6 Terra","description":"GPT-5.6 Terra Pro is the same underlying model as [GPT-5.6 Terra](https://openrouter.ai/openai/gpt-5.6-terra), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-r1-0528":{"id":"deepseek/deepseek-r1-0528","name":"DeepSeek R1 0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.5,"output":2.15,"cache_read":0.35},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek/deepseek-v3.2-exp":{"id":"deepseek/deepseek-v3.2-exp","name":"DeepSeek: DeepSeek V3.2 Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":147456},"cost":{"input":0.27,"output":0.41}},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048575,"output":943717},"cost":{"input":0.091,"output":0.182,"cache_read":0.0182},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v3.1-terminus":{"id":"deepseek/deepseek-v3.1-terminus","name":"DeepSeek: DeepSeek V3.1 Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-22","last_updated":"2025-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":147456},"cost":{"input":0.27,"output":1}},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.28,"output":0.42,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"DeepSeek V4.1 Flash 1017 is the October 17, 2026 snapshot of DeepSeek V4.1 Flash, served by DeepSeek under the `deepseek-flash` API model name. V4.1 Flash is a sparse mixture-of-experts model...","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro 0813 is a large-scale mixture-of-experts model from DeepSeek. This is the GA release of DeepSeek V4 Pro.","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-chat-v3-0324":{"id":"deepseek/deepseek-chat-v3-0324","name":"DeepSeek V3 0324","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07-31","release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":115200},"cost":{"input":0.29,"output":1.14,"cache_read":0.11},"canonical_model_id":"deepseek/deepseek-chat-v3-0324"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"DeepSeek V4 Flash Vision Exp is an experimental vision-enabled version of [DeepSeek V4 Flash 0731](https://openrouter.ai/deepseek/deepseek-v4-flash-0731) from DeepSeek, adding image understanding while matching the base model on text capabilities including agents,...","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0.2156,"output":0.6468,"cache_read":0.00686},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":384000},"cost":{"input":1.6,"output":3.2,"cache_read":0.135},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-chat":{"id":"deepseek/deepseek-chat","name":"DeepSeek Chat","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16000},"cost":{"input":0.2574,"output":1.0287},"canonical_model_id":"deepseek/deepseek-chat"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the cost-efficient tier of the V4.1 family. DeepSeek reports that it exceeds V4 Pro on performance, speed, and task...","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows.","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.44,"output":1.32,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-chat-v3.1":{"id":"deepseek/deepseek-chat-v3.1","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.25,"output":0.95,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-chat-v3.1"},"anthropic/claude-opus-4.1":{"id":"anthropic/claude-opus-4.1","name":"Claude Opus 4.1 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Claude Sonnet 5.5 is Anthropic's Sonnet-class model for well-scoped everyday work, succeeding Claude Sonnet 5 as a direct upgrade. It is especially strong at building features, fixing bugs, and producing...","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude Fable 5 is a Mythos-class model from Anthropic, built for autonomous knowledge work and coding. It supports text, image, and file inputs with text output, with reasoning support and...","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Claude Opus 5 is Anthropic’s flagship model for demanding reasoning, coding, and long-horizon agentic work. It is particularly strong at end-to-end software tasks, code review and bug finding, visual analysis...","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude Opus 5.5 is Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work, succeeding Claude Opus 5. It is particularly strong at multi-step changes in large codebases, code...","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Claude Opus 4.8 is Anthropic's most capable generally available model in the Opus family. It supports text, image, and file inputs with text output, with reasoning support and a 1M-token...","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Claude Haiku 5.5 is Anthropic's small, fast model for high-volume, cost-sensitive work such as summarization, subagents, and browser use. It succeeds Claude Haiku 4.5 with stronger coding, computer use, and...","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Sonnet 5 is Anthropic's most capable Sonnet-class model, with frontier performance across coding, agents, and professional work. It supports adaptive thinking with selectable reasoning effort levels (low, medium, high, max,...","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude Fable 5.1 improves on Claude Fable 5 across the board, with the biggest gains in agentic coding, long-running agentic workflows, and knowledge work: long code refactors, front-end and visual...","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"~x-ai/grok-latest":{"id":"~x-ai/grok-latest","name":"xAI: Grok Latest","description":"This model always redirects to the latest Grok model from xAI.","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5}},"nousresearch/hermes-3-llama-3.1-405b":{"id":"nousresearch/hermes-3-llama-3.1-405b","name":"Nous: Hermes 3 405B Instruct (retires Nov 2)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-16","last_updated":"2024-08-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":1,"output":1},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-405b"},"nousresearch/hermes-4-405b":{"id":"nousresearch/hermes-4-405b","name":"Nous: Hermes 4 405B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"hermes","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":1,"output":3},"canonical_model_id":"nousresearch/hermes-4-405b"},"nousresearch/hermes-3-llama-3.1-70b":{"id":"nousresearch/hermes-3-llama-3.1-70b","name":"Nous: Hermes 3 70B Instruct (retires Nov 2)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-18","last_updated":"2024-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.7,"output":0.7},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-70b"},"~z-ai/glm-latest":{"id":"~z-ai/glm-latest","name":"Z.ai: GLM Latest","description":"This model always redirects to the latest GLM model from Z.ai.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-19","last_updated":"2026-08-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.079,"output":1.4,"cache_read":0.078}},"~z-ai/glm-flash-latest":{"id":"~z-ai/glm-flash-latest","name":"Z.ai: GLM Flash Latest","description":"This model always redirects to the latest model in the GLM Flash family.","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.0413,"output":0.0773,"cache_read":0.03}},"tencent/hy-mt2-30b-a3b":{"id":"tencent/hy-mt2-30b-a3b","name":"Hy-MT2-30B-A3B","description":"Hy-MT2-30B-A3B is Tencent's flagship translation model in the Hy-MT2 family. It supports 33 language pairs and five Chinese dialect and minority-language pairs, with workflows for structured, delimiter-based, contextual, glossary-based, and...","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.074,"output":0.295},"canonical_model_id":"tencent/hy-mt2-30b-a3b"},"tencent/hunyuan-a13b-instruct":{"id":"tencent/hunyuan-a13b-instruct","name":"Tencent: Hunyuan A13B Instruct","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-08","last_updated":"2025-07-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":117964},"cost":{"input":0.14,"output":0.57}},"tencent/hy3-preview":{"id":"tencent/hy3-preview","name":"Hy3 preview","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.18,"output":0.6,"cache_read":0.06},"canonical_model_id":"tencent/hy3-preview"},"tencent/hy-mt2-7b":{"id":"tencent/hy-mt2-7b","name":"Hy-MT2-7B","description":"Hy-MT2-7B is a 7B-parameter translation model from Tencent. It supports 33 language pairs and five Chinese dialect and minority-language pairs, with workflows for structured, delimiter-based, contextual, glossary-based, and style-guided translation.","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.074,"output":0.295},"canonical_model_id":"tencent/hy-mt2-7b"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 preview","description":"Tencent: Hy4 preview is a mixture-of-experts model from Tencent, with 49B active parameters out of 770B total. It is designed for coding agents, complex tool-use workflows, and productivity tasks that...","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":0.7506,"output":2.2509,"cache_read":0.0378},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3","description":"Hy3 is a 295B-parameter Mixture-of-Experts model from Tencent (21B active, 192 experts with top-8 routing) built for reasoning, agentic workflows, and real-world production use. It supports a configurable reasoning effort:...","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":128000},"cost":{"input":0.0825,"output":0.33,"cache_read":0.020625},"canonical_model_id":"tencent/hy3"},"tencent/hy-mt2-1.8b":{"id":"tencent/hy-mt2-1.8b","name":"Hy-MT2-1.8B","description":"Hy-MT2-1.8B is a compact 1.8B-parameter translation model from Tencent. It supports 33 language pairs and five Chinese dialect and minority-language pairs, with workflows for structured, delimiter-based, contextual, glossary-based, and style-guided...","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.044,"output":0.177},"canonical_model_id":"tencent/hy-mt2-1.8b"},"ibm-granite/granite-4.0-h-micro":{"id":"ibm-granite/granite-4.0-h-micro","name":"IBM: Granite 4.0 Micro","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"granite","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-10-20","last_updated":"2025-10-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":117900},"cost":{"input":0.017,"output":0.112}},"ibm-granite/granite-4.2-8b":{"id":"ibm-granite/granite-4.2-8b","name":"IBM: Granite 4.2 8B","description":"Granite 4.2 8B is a dense reasoning model from IBM. It is suited for mathematics, code generation, multilingual dialogue, and agentic workflows that need multi-step reasoning. It supports full, low-effort,...","family":"granite","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-25","last_updated":"2026-08-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.06,"output":0.25,"cache_read":0.015},"canonical_model_id":"ibm/granite-4.2-8b"},"moondream/moondream3.1-9b-a2b":{"id":"moondream/moondream3.1-9b-a2b","name":"Moondream: Moondream 3.1","description":"Moondream 3.1 is a vision-language model from Moondream, the successor to Moondream 3 Preview, with 9B total parameters and 2B active (mixture-of-experts). It handles visual question answering, captioning, OCR, and...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":4096},"cost":{"input":0.3,"output":2.5}},"moondream/moondream3-preview":{"id":"moondream/moondream3-preview","name":"Moondream: Moondream 3 Preview","description":"Moondream 3 Preview is a vision-language model from Moondream with 9B total parameters and 2B active (mixture-of-experts). It handles visual question answering, captioning, OCR, and visual reasoning over images, with...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":4096},"cost":{"input":0.3,"output":2.5}},"stealth/claude-opus-4.7":{"id":"stealth/claude-opus-4.7","name":"Stealth: Claude Opus 4.7 (20% off)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"reasoning":0,"cache_read":0.4,"cache_write":5}},"stealth/claude-sonnet-4.6":{"id":"stealth/claude-sonnet-4.6","name":"Stealth: Claude Sonnet 4.6 (20% off)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2.4,"output":12,"reasoning":0,"cache_read":0.24,"cache_write":3}},"stealth/claude-opus-4.6":{"id":"stealth/claude-opus-4.6","name":"Stealth: Claude Opus 4.6 (20% off)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"reasoning":0,"cache_read":0.4,"cache_write":5}},"stealth/glyph-cluster":{"id":"stealth/glyph-cluster","name":"Stealth: Glyph Cluster (free)","description":"Your prompts and completions may be retained and used to train or improve the provider's services. Glyph Cluster is a reasoning model for coding and long-context analysis, served through Vercel AI Gateway during its stealth period at no cost. It supports function calling and streaming, takes text input only, and does not support structured outputs. Zero Data Retention is not available for this model.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0,"output":0,"reasoning":0,"cache_read":0}},"stealth/claude-opus-4.8":{"id":"stealth/claude-opus-4.8","name":"Stealth: Claude Opus 4.8 (20% off)","description":"Your prompts and completions may be retained and used to train or improve the provider's services. This third-party-served variant of Claude Opus 4.8 is offered at 20% lower cost than standard Claude Opus 4.8 pricing and is not served by Anthropic or Kilo Code.","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"reasoning":0,"cache_read":0.4,"cache_write":5}},"stealth/qwen3.6-plus":{"id":"stealth/qwen3.6-plus","name":"Stealth: Qwen3.6 Plus (50% off)","description":"Your prompts and completions may be retained and used to train or improve the provider's services. This third-party-served variant of Qwen3.6 Plus is offered at 50% lower cost than standard Qwen3.6 Plus pricing and is not served by Alibaba or Kilo Code. Note: a surcharge applies to long-context workloads exceeding 256K input tokens.","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":0,"cache_read":0.025,"cache_write":0.3125}},"meta-llama/llama-3.1-8b-instruct":{"id":"meta-llama/llama-3.1-8b-instruct","name":"Llama-3.1-8B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.02,"output":0.04},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta-llama/llama-3.2-3b-instruct":{"id":"meta-llama/llama-3.2-3b-instruct","name":"Meta: Llama 3.2 3B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":80000,"output":72000},"cost":{"input":0.05,"output":0.33},"canonical_model_id":"meta/llama-3.2-3b-instruct"},"meta-llama/llama-guard-4-12b":{"id":"meta-llama/llama-guard-4-12b","name":"Meta: Llama Guard 4 12B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":16384},"cost":{"input":0.144,"output":0.144},"canonical_model_id":"meta/llama-guard-4-12b"},"meta-llama/llama-3.1-70b-instruct":{"id":"meta-llama/llama-3.1-70b-instruct","name":"Llama-3.1-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.4,"output":0.4},"canonical_model_id":"meta/llama-3.1-70b-instruct"},"meta-llama/llama-3.2-1b-instruct":{"id":"meta-llama/llama-3.2-1b-instruct","name":"Llama 3.2 1B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":60000,"output":54000},"cost":{"input":0.027,"output":0.201},"canonical_model_id":"meta/llama-3.2-1b-instruct"},"meta-llama/llama-3.3-70b-instruct":{"id":"meta-llama/llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.1,"output":0.32},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/llama-4-scout":{"id":"meta-llama/llama-4-scout","name":"Meta: Llama 4 Scout","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":327680,"output":16384},"cost":{"input":0.08,"output":0.24},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"meta-llama/llama-4-maverick":{"id":"meta-llama/llama-4-maverick","name":"Meta: Llama 4 Maverick","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.1875,"output":0.6525,"cache_read":0.05},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"xiaomi/mimo-v2.6-pro-ultraspeed":{"id":"xiaomi/mimo-v2.6-pro-ultraspeed","name":"MiMo-V2.6-Pro-UltraSpeed","description":"MiMo-V2.6-Pro-UltraSpeed is the fast speed edition of Xiaomi's flagship foundation model, MiMo-V2.6-Pro. Built from the same 1T MiMo-V2.6-Pro checkpoint, it matches the original model in quality while delivering roughly 10x...","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.35,"output":8.7,"cache_read":0.036},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":945000},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.003,"tiers":[{"input":0.8,"output":4,"cache_read":0.16,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.8,"output":4,"cache_read":0.16}},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2.5-pro":{"id":"xiaomi/mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":945000},"status":"deprecated","cost":{"input":0.435,"output":0.87,"cache_read":0.004,"tiers":[{"input":2,"output":6,"cache_read":0.4,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.4}},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"MiMo-V2.6-Pro is the flagship foundation model developed by Xiaomi. Built at a scale of over 1T parameters, it is designed to push the ceiling of capability for the most demanding...","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.43,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"xiaomi/mimo-v2.6-flash":{"id":"xiaomi/mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo-V2.6-Flash is an open-source foundation model developed by Xiaomi. Built on a Mixture-of-Experts architecture with 309B total parameters and 15B activated per token, it employs a hybrid attention mechanism for...","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.28,"cache_read":0.05},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"~google/gemini-pro-latest":{"id":"~google/gemini-pro-latest","name":"Google: Gemini Pro Latest","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["audio","pdf","image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":2,"output":10,"reasoning":10,"cache_read":0.1,"cache_write":0.1875}},"~google/gemini-flash-latest":{"id":"~google/gemini-flash-latest","name":"Google: Gemini Flash Latest","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667}},"anthracite-org/magnum-v4-72b":{"id":"anthracite-org/magnum-v4-72b","name":"Magnum v4 72B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":4096},"cost":{"input":2.5,"output":5},"canonical_model_id":"anthracite-org/magnum-v4-72b"},"stepfun/step-3.7-flash":{"id":"stepfun/step-3.7-flash","name":"Step 3.7 Flash","description":"Step 3.7 Flash is StepFun's latest high-efficiency multimodal Mixture-of-Experts model. It pairs a 196B-parameter language backbone with a vision encoder for native image and video understanding, activating roughly 11B parameters...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":230400},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"stepfun/step-5-preview":{"id":"stepfun/step-5-preview","name":"Step 5 Preview","description":"Step 5 Preview is StepFun's flagship model for agentic work, built on a sparse Mixture-of-Experts architecture (27B active / 600B total parameters). It performs strongly in software engineering and professional...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":64000},"cost":{"input":1,"output":2.7,"cache_read":0.1},"canonical_model_id":"stepfun/step-5-preview"},"stepfun/step-3.5-flash":{"id":"stepfun/step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"stepfun/step-3.5-flash"},"~openai/gpt-terra-latest":{"id":"~openai/gpt-terra-latest","name":"OpenAI: GPT Terra Latest","description":"This model always redirects to the latest model in the OpenAI GPT Terra family.","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5}},"~openai/gpt-mini-latest":{"id":"~openai/gpt-mini-latest","name":"OpenAI: GPT Mini Latest","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075}},"~openai/gpt-astra-latest":{"id":"~openai/gpt-astra-latest","name":"OpenAI: GPT Astra Latest ($$$$)","description":"This model always redirects to the latest model in the OpenAI GPT Astra family.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5}},"~openai/gpt-luna-latest":{"id":"~openai/gpt-luna-latest","name":"OpenAI: GPT Luna Latest","description":"This model always redirects to the latest model in the OpenAI GPT Luna family.","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125}},"~openai/gpt-sol-latest":{"id":"~openai/gpt-sol-latest","name":"OpenAI: GPT Sol Latest","description":"This model always redirects to the latest model in the OpenAI GPT Sol family.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5}},"reflection/beam-free":{"id":"reflection/beam-free","name":"Reflection: Beam 1.0 (free)","description":"Your prompts and completions may be retained and used to train or improve the provider's services. Beam is Reflection's reasoning model for coding and agentic workflows, combining strong coding capabilities with fast, low-latency inference, served through Vercel AI Gateway at no cost. It supports function calling, structured outputs, and streaming, and takes text input only. Zero Data Retention is not available for this model.","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"reasoning":0,"cache_read":0}},"reflection/beam-1.0-501b-a23b":{"id":"reflection/beam-1.0-501b-a23b","name":"Reflection: Beam 1.0 501B A23B","description":"Beam 1.0 is Reflection's open-weight reasoning model for coding and long-horizon agentic workflows. It uses a sparse mixture-of-experts architecture with 501B total and 23B active parameters per token, and a...","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":0.25,"output":1,"cache_read":0.05},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"elorian/glimpse":{"id":"elorian/glimpse","name":"Elorian: Glimpse","description":"Glimpse is a multimodal visual reasoning model from Elorian, built for visual understanding and complex visual reasoning in the physical world. It reasons directly over images and video rather than...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":256000},"cost":{"input":0.6,"output":3,"cache_read":0.06}},"~deepseek/deepseek-v4-flash-latest":{"id":"~deepseek/deepseek-v4-flash-latest","name":"DeepSeek: DeepSeek V4 Flash Latest","description":"This model always redirects to the latest model in the DeepSeek V4 Flash family.","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-01","last_updated":"2026-08-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.0422,"output":0.1729,"cache_read":0.0275}},"~deepseek/deepseek-pro-latest":{"id":"~deepseek/deepseek-pro-latest","name":"DeepSeek: DeepSeek Pro Latest","description":"This model always redirects to the latest model in the DeepSeek Pro family.","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-14","last_updated":"2026-09-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":393216},"cost":{"input":0.1307,"output":8,"cache_read":0.13}},"~deepseek/deepseek-flash-latest":{"id":"~deepseek/deepseek-flash-latest","name":"DeepSeek: DeepSeek Flash Latest","description":"This model always redirects to the latest model in the DeepSeek Flash family.","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-14","last_updated":"2026-09-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.07,"output":0.5,"cache_read":0.008}},"~anthropic/claude-fable-latest":{"id":"~anthropic/claude-fable-latest","name":"Anthropic: Claude Fable Latest ($$$$)","description":"This model always redirects to the latest model in the Claude Fable family.","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5}},"~anthropic/claude-haiku-latest":{"id":"~anthropic/claude-haiku-latest","name":"Anthropic: Claude Haiku Latest","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125}},"~anthropic/claude-opus-latest":{"id":"~anthropic/claude-opus-latest","name":"Anthropic: Claude Opus Latest","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5}},"~anthropic/claude-sonnet-latest":{"id":"~anthropic/claude-sonnet-latest","name":"Anthropic: Claude Sonnet Latest","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5}},"writer/palmyra-x5":{"id":"writer/palmyra-x5","name":"Writer: Palmyra X5","description":"General-purpose chat model for instruction following, writing, and analysis","family":"palmyra","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1040000,"output":8192},"cost":{"input":0.6,"output":6},"canonical_model_id":"writer/palmyra-x5"},"fireworks/ember-1":{"id":"fireworks/ember-1","name":"Fireworks: Ember-1","description":"Ember-1 is a specialized reasoning model from Fireworks Research, built on [Kimi K3](https://openrouter.ai/moonshotai/kimi-k3). It is designed to make every token go further: it produces shorter reasoning traces, using roughly 40%...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"fireworks/ember-1"},"amazon/nova-2-lite-v1":{"id":"amazon/nova-2-lite-v1","name":"Amazon: Nova 2 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.3,"output":2.5},"canonical_model_id":"amazon/nova-2-lite-v1"},"amazon/nova-pro-v1":{"id":"amazon/nova-pro-v1","name":"Amazon: Nova Pro 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":5120},"cost":{"input":0.8,"output":3.2},"canonical_model_id":"amazon/nova-pro-v1"},"amazon/nova-lite-v1":{"id":"amazon/nova-lite-v1","name":"Amazon: Nova Lite 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":5120},"cost":{"input":0.06,"output":0.24},"canonical_model_id":"amazon/nova-lite-v1"},"amazon/nova-premier-v1":{"id":"amazon/nova-premier-v1","name":"Amazon: Nova Premier 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2025-04-30","last_updated":"2025-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"status":"deprecated","cost":{"input":2.5,"output":12.5,"cache_read":0.625},"canonical_model_id":"amazon/nova-premier"},"amazon/nova-micro-v1":{"id":"amazon/nova-micro-v1","name":"Amazon: Nova Micro 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":5120},"cost":{"input":0.035,"output":0.14},"canonical_model_id":"amazon/nova-micro-v1"},"~moonshotai/kimi-latest":{"id":"~moonshotai/kimi-latest","name":"MoonshotAI: Kimi Latest","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.238,"output":16,"cache_read":0.228}},"perplexity/sonar-reasoning-pro":{"id":"perplexity/sonar-reasoning-pro","name":"Perplexity: Sonar Reasoning Pro","description":"Web-grounded reasoning model for multi-step research and cited answers","family":"sonar-reasoning","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-07","last_updated":"2025-03-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":115200},"cost":{"input":2,"output":8}},"perplexity/sonar-deep-research":{"id":"perplexity/sonar-deep-research","name":"Perplexity: Sonar Deep Research","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar-deep-research","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-07","last_updated":"2025-03-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":115200},"cost":{"input":2,"output":8,"reasoning":3}},"perplexity/sonar-pro-search":{"id":"perplexity/sonar-pro-search","name":"Perplexity: Sonar Pro Search","description":"Advanced Sonar search model for deeper research and cited synthesis","family":"sonar-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-10-30","last_updated":"2025-10-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8000},"cost":{"input":3,"output":15}},"perplexity/sonar-pro":{"id":"perplexity/sonar-pro","name":"Perplexity: Sonar Pro","description":"Advanced Sonar search model for deeper research and cited synthesis","family":"sonar-pro","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-07","last_updated":"2025-03-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8000},"cost":{"input":3,"output":15}},"perplexity/sonar":{"id":"perplexity/sonar","name":"Perplexity: Sonar","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-01-27","last_updated":"2025-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":127072,"output":114364},"cost":{"input":1,"output":1}},"xingchen-agi/xing4.0-29b-a4b:free":{"id":"xingchen-agi/xing4.0-29b-a4b:free","name":"XingChen: Xing4.0 29B A4B (free)","description":"Xing4.0-29B-A4B is an open-weight mixture-of-experts language model from China Telecom's AI team (TeleAI) and the successor to the TeleChat series. It has 29B total parameters with 4B active per token...","family":"agi","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0,"output":0}},"relace/relace-search":{"id":"relace/relace-search","name":"Relace: Relace Search","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":1,"output":3},"canonical_model_id":"relace/relace-search"},"relace/relace-apply-3":{"id":"relace/relace-apply-3","name":"Relace: Relace Apply 3","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2025-09-26","last_updated":"2025-09-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.85,"output":1.25},"canonical_model_id":"relace/relace-apply-3"},"perceptron/perceptron-mk1":{"id":"perceptron/perceptron-mk1","name":"Perceptron: Perceptron Mk1","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":0.15,"output":1.5},"canonical_model_id":"perceptron/perceptron-mk1"},"perceptron/perceptron-mk1.5":{"id":"perceptron/perceptron-mk1.5","name":"Perceptron: Perceptron Mk1.5","description":"Perceptron Mk1.5 is Perceptron's embodied reasoning model for physical agents. It accepts text, image, video, and audio input, and answers with text plus optional structured annotations: points, boxes, polygons, tracks,...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":36864,"output":8192},"cost":{"input":0.15,"output":1.5}},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax-M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-m1":{"id":"minimax/minimax-m1","name":"MiniMax: MiniMax M1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":40000},"cost":{"input":0.4,"output":2.2}},"minimax/minimax-01":{"id":"minimax/minimax-01","name":"MiniMax: MiniMax-01","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-01-15","last_updated":"2025-01-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000192,"output":900172},"cost":{"input":0.2,"output":1.1}},"minimax/minimax-m2-her":{"id":"minimax/minimax-m2-her","name":"MiniMax-M2 Her","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-01-23","last_updated":"2026-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":2048},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2-Her"},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax-M2","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m3.1-flash:free":{"id":"minimax/minimax-m3.1-flash:free","name":"MiniMax: MiniMax M3.1 Flash (free)","description":"MiniMax-M3.1-Flash is a multimodal model from MiniMax. Compared with M3, it delivers significant improvements across multiple benchmarks, particularly in complex agentic tasks, coding, and cybersecurity. It is built for complex,...","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0,"output":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":32768},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"upstage/solar-mini4":{"id":"upstage/solar-mini4","name":"Upstage: Solar Mini 4","description":"Solar Mini 4 is Upstage's compact, cost-efficient language model, a 35B-parameter mixture-of-experts with 3B active parameters and a 524K context window. It is built for agentic use cases where response...","family":"solar-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"upstage/solar-mini4"},"upstage/solar-pro-3":{"id":"upstage/solar-pro-3","name":"Upstage: Solar Pro 3","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07","release_date":"2026-01-26","last_updated":"2026-03-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":117964},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"upstage/solar-pro3"},"upstage/solar-pro4":{"id":"upstage/solar-pro4","name":"Upstage: Solar Pro 4","description":"Solar Pro 4 is a large language model from Upstage. It is suited for agentic workflows, office productivity, document-intensive work, and coding.","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"upstage/solar-pro4"},"inference-net/schematron-v2-small":{"id":"inference-net/schematron-v2-small","name":"Inference.net: Schematron V2 Small","description":"Schematron V2 Small is a 3B-parameter HTML-to-JSON extraction model from Inference.net. It prioritizes extraction quality for complex schemas and long pages. Extraction instructions must be supplied through a JSON schema...","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":0.05,"output":0.23,"cache_read":0.05},"canonical_model_id":"inference-net/schematron-v2-small"},"inference-net/schematron-v2-turbo":{"id":"inference-net/schematron-v2-turbo","name":"Inference.net: Schematron V2 Turbo","description":"Schematron V2 Turbo is a 3B-parameter HTML-to-JSON extraction model from Inference.net. It prioritizes throughput for high-volume extraction workloads. Extraction instructions must be supplied through a JSON schema in response_format rather...","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.03,"output":0.15,"cache_read":0.03},"canonical_model_id":"inference-net/schematron-v2-turbo"},"scaledown/summarize":{"id":"scaledown/summarize","name":"ScaleDown: Summarize","description":"Summarize is an abstractive summarization model from ScaleDown. It condenses long text and documents into a fluent, faithful rewrite rather than lifting sentences from the source, and is suited for...","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":4000},"cost":{"input":0.05,"output":0}},"scaledown/extract":{"id":"scaledown/extract","name":"ScaleDown: Extract","description":"Extract is a description-driven entity extraction model from ScaleDown. Instead of a fixed set of entity types, callers describe the fields they want in plain English through `provider.options.scaledown.entities`, and the...","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":900000},"cost":{"input":0.05,"output":0}},"microsoft/phi-4":{"id":"microsoft/phi-4","name":"Microsoft: Phi 4","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"phi","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-01-10","last_updated":"2025-01-10","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":14745},"cost":{"input":0.056,"output":0.112}},"nvidia/nemotron-3.5-lightning":{"id":"nvidia/nemotron-3.5-lightning","name":"Nemotron 3.5 Lightning 30B A3B","description":"NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that...","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.025,"output":0.12,"cache_read":0.0125},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-3-super-120b-a12b":{"id":"nvidia/nemotron-3-super-120b-a12b","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.08,"output":0.45},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/nemotron-3.5-content-safety:free":{"id":"nvidia/nemotron-3.5-content-safety:free","name":"NVIDIA: Nemotron 3.5 Content Safety (free)","description":"NVIDIA Nemotron 3.5 Content Safety is a compact 4B-parameter multimodal guardrail model from NVIDIA, fine-tuned from Google Gemma-3-4B. It moderates both inputs to and responses from LLMs and VLMs, accepting... **Terms of service** For NVIDIA free endpoints (Super/Ultra/etc): Trial use only - do not submit personal or confidential data. Your use is logged for security purposes and to improve NVIDIA products and services. The logged session data for improvement purposes is not linked to your identity or any persistent identifier. For more information about our data processing practices, see our [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/). By interacting with this endpoint, you consent to our collection, recording, and use of such information and the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"nvidia/nemotron-3-ultra-550b-a55b":{"id":"nvidia/nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it...","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":182520},"cost":{"input":0.4,"output":1.76,"cache_read":0.08},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/nemotron-3.5-lightning:free":{"id":"nvidia/nemotron-3.5-lightning:free","name":"NVIDIA: Nemotron 3.5 Lightning (free)","description":"NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that... **Terms of service** For NVIDIA free endpoints (Super/Ultra/etc): Trial use only - do not submit personal or confidential data. Your use is logged for security purposes and to improve NVIDIA products and services. The logged session data for improvement purposes is not linked to your identity or any persistent identifier. For more information about our data processing practices, see our [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/). By interacting with this endpoint, you consent to our collection, recording, and use of such information and the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-3-nano-30b-a3b":{"id":"nvidia/nemotron-3-nano-30b-a3b","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.05,"output":0.2},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/nemotron-3.5-super-vl":{"id":"nvidia/nemotron-3.5-super-vl","name":"NVIDIA: Nemotron 3.5 Super VL","description":"NVIDIA Nemotron 3.5 Super VL is an open multimodal mixture-of-experts model from NVIDIA, with 12B active parameters out of 120B total. It takes text, images, documents, GUI screenshots, and video...","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.1,"output":0.4,"cache_read":0.08},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/nemotron-3.5-super-vl:free":{"id":"nvidia/nemotron-3.5-super-vl:free","name":"NVIDIA: Nemotron 3.5 Super VL (free)","description":"NVIDIA Nemotron 3.5 Super VL is an open multimodal mixture-of-experts model from NVIDIA, with 12B active parameters out of 120B total. It takes text, images, documents, GUI screenshots, and video... **Terms of service** For NVIDIA free endpoints (Super/Ultra/etc): Trial use only - do not submit personal or confidential data. Your use is logged for security purposes and to improve NVIDIA products and services. The logged session data for improvement purposes is not linked to your identity or any persistent identifier. For more information about our data processing practices, see our [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/). By interacting with this endpoint, you consent to our collection, recording, and use of such information and the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/nemotron-3-super-120b-a12b:free":{"id":"nvidia/nemotron-3-super-120b-a12b:free","name":"NVIDIA: Nemotron 3 Super (free)","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/nemotron-3-ultra-550b-a55b:free":{"id":"nvidia/nemotron-3-ultra-550b-a55b:free","name":"NVIDIA: Nemotron 3 Ultra (free)","description":"NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it... **Terms of service** For NVIDIA free endpoints (Super/Ultra/etc): Trial use only - do not submit personal or confidential data. Your use is logged for security purposes and to improve NVIDIA products and services. The logged session data for improvement purposes is not linked to your identity or any persistent identifier. For more information about our data processing practices, see our [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/). By interacting with this endpoint, you consent to our collection, recording, and use of such information and the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/nemotron-3.5-content-safety":{"id":"nvidia/nemotron-3.5-content-safety","name":"Nemotron 3.5 Content Safety","description":"NVIDIA Nemotron 3.5 Content Safety is a compact 4B-parameter multimodal guardrail model from NVIDIA, fine-tuned from Google Gemma-3-4B. It moderates both inputs to and responses from LLMs and VLMs, accepting...","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.2,"output":0.2},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free":{"id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free","name":"NVIDIA: Nemotron 3 Nano Omni (free)","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"mancer/weaver":{"id":"mancer/weaver","name":"Mancer: Weaver (alpha)","description":"An attempt to recreate Claude-style verbosity, but don't expect the same level of coherence or memory. Meant for use in roleplay/narrative situations.","family":"alpha","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-06-30","release_date":"2023-08-02","last_updated":"2023-08-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":6000},"cost":{"input":0.4,"output":0.75},"canonical_model_id":"mancer/weaver"},"bytedance/ui-tars-1.5-7b":{"id":"bytedance/ui-tars-1.5-7b","name":"ByteDance: UI-TARS 7B ","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":2048},"cost":{"input":0.1,"output":0.2,"cache_read":0.1}},"sao10k/l3.1-euryale-70b":{"id":"sao10k/l3.1-euryale-70b","name":"Sao10K: Llama 3.1 Euryale 70B v2.2","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-28","last_updated":"2024-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.85,"output":0.85},"canonical_model_id":"sao10k/l3.1-euryale-70b"},"sao10k/l3-lunaris-8b":{"id":"sao10k/l3-lunaris-8b","name":"Sao10K: Llama 3 8B Lunaris","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-13","last_updated":"2024-08-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":7372},"cost":{"input":0.04,"output":0.05},"canonical_model_id":"sao10k/l3-lunaris-8b"},"meituan/longcat-2.0":{"id":"meituan/longcat-2.0","name":"Meituan: LongCat 2.0","description":"LongCat 2.0 is a sparse mixture-of-experts language model from Meituan, with 48B active parameters out of 1.6T total. It is suited for coding, repository-level changes, long-horizon problem solving, and agentic...","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048756,"output":262144},"cost":{"input":0.75,"output":3,"cache_read":0.015},"canonical_model_id":"meituan/longcat-2.0"},"nex-agi/nex-n2.5-pro":{"id":"nex-agi/nex-n2.5-pro","name":"Nex AGI: Nex-N2.5-Pro","description":"Nex-N2.5 is an agentic model built to turn goals into working, verified outcomes. Its core strength is agentic coding within a visual feedback loop: it can explore codebases, implement multi-file...","family":"agi","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0.075,"output":0.25,"cache_read":0.015}},"nex-agi/nex-n2.5-mini":{"id":"nex-agi/nex-n2.5-mini","name":"Nex AGI: Nex-N2.5-Mini","description":"Nex-N2.5 is an agentic model built to turn goals into working, verified outcomes. Its core strength is agentic coding within a visual feedback loop: it can explore codebases, implement multi-file...","family":"agi","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0.025,"output":0.1,"cache_read":0.0025}},"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":81920},"cost":{"input":0.1625,"output":1.3},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3-vl-30b-a3b-instruct":{"id":"qwen/qwen3-vl-30b-a3b-instruct","name":"Qwen3 VL 30B A3B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"qwen/qwen3-30b-a3b":{"id":"qwen/qwen3-30b-a3b","name":"Qwen3 30B A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.5},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.475,"output":4.425,"cache_read":0.295,"cache_write":1.84375},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Qwen3.7-Plus is a cost-effective model in Alibaba's Qwen3.7 series. It supports text and image input with text output, building on the series' text capabilities with a comprehensive upgrade to its...","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.32,"output":1.28,"cache_read":0.064,"cache_write":0.4},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3-14b":{"id":"qwen/qwen3-14b","name":"Qwen: Qwen3 14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.24}},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Qwen3.8 27B is an open-weight dense vision-language model from Qwen. It is suited for coding, professional workflows, research, multimodal interaction, and long-running agent tasks, with flexible thinking that can be...","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.425,"output":2.55,"cache_read":0.085,"cache_write":0.53125},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.8-max-prime":{"id":"qwen/qwen3.8-max-prime","name":"Qwen 3.8 Max Prime","description":"Qwen3.8 Max Prime is a higher-throughput variant of Qwen3.8 Max from Alibaba's Qwen team, served as a separate SKU at a higher price point. It accepts text, image, and video...","family":"qwen3.8-max","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":4,"output":12,"cache_read":0.5},"canonical_model_id":"alibaba/qwen3.8-max-prime"},"qwen/qwen3.6-35b-a3b":{"id":"qwen/qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.1,"output":0.95,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.26,"output":2.08},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3.8-2.4t-a95b":{"id":"qwen/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1010000,"output":909000},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"qwen/qwen3-coder-next":{"id":"qwen/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.12,"output":0.8,"cache_read":0.07},"canonical_model_id":"alibaba/qwen3-coder-next"},"qwen/qwen3-30b-a3b-instruct-2507":{"id":"qwen/qwen3-30b-a3b-instruct-2507","name":"Qwen: Qwen3 30B A3B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.09,"output":0.3},"canonical_model_id":"alibaba/qwen3-30b-a3b-instruct-2507"},"qwen/qwen-2.5-7b-instruct":{"id":"qwen/qwen-2.5-7b-instruct","name":"Qwen: Qwen2.5 7B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-10-16","last_updated":"2024-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":29491},"cost":{"input":0.1,"output":0.2}},"qwen/qwen3.8-max-0902":{"id":"qwen/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"Qwen3.8 Max 0902 is an updated snapshot of Qwen3.8 Max from Alibaba's Qwen team. It is a 2.4-trillion-parameter mixture-of-experts model that accepts text, image, and video input and returns text,...","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"qwen/qwen-2.5-72b-instruct":{"id":"qwen/qwen-2.5-72b-instruct","name":"Qwen2.5 72B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-09-19","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":16384},"cost":{"input":0.36,"output":0.4}},"qwen/qwen3.5-flash-02-23":{"id":"qwen/qwen3.5-flash-02-23","name":"Qwen: Qwen3.5-Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-25","last_updated":"2026-02-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.065,"output":0.26}},"qwen/qwen3-vl-8b-instruct":{"id":"qwen/qwen3-vl-8b-instruct","name":"Qwen: Qwen3 VL 8B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-14","last_updated":"2025-10-14","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0.25,"output":0.75,"cache_read":0.12}},"qwen/qwen3.5-plus-02-15":{"id":"qwen/qwen3.5-plus-02-15","name":"Qwen: Qwen3.5 Plus 2026-02-15","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.26,"output":1.56}},"qwen/qwen2.5-vl-72b-instruct":{"id":"qwen/qwen2.5-vl-72b-instruct","name":"Qwen: Qwen2.5 VL 72B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-01-26","last_updated":"2025-01-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":115200},"cost":{"input":0.8,"output":1,"cache_read":0.4},"canonical_model_id":"alibaba/qwen2-5-vl-72b-instruct"},"qwen/qwen3.7-flash":{"id":"qwen/qwen3.7-flash","name":"Qwen3.7 Flash","description":"Qwen3.7 Flash is a vision-language reasoning model from Alibaba. It is suited for multimodal agents, visual coding, search, and computer interaction, with strengths in object recognition, spatial understanding, and real-world...","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":65536},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.038},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.195,"output":1.56},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1875,"output":1.125,"cache_write":0.234375},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.39,"output":2.34},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3-coder-30b-a3b-instruct":{"id":"qwen/qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.07,"output":0.28},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"qwen/qwen3-next-80b-a3b-thinking":{"id":"qwen/qwen3-next-80b-a3b-thinking","name":"Qwen3-Next 80B-A3B (Thinking)","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"qwen/qwen3.8-omni-flash":{"id":"qwen/qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen3.8 Omni Flash is an omni-modal reasoning model from Alibaba, the first Qwen model built around agentic capabilities with native audio-video understanding. It is suited for audio-video analysis and summarization,...","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen3.8 Flash is a multimodal reasoning model from Alibaba. It is suited for coding assistance, agentic workflows, visual understanding, document and codebase analysis, desktop interaction, chart analysis, and long-video analysis.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3-next-80b-a3b-instruct":{"id":"qwen/qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.09,"output":1.1},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"qwen/qwen3-vl-30b-a3b-thinking":{"id":"qwen/qwen3-vl-30b-a3b-thinking","name":"Qwen: Qwen3 VL 30B A3B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0.29,"output":1}},"qwen/qwen3-235b-a22b-2507":{"id":"qwen/qwen3-235b-a22b-2507","name":"Qwen: Qwen3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.09,"output":0.55},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"qwen/qwen3-32b":{"id":"qwen/qwen3-32b","name":"Qwen3 32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.08,"output":0.28},"canonical_model_id":"alibaba/qwen3-32b"},"qwen/qwen3-235b-a22b-thinking-2507":{"id":"qwen/qwen3-235b-a22b-thinking-2507","name":"Qwen: Qwen3 235B A22B Thinking 2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.45,"output":3.5},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"qwen/qwen3.5-9b":{"id":"qwen/qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.08,"output":0.13,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.5-9b"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.325,"output":1.95,"cache_write":0.40625},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen/qwen3-vl-235b-a22b-instruct":{"id":"qwen/qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.21,"output":1.9,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"qwen/qwen-2.5-coder-32b-instruct":{"id":"qwen/qwen-2.5-coder-32b-instruct","name":"Qwen2.5 Coder 32B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":29491},"cost":{"input":0.66,"output":1},"canonical_model_id":"alibaba/qwen2.5-coder-32b-instruct"},"qwen/qwen-plus":{"id":"qwen/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.26,"output":0.78,"cache_read":0.052,"cache_write":0.325},"canonical_model_id":"alibaba/qwen-plus"},"qwen/qwen3.5-plus-20260420":{"id":"qwen/qwen3.5-plus-20260420","name":"Qwen: Qwen3.5 Plus 2026-04-20","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":1.8,"cache_write":0.375}},"qwen/qwen3-coder-flash":{"id":"qwen/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.195,"output":0.975,"cache_read":0.039,"cache_write":0.24375},"canonical_model_id":"alibaba/qwen3-coder-flash"},"qwen/qwen3-coder":{"id":"qwen/qwen3-coder","name":"Qwen: Qwen3 Coder 480B A35B","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":1,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.45,"output":2.7},"canonical_model_id":"alibaba/qwen3.6-27b"},"cohere/command-r7b-12-2024":{"id":"cohere/command-r7b-12-2024","name":"Command R7B","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-r","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-12-13","last_updated":"2024-12-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.0375,"output":0.15},"canonical_model_id":"cohere/command-r7b-12-2024"},"cohere/command-r-plus-08-2024":{"id":"cohere/command-r-plus-08-2024","name":"Command R+","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-r-plus-08-2024"},"cohere/north-mini-code:free":{"id":"cohere/north-mini-code:free","name":"Cohere: North Mini Code (free)","description":"North Mini Code is Cohere's first agentic coding model and the debut of its North family. A sparse mixture-of-experts model with 30B total parameters and 3B active, it is optimized...","family":"north","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09-23","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0,"output":0},"canonical_model_id":"cohere/north-mini-code-1-0"},"cohere/command-a-plus":{"id":"cohere/command-a-plus","name":"Cohere: Command A+","description":"Command A+ is Cohere's flagship model for enterprise agentic workflows. It accepts text and image inputs with a 192K context window, supports native tool calling with strict tool schemas, structured...","family":"command-a","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04-01","release_date":"2026-05-20","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":192000,"output":64000},"cost":{"input":0.3,"output":1.5,"cache_read":0.15},"canonical_model_id":"cohere/command-a-plus-05-2026"},"cohere/command-a":{"id":"cohere/command-a","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a"},"cohere/command-r-08-2024":{"id":"cohere/command-r-08-2024","name":"Command R","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"cohere/command-r-08-2024"},"recursive/flint-v1.0:free":{"id":"recursive/flint-v1.0:free","name":"Recursive: Flint (free)","description":"Flint is an open-weight mixture-of-experts model from Recursive, with about 2.5B active parameters out of 40B total, built by Recursive's autoresearch system Eureka Machine for coding agents. It is suited...","family":"flint","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-21","last_updated":"2026-10-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0,"output":0}},"unbiased/pareto-26.10":{"id":"unbiased/pareto-26.10","name":"Pareto 26.10","description":"Pareto is a multimodal composite model built for research, coding, and agentic workflows, while delivering frontier-level performance across a broad range of general-purpose tasks.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.8,"output":3.2,"cache_read":0.03}},"unbiased/pareto-26.9":{"id":"unbiased/pareto-26.9","name":"Pareto 26.9","description":"Pareto is a multimodal composite model built for research, coding, and agentic workflows, while delivering frontier-level performance across a broad range of general-purpose tasks.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.25}},"prism-ml/ternary-bonsai-2-27b":{"id":"prism-ml/ternary-bonsai-2-27b","name":"PrismML: Ternary Bonsai 2 27B","description":"Bonsai 2 27B is a 27B-parameter reasoning model from PrismML derived from Qwen3.8-27B. It supports coding, mathematics, tool calling, and image understanding with a 262K-token context window. Ternary compression shrinks...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.075,"output":0.5,"cache_read":0.0375}},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Gemini 3.8 Flash is Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows.","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.125,"output":0.75,"reasoning":0.75,"cache_read":0.0125,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"status":"deprecated","cost":{"input":0.15,"output":1.25,"reasoning":1.25,"cache_read":0.015,"cache_write":0.041667},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Gemini 4 Argon is Google's flagship reasoning model, built for complex software engineering, deep mathematical and scientific problem solving, and autonomous multi-step agents. Google trained it to keep context across...","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":2,"output":10,"reasoning":10,"cache_read":0.1,"cache_write":0.1875},"canonical_model_id":"google/gemini-4-argon"},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemini-3.1-flash-lite-preview":{"id":"google/gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.125,"output":0.75,"reasoning":0.75,"cache_read":0.0125,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.08,"output":0.16},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemini-3.1-flash-lite-image":{"id":"google/gemini-3.1-flash-lite-image","type":"image","name":"Nano Banana 2 Lite","description":"Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) is Google's fastest, most cost-efficient Gemini image model, built for high-velocity developer pipelines and rapid-fire visual exploration. It delivers text-to-image generation...","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["image","text","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":58982},"cost":{"input":0.25,"output":1.5},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"google/gemini-3-pro-image-preview":{"id":"google/gemini-3-pro-image-preview","type":"image","name":"Nano Banana Pro Preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["image","text","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":1,"output":6,"reasoning":6,"cache_read":0.1,"cache_write":0.1875},"canonical_model_id":"google/gemini-3-pro-image-preview"},"google/gemma-3-4b-it":{"id":"google/gemma-3-4b-it","name":"Gemma 3 4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.04,"output":0.05,"cache_read":0.02},"canonical_model_id":"google/gemma-3-4b-it"},"google/gemini-2.5-flash-image":{"id":"google/gemini-2.5-flash-image","type":"image","name":"Nano Banana","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["image","text","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.083333},"canonical_model_id":"google/gemini-2.5-flash-image"},"google/gemini-2.5-pro-preview":{"id":"google/gemini-2.5-pro-preview","name":"Gemini 2.5 Pro Preview 06-05","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01-31","release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["pdf","image","text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.625,"output":5,"reasoning":5,"cache_read":0.0625,"cache_write":0.1875},"canonical_model_id":"google/gemini-2.5-pro-preview"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Gemini 3.5 Flash Lite is a high-efficiency model from Google with upgraded agentic capabilities. It is suited for subagents that execute focused tasks within complex, multi-agent workflows.","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.15,"output":1.25,"reasoning":1.25,"cache_read":0.015,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-3.1-flash-image-preview":{"id":"google/gemini-3.1-flash-image-preview","type":"image","name":"Nano Banana 2 Preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":58982},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image-preview"},"google/gemini-nano-banana-2.1":{"id":"google/gemini-nano-banana-2.1","type":"image","name":"Nano Banana 2.1","description":"Nano Banana 2.1 (Gemini Nano Banana 2.1) is Google's image generation and editing model on the Flash tier, succeeding Nano Banana 2 and Nano Banana Pro. It improves product recontextualization,...","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["image","text","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":58982},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"google/gemini-nano-banana-2.1"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1,"output":6,"reasoning":6,"cache_read":0.1,"cache_write":0.1875},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.042,"output":0.22,"cache_read":0.021},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Gemini 3.6 Flash is a high-efficiency model from Google for coding, agentic workflows, and web and app development. It is designed to produce polished outputs with fewer unnecessary edits and...","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.375,"output":1.875,"reasoning":1.875,"cache_read":0.0375,"cache_write":0.020833},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"Gemini 3.7 Flash is a multimodal model from Google for fast agentic workflows, coding, and complex multi-step reasoning. It is designed for tasks that require responsive performance and reliable multi-step...","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.04,"output":0.12},"canonical_model_id":"google/gemma-3-12b-it"},"google/gemma-2-27b-it":{"id":"google/gemma-2-27b-it","name":"Gemma 2 27B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-07-13","last_updated":"2024-07-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":2048},"cost":{"input":0.65,"output":0.65},"canonical_model_id":"google/gemma-2-27b-it"},"google/lyria-3-clip-preview":{"id":"google/lyria-3-clip-preview","name":"Lyria 3 Clip Preview","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"lyria","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-25","last_updated":"2026-03-25","modalities":{"input":["text","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/lyria-3-clip-preview"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":4.5,"reasoning":4.5,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-3.1-flash-image":{"id":"google/gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Gemini 3.1 Flash Image, a.k.a. \"Nano Banana 2,\" is Google’s latest state of the art image generation and editing model, delivering Pro-level visual quality at Flash speed. It combines advanced...","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["image","text","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image"},"google/lyria-3-pro-preview":{"id":"google/lyria-3-pro-preview","name":"Lyria 3 Pro Preview","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"lyria","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-25","last_updated":"2026-03-25","modalities":{"input":["text","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/lyria-3-pro-preview"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.041667},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.1,"output":0.34,"cache_read":0.1},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"status":"deprecated","cost":{"input":0.05,"output":0.2,"reasoning":0.2,"cache_read":0.005,"cache_write":0.041667},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-3-pro-image":{"id":"google/gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro is Google’s most advanced image-generation and editing model, built on Gemini 3 Pro. It extends the original Nano Banana with significantly improved multimodal reasoning, real-world grounding, and...","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["image","text","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3-pro-image"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.625,"output":5,"reasoning":5,"cache_read":0.0625,"cache_write":0.1875},"canonical_model_id":"google/gemini-2.5-pro"},"dots-studio/dots-3-note-preview:free":{"id":"dots-studio/dots-3-note-preview:free","name":"Dots Studio: Dots3-Note Preview (free)","description":"Dots3-Note Preview is an open-weight mixture-of-experts model from Dots Studio, with 16B active parameters out of 280B total. It is the lightest model in the Dots 3 family and is...","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":512000,"output":460800},"cost":{"input":0,"output":0}},"x-ai/grok-4.5":{"id":"x-ai/grok-4.5","name":"Grok 4.5","description":"Grok 4.5 is SpaceXAI's smartest model with frontier performance on coding, knowledge work, and STEM.","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.3},"canonical_model_id":"xai/grok-4.5"},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"Grok 4.6 is SpaceXAI's smartest model with frontier performance on coding, knowledge work, and STEM.","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"x-ai/grok-4.7":{"id":"x-ai/grok-4.7","name":"Grok 4.7","description":"Grok 4.7 is SpaceXAI's smartest model with frontier performance on coding, knowledge work, and STEM.","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"x-ai/grok-4.20-multi-agent":{"id":"x-ai/grok-4.20-multi-agent","name":"SpaceXAI: Grok 4.20 Multi-Agent","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":1800000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2}},"x-ai/grok-4.20":{"id":"x-ai/grok-4.20","name":"SpaceXAI: Grok 4.20","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":1800000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2}},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":900000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"x-ai/grok-build-0.1":{"id":"x-ai/grok-build-0.1","name":"Grok Build 0.1","description":"Grok coding model for agentic engineering, edits, and codebase workflows","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":230400},"cost":{"input":1,"output":2,"cache_read":0.2},"canonical_model_id":"xai/grok-build-0.1"},"openrouter/free":{"id":"openrouter/free","name":"OpenRouter Free Models Router","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-01","last_updated":"2026-02-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32768},"cost":{"input":0,"output":0}},"openrouter/bodybuilder":{"id":"openrouter/bodybuilder","name":"Body Builder (beta)","description":"Preview model for early access evaluation, prototyping, and compatibility testing","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-03-15","last_updated":"2026-03-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"status":"beta","cost":{"input":0,"output":0}},"openrouter/pareto-code":{"id":"openrouter/pareto-code","name":"Pareto Code Router","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-05-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":65536},"cost":{"input":0,"output":0}},"openrouter/auto":{"id":"openrouter/auto","name":"Auto Router","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-15","last_updated":"2026-03-15","modalities":{"input":["audio","image","pdf","text","video"],"output":["image","text"]},"open_weights":false,"limit":{"context":2000000,"output":32768},"cost":{"input":0,"output":0}},"arcee-ai/trinity-large-thinking":{"id":"arcee-ai/trinity-large-thinking","name":"Trinity Large Thinking","description":"Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use","family":"trinity","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":80000},"cost":{"input":0.25,"output":0.8,"cache_read":0.06},"canonical_model_id":"arcee-ai/trinity-large-thinking"},"rekaai/reka-edge":{"id":"rekaai/reka-edge","name":"Reka Edge","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"reka","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":14745},"cost":{"input":0.1,"output":0.1},"canonical_model_id":"rekaai/reka-edge"},"rekaai/reka-flash-3":{"id":"rekaai/reka-flash-3","name":"Reka Flash 3","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"reka","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01-31","release_date":"2025-03-10","last_updated":"2025-03-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":58982},"cost":{"input":0.1,"output":0.2},"canonical_model_id":"rekaai/reka-flash-3"}}},"subconscious":{"id":"subconscious","env":["SUBCONSCIOUS_API_KEY"],"npm":"@ai-sdk/anthropic","api":"https://api.subconscious.dev/v1","name":"Subconscious","doc":"https://docs.subconscious.dev","models":{"subconscious/glm-5.3-marathon":{"id":"subconscious/glm-5.3-marathon","name":"GLM-5.3 Marathon","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":5000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"subconscious/deepseek-v4.1-flash-marathon":{"id":"subconscious/deepseek-v4.1-flash-marathon","name":"DeepSeek V4.1 Flash Marathon","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":5000000,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4.1-flash"}}},"cline-pass":{"id":"cline-pass","env":["CLINE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.cline.bot/api/v1","name":"ClinePass","doc":"https://docs.cline.bot/getting-started/clinepass","models":{"cline-pass/qwen3.7-max":{"id":"cline-pass/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.5,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.7-max"},"cline-pass/qwen3.7-plus":{"id":"cline-pass/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":1.6,"cache_read":0.04,"cache_write":0.5,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.12,"cache_write":1.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.12,"cache_write":1.5}},"canonical_model_id":"alibaba/qwen3.7-plus"},"cline-pass/mimo-v2.5":{"id":"cline-pass/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.5"},"cline-pass/mimo-v2.5-pro":{"id":"cline-pass/mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.74,"output":3.48,"cache_read":0.0145},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"cline-pass/deepseek-v4-flash":{"id":"cline-pass/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"cline-pass/kimi-k2.6":{"id":"cline-pass/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"cline-pass/muse-spark-1.3-contributor":{"id":"cline-pass/muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.002},"canonical_model_id":"meta/muse-spark-1.3"},"cline-pass/mimo-v2.6-pro":{"id":"cline-pass/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"cline-pass/mimo-v2.6-flash":{"id":"cline-pass/mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"cline-pass/deepseek-v4-pro":{"id":"cline-pass/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro"},"cline-pass/qwen3.8-max":{"id":"cline-pass/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"cline-pass/kimi-k2.7-code":{"id":"cline-pass/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"cline-pass/kimi-k3":{"id":"cline-pass/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"cline-pass/glm-5.3-flash":{"id":"cline-pass/glm-5.3-flash","name":"cline-pass/glm-5.3-flash","description":"Latest natively multimodal model in the GLM-5 series","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"cline-pass/minimax-m3":{"id":"cline-pass/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"cline-pass/glm-5.2":{"id":"cline-pass/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"cline-pass/deepseek-v4.1-flash":{"id":"cline-pass/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"cline-pass/glm-5.3":{"id":"cline-pass/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"}}},"wandb":{"id":"wandb","env":["WANDB_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.inference.wandb.ai/v1","name":"CoreWeave","doc":"https://docs.wandb.ai/inference","models":{"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Kimi K2.7 Code is a 1T-parameter MoE model with 32B active parameters purpose-built for long-horizon agentic coding and software engineering.","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.71,"output":3.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Kimi K2.6 is a multimodal Mixture-of-Experts language model featuring 32 billion activated parameters and a total of 1 trillion parameters.","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.65,"output":3.41,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2.6"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax M3","description":"MiniMax M3 is a multimodal MoE model with 23B active parameters optimized for coding and agentic workflows.","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.23,"output":0.96,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M3"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"gpt-oss-20b","description":"Lower latency Mixture-of-Experts model trained on OpenAI's Harmony response format with reasoning capabilities.","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.03,"output":0.13,"cache_read":0.03},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"gpt-oss-120b","description":"Efficient Mixture-of-Experts model designed for high-reasoning, agentic and general-purpose use cases.","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.03,"output":0.17,"cache_read":0.03},"canonical_model_id":"openai/gpt-oss-120b"},"ibm-granite/granite-4.2-8b":{"id":"ibm-granite/granite-4.2-8b","name":"Granite 4.2 8B","description":"Granite 4.2 8B is an instruct model capable of enhanced tool calling, instruction following, and chat capabilities.","family":"granite","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-25","last_updated":"2026-08-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.1,"output":0.15,"cache_read":0.05},"canonical_model_id":"ibm/granite-4.2-8b"},"meta-llama/Llama-3.3-70B-Instruct":{"id":"meta-llama/Llama-3.3-70B-Instruct","name":"Llama 3.3 70B","description":"Multilingual model excelling in conversational tasks, detailed instruction-following, and coding.","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.71,"output":0.71,"cache_read":0.71},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/Llama-3.1-8B-Instruct":{"id":"meta-llama/Llama-3.1-8B-Instruct","name":"Llama 3.1 8B","description":"Efficient conversational model optimized for responsive multilingual chatbot interactions.","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.22,"output":0.22,"cache_read":0.22},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM 5.2","description":"GLM-5.2 is a Mixture-of-Experts language model featuring 40 billion activated parameters and a total of 744 billion parameters.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.62,"output":2.42,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM 5.3","description":"GLM-5.3 is a 744B-parameter MoE model built for advanced reasoning, software engineering, and long-horizon agentic tasks.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.2,"output":4,"cache_read":0.19},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM 5.3 Flash","description":"GLM-5.3-Flash is a natively multimodal model with 320B total parameters and 18B active parameters.","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.15,"output":0.5,"cache_read":0.02},"canonical_model_id":"zhipuai/glm-5.3-flash"},"reflection/Beam-1.0-501B-A23B":{"id":"reflection/Beam-1.0-501B-A23B","name":"Beam 1.0 501B A23B","description":"Beam is a 501B-parameter MoE model with 23B active parameters built for coding, reasoning, and agentic workflows.","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.3,"output":1,"cache_read":0.04},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B":{"id":"nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B","name":"Nemotron 3.5 Super VL","description":"Nemotron 3.5 Super VL is a multimodal MoE model built for visual understanding, reasoning, and agentic workflows.","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.15,"output":0.5,"cache_read":0.02},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B":{"id":"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B","name":"Nemotron 3 Ultra","description":"Nemotron 3 Ultra is a powerful MoE model designed for long-running agents across coding, deep research, and enterprise automation.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":2.15,"cache_read":0.1},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B":{"id":"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B","name":"Nemotron 3.5 Lightning","description":"Nemotron 3.5 Lightning is an MoE model built for fast, reliable agentic tasks across use cases such as financial services, cybersecurity, telecom, and retail.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.07,"output":0.2,"cache_read":0.04},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"DeepSeek V4-Flash-0731 is an MoE model great for coding, reasoning, and agentic workloads.","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.13,"output":0.28,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek V3.1","description":"A large hybrid model that supports both thinking and non-thinking modes via prompt templates.","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":161000,"output":161000},"cost":{"input":0.55,"output":1.65,"cache_read":0.55},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4-Pro-0813 is a 1.6T-parameter MoE model excelling at advanced reasoning, coding, and complex agentic workloads.","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.31,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek-ai/DeepSeek-V4.1-Flash":{"id":"deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash is a multimodal MoE model for coding, reasoning, and agentic workloads with long contexts.","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.15,"output":0.65,"cache_read":0.002},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"Qwen/Qwen3.6-35B-A3B":{"id":"Qwen/Qwen3.6-35B-A3B","name":"Qwen3.6 35B A3B","description":"Qwen3.6-35B-A3B is an MoE multimodal model with 262K context optimized for agentic coding workflows.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.25,"output":1.25,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8 27B","description":"Qwen3.8-27B is a dense multimodal model suited for coding, research, vision, and long-running agent tasks.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.4,"output":3,"cache_read":0.15},"canonical_model_id":"alibaba/qwen3.8-27b"},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B","description":"Gemma 4 31B Dense is designed for advanced reasoning, agentic workflows, and longer context and is natively trained on 140+ languages.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.34,"cache_read":0.1},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemma-4-26B-A4B-it":{"id":"google/gemma-4-26B-A4B-it","name":"Gemma 4 26B A4B","description":"Gemma 4 26B A4B is a multimodal MoE model with LoRA support and function calling for agentic workflows.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.3,"cache_read":0.05},"canonical_model_id":"google/gemma-4-26b-a4b-it"}}},"baseten":{"id":"baseten","env":["BASETEN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.baseten.co/v1","name":"Baseten","doc":"https://docs.baseten.co/inference/model-apis/overview","models":{"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"status":"deprecated","cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K2.5":{"id":"moonshotai/Kimi-K2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"status":"deprecated","cost":{"input":0.6,"output":3,"cache_read":0.12},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"status":"deprecated","cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax-M2.5","description":"Legacy model retained for compatibility with older integrations","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2026-01","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204000,"output":204000},"status":"deprecated","cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.5"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"status":"deprecated","cost":{"input":0.5,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"status":"deprecated","cost":{"input":1,"output":4.05},"canonical_model_id":"thinkingmachines/inkling"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"OpenAI GPT 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128072,"output":128072},"cost":{"input":0.1,"output":0.5,"cache_read":0.1},"canonical_model_id":"openai/gpt-oss-120b"},"zai-org/GLM-5":{"id":"zai-org/GLM-5","name":"GLM 5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":202800},"status":"deprecated","cost":{"input":0.95,"output":3.15,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.2-Fast":{"id":"zai-org/GLM-5.2-Fast","name":"GLM 5.2 Fast","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":2.1,"output":6.6,"cache_read":0.21},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM 5.3","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"GLM 5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":202800},"status":"deprecated","cost":{"input":1.3,"output":4.3,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-5.3-Fast":{"id":"zai-org/GLM-5.3-Fast","name":"GLM 5.3 Fast","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":2.1,"output":6.6,"cache_read":0.21},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-4.7":{"id":"zai-org/GLM-4.7","name":"GLM 4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":200000},"status":"deprecated","cost":{"input":0.6,"output":2.2,"cache_read":0.12},"canonical_model_id":"zhipuai/glm-4.7"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM 5.3 Flash","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"nvidia/Nemotron-120B-A12B":{"id":"nvidia/Nemotron-120B-A12B","name":"Nemotron Super","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":202800},"status":"deprecated","cost":{"input":0.3,"output":0.75,"cache_read":0.06},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B":{"id":"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B","name":"Nemotron Ultra","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":202800},"cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"deepseek-ai/DeepSeek-V4.1-Flash-1017":{"id":"deepseek-ai/DeepSeek-V4.1-Flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0.3,"output":1.2,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.13,"output":0.26,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"status":"deprecated","cost":{"input":1.74,"output":3.48,"cache_read":0.145},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/DeepSeek-V4.1-Flash-1017-Fast":{"id":"deepseek-ai/DeepSeek-V4.1-Flash-1017-Fast","name":"DeepSeek V4.1 Flash 1017 Fast","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0.6,"output":2.4,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek V3.1","description":"Legacy model retained for compatibility with older integrations","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":131000},"status":"deprecated","cost":{"input":0.5,"output":1.5},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek-ai/DeepSeek-V4.1-Flash-Fast":{"id":"deepseek-ai/DeepSeek-V4.1-Flash-Fast","name":"DeepSeek V4.1 Flash Fast","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.6,"output":2.4,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":1.32,"output":3.96,"cache_read":0.132},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek-ai/DeepSeek-V4.1-Flash":{"id":"deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.3,"output":1.2,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4.1-flash"}}},"engy":{"id":"engy","env":["ENGY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.engy.ai/v1","name":"engy","doc":"https://engy.ai/pricing","models":{"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1001536,"input":936000,"output":65536},"cost":{"input":0.045,"output":0.32,"cache_read":0.015},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":208192,"input":200000,"output":8192},"cost":{"input":0.045,"output":0.3,"cache_read":0.015},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":983040,"output":65536},"cost":{"input":0.04,"output":0.08,"cache_read":0.008},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1113088,"input":1047552,"output":65536},"cost":{"input":1.95,"output":9.75,"cache_read":0.195},"canonical_model_id":"moonshotai/kimi-k3"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":327680,"input":294912,"output":32768},"cost":{"input":0.135,"output":0.45,"cache_read":0.027},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":229376,"output":32768},"cost":{"input":0.68,"output":1.5,"cache_read":0.18},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":983040,"output":65536},"cost":{"input":0.04,"output":0.08,"cache_read":0.008},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":920576,"output":128000},"cost":{"input":0.045,"output":0.09,"cache_read":0.009},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":327680,"input":294912,"output":32768},"cost":{"input":0.98,"output":3.08,"cache_read":0.18},"canonical_model_id":"zhipuai/glm-5.3"}}},"openai":{"id":"openai","env":["OPENAI_API_KEY"],"npm":"@ai-sdk/openai","name":"OpenAI","doc":"https://platform.openai.com/docs/models","models":{"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}}}},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":2.5,"output":20,"cache_read":0.25},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.625,"output":5,"cache_read":0.0625},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"experimental":{"modes":{"fast":{"cost":{"input":0.7,"output":2.8,"cache_read":0.175},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.4,"output":1.6,"cache_read":0.1}},"gpt-4o-2024-05-13":{"id":"gpt-4o-2024-05-13","name":"GPT-4o (2024-05-13)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-05-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","experimental":{"modes":{"fast":{"cost":{"input":8.75,"output":26.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":15}},"gpt-5-pro":{"id":"gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":15,"output":120}},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":20,"output":100,"cache_read":2,"cache_write":25},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}},"ultrafast":{"cost":{"input":60,"output":300,"cache_read":6,"cache_write":75},"provider":{"body":{"service_tier":"ultrafast"}}}}},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"chatgpt-image-latest":{"id":"chatgpt-image-latest","type":"image","name":"chatgpt-image-latest","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":0,"input":0,"output":0}},"o3-mini":{"id":"o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.55}},"gpt-5.5-pro":{"id":"gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}}},"gpt-4":{"id":"gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"status":"deprecated","cost":{"input":30,"output":60}},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":4,"output":24,"cache_read":0.4,"cache_write":5},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}}}},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":0.45,"output":3.6,"cache_read":0.045},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.125,"output":1,"cache_read":0.0125},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":28,"cache_read":0.35},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"o1-pro":{"id":"o1-pro","name":"o1-pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2025-03-19","last_updated":"2025-03-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":150,"output":600}},"text-embedding-3-large":{"id":"text-embedding-3-large","type":"embedding","name":"text-embedding-3-large","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-01","release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8191,"output":3072},"cost":{"input":0.13,"output":0}},"gpt-5.4-pro":{"id":"gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}}},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"flex":{"cost":{"input":0.025,"output":0.2,"cache_read":0.0025},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":0.05,"output":0.4,"cache_read":0.005}},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":28,"cache_read":0.35},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.875,"output":7,"cache_read":0.0875},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}}},"o3":{"id":"o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":14,"cache_read":0.875},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":1,"output":4,"cache_read":0.25},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":2,"output":8,"cache_read":0.5}},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}}},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":0.2,"output":1,"cache_read":0.02,"cache_write":0.25},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}}}},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"gpt-5.3-codex-spark":{"id":"gpt-5.3-codex-spark","name":"GPT-5.3 Codex Spark","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex-spark","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":100000,"output":32000},"cost":{"input":1.75,"output":14,"cache_read":0.175}},"gpt-image-2":{"id":"gpt-image-2","type":"image","name":"gpt-image-2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"input":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":8,"output":40,"cache_read":0.8,"cache_write":10},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}}}},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"gpt-5.6":{"id":"gpt-5.6","name":"GPT-5.6","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":8,"output":40,"cache_read":0.8,"cache_write":10},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}}}},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"text-embedding-3-small":{"id":"text-embedding-3-small","type":"embedding","name":"text-embedding-3-small","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-01","release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8191,"output":1536},"cost":{"input":0.02,"output":0}},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}},"ultrafast":{"cost":{"input":12,"output":60,"cache_read":0.6,"cache_write":15},"provider":{"body":{"service_tier":"ultrafast"}}}}},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":0.4,"output":2.4,"cache_read":0.04,"cache_write":0.5},"provider":{"body":{"service_tier":"priority"}}},"pro":{"provider":{"body":{"reasoning":{"mode":"pro"}}}}}},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"gpt-image-1-mini":{"id":"gpt-image-1-mini","type":"image","name":"gpt-image-1-mini","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-09-26","last_updated":"2025-09-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":0,"input":0,"output":0}},"gpt-image-1":{"id":"gpt-image-1","type":"image","name":"gpt-image-1","description":"OpenAI image model for production generation, edits, and brand-safe visual workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-24","last_updated":"2025-04-24","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"input":0,"output":0},"status":"deprecated","canonical_model_id":"openai/gpt-image-1"},"gpt-realtime-2.1":{"id":"gpt-realtime-2.1","type":"realtime","name":"GPT-Realtime-2.1","description":"Realtime speech-to-speech model with configurable reasoning, tool use, and robust voice-agent behavior","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2024-09-30","release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text","audio","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":128000,"input":96000,"output":32000},"cost":{"input":4,"output":24,"cache_read":0.4,"input_audio":32,"output_audio":64},"canonical_model_id":"openai/gpt-realtime-2.1"},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"experimental":{"modes":{"fast":{"cost":{"input":4.25,"output":17,"cache_read":2.125},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":10,"cache_read":1.25}},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","experimental":{"modes":{"fast":{"cost":{"input":0.2,"output":0.8,"cache_read":0.05},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.1,"output":0.4,"cache_read":0.025}},"gpt-4o-mini":{"id":"gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"experimental":{"modes":{"fast":{"cost":{"input":0.25,"output":1,"cache_read":0.125},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.15,"output":0.6,"cache_read":0.075}},"gpt-image-1.5":{"id":"gpt-image-1.5","type":"image","name":"gpt-image-1.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":0,"input":0,"output":0},"canonical_model_id":"openai/gpt-image-1.5"},"gpt-4o-2024-11-20":{"id":"gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25}},"gpt-4o-2024-08-06":{"id":"gpt-4o-2024-08-06","name":"GPT-4o (2024-08-06)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25}},"gpt-4-turbo":{"id":"gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":10,"output":30}},"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","experimental":{"modes":{"fast":{"cost":{"input":2,"output":8,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.55,"output":2.2,"cache_read":0.1375},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":1.1,"output":4.4,"cache_read":0.275}},"gpt-3.5-turbo":{"id":"gpt-3.5-turbo","name":"GPT-3.5-turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"status":"deprecated","cost":{"input":0.5,"output":1.5,"cache_read":0}},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":14,"cache_read":0.875},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2,"output":8,"cache_read":0.5}},"o1":{"id":"o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":15,"output":60,"cache_read":7.5}},"gpt-daybreak-red-latest":{"id":"gpt-daybreak-red-latest","name":"Daybreak Red","description":"Alias for advanced cybersecurity models for authorized vulnerability research and security testing","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-08-07","last_updated":"2026-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":12.5,"output":75,"cache_read":1.25,"cache_write":15.625},"canonical_model_id":"openai/gpt-5.6-cyber"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":2.5,"output":20,"cache_read":0.25},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.625,"output":5,"cache_read":0.0625},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"o3-pro":{"id":"o3-pro","name":"o3-pro","description":"High-effort o3 tier for difficult technical reasoning and careful answers","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":20,"output":80}},"gpt-daybreak-blue-latest":{"id":"gpt-daybreak-blue-latest","name":"Daybreak Blue","description":"Alias for flagship general-purpose models with safeguards for defensive cybersecurity work","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-08-07","last_updated":"2026-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"gpt-5.2-pro":{"id":"gpt-5.2-pro","name":"GPT-5.2 Pro","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":21,"output":168}},"text-embedding-ada-002":{"id":"text-embedding-ada-002","type":"embedding","name":"text-embedding-ada-002","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2022-12","release_date":"2022-12-15","last_updated":"2022-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536},"cost":{"input":0.1,"output":0}},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":1.5,"output":9,"cache_read":0.15},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.375,"output":2.25,"cache_read":0.0375},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":0.75,"output":4.5,"cache_read":0.075}},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"flex":{"cost":{"input":0.1,"output":0.625,"cache_read":0.01},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":0.2,"output":1.25,"cache_read":0.02}}}},"scnet-token-plan":{"id":"scnet-token-plan","env":["SCNET_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.scnet.cn/api/llm/v1","name":"SCNet Token Plan","doc":"https://www.scnet.cn/ac/openapi/doc/2.0/moduleapi/plans/token-plan.html","models":{"Kimi-K3":{"id":"Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k3"},"Kimi-K2.7-Code":{"id":"Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"GLM-5":{"id":"GLM-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5"},"Kimi-K2.5":{"id":"Kimi-K2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k2.5"},"DeepSeek-V4-Flash-0731":{"id":"DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"DeepSeek-V4-Pro":{"id":"DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-pro"},"GLM-5.2":{"id":"GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.2"},"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"minimax/MiniMax-M2.7"},"DeepSeek-V4-Flash":{"id":"DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"Qwen3.8-Flash":{"id":"Qwen3.8-Flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"alibaba/qwen3.8-flash"},"DeepSeek-V4-Pro-0813":{"id":"DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"DeepSeek-V4.1-Flash":{"id":"DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"Kimi-K2.6":{"id":"Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"Qwen3.8-Max":{"id":"Qwen3.8-Max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"alibaba/qwen3.8-max"},"GLM-5.3":{"id":"GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.3"},"GLM-5.1":{"id":"GLM-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.1"},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"minimax/MiniMax-M2.5"},"GLM-5.3-Flash":{"id":"GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.3-flash"}}},"inference":{"id":"inference","env":["INFERENCE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.net/v1","name":"Inference","doc":"https://inference.net/models","models":{"meta/llama-3.1-8b-instruct":{"id":"meta/llama-3.1-8b-instruct","name":"Llama 3.1 8B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4096},"status":"deprecated","cost":{"input":0.025,"output":0.025},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta/llama-3.2-3b-instruct":{"id":"meta/llama-3.2-3b-instruct","name":"Llama 3.2 3B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4096},"cost":{"input":0.02,"output":0.02}},"meta/llama-3.2-11b-vision-instruct":{"id":"meta/llama-3.2-11b-vision-instruct","name":"Llama 3.2 11B Vision Instruct","description":"Open Llama multimodal model for image understanding and text reasoning","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4096},"status":"deprecated","cost":{"input":0.055,"output":0.055},"canonical_model_id":"meta/llama-3.2-11b-vision-instruct"},"meta/llama-3.2-1b-instruct":{"id":"meta/llama-3.2-1b-instruct","name":"Llama 3.2 1B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4096},"cost":{"input":0.01,"output":0.01},"canonical_model_id":"meta/llama-3.2-1b-instruct"},"mistral/mistral-nemo-12b-instruct":{"id":"mistral/mistral-nemo-12b-instruct","name":"Mistral Nemo 12B Instruct","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4096},"status":"deprecated","cost":{"input":0.038,"output":0.1},"canonical_model_id":"mistral/mistral-nemo"},"osmosis/osmosis-structure-0.6b":{"id":"osmosis/osmosis-structure-0.6b","name":"Osmosis Structure 0.6B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"osmosis","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4000,"output":2048},"status":"deprecated","cost":{"input":0.1,"output":0.5}},"qwen/qwen3-embedding-4b":{"id":"qwen/qwen3-embedding-4b","type":"embedding","name":"Qwen 3 Embedding 4B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-12","release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":2048},"status":"deprecated","cost":{"input":0.01,"output":0},"canonical_model_id":"alibaba/qwen3-embedding-4b"},"qwen/qwen-2.5-7b-vision-instruct":{"id":"qwen/qwen-2.5-7b-vision-instruct","name":"Qwen 2.5 7B Vision Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":125000,"output":4096},"status":"deprecated","cost":{"input":0.2,"output":0.2}},"google/gemma-3":{"id":"google/gemma-3","name":"Google Gemma 3","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":125000,"output":4096},"cost":{"input":0.15,"output":0.3}}}},"deepseek":{"id":"deepseek","env":["DEEPSEEK_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.deepseek.com","name":"DeepSeek","doc":"https://api-docs.deepseek.com/quick_start/pricing","models":{"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"status":"deprecated","cost":{"input":0.15,"output":0.6,"reasoning":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-flash":{"id":"deepseek-flash","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.15,"output":0.6,"reasoning":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-v4-flash-vision-exp":{"id":"deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"status":"deprecated","cost":{"input":0.15,"output":0.6,"reasoning":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.66,"output":1.98,"reasoning":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"}}},"anthropic":{"id":"anthropic","env":["ANTHROPIC_API_KEY"],"npm":"@ai-sdk/anthropic","name":"Anthropic","doc":"https://docs.anthropic.com/en/docs/about-claude/models","models":{"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-14","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-07","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5}},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":8,"output":40,"cache_read":0.4,"cache_write":10},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"claude-opus-4-5-20251101":{"id":"claude-opus-4-5-20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-04","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25}},"claude-sonnet-4-5-20250929":{"id":"claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25}},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-29","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5}}}},"evroc":{"id":"evroc","env":["EVROC_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://models.think.evroc.com/v1","name":"evroc","doc":"https://docs.evroc.com/products/think/overview.html","models":{"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.4375,"output":5.75},"canonical_model_id":"moonshotai/kimi-k2.6"},"mistralai/Mistral-Medium-3.5-128B":{"id":"mistralai/Mistral-Medium-3.5-128B","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.725,"output":6.9},"canonical_model_id":"mistral/mistral-medium-2604"},"mistralai/Voxtral-Small-24B-2507":{"id":"mistralai/Voxtral-Small-24B-2507","type":"transcription","name":"Voxtral Small 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"voxtral","attachment":true,"reasoning":false,"tool_call":false,"release_date":"2025-03-01","last_updated":"2025-03-01","modalities":{"input":["audio","text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000}},"intfloat/multilingual-e5-large-instruct":{"id":"intfloat/multilingual-e5-large-instruct","type":"embedding","name":"E5 Multi-Lingual Large Embeddings 0.6B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2024-06-01","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":512},"cost":{"input":0.114,"output":0.114}},"openai/whisper-large-v3-turbo":{"id":"openai/whisper-large-v3-turbo","type":"transcription","name":"Whisper Large v3 Turbo","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-10-01","last_updated":"2024-10-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":448},"cost":{"input":0.0023,"output":0.0023,"output_audio":2.3},"canonical_model_id":"openai/whisper-large-v3-turbo"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.23,"output":0.92},"canonical_model_id":"openai/gpt-oss-120b"},"openai/whisper-large-v3":{"id":"openai/whisper-large-v3","type":"transcription","name":"Whisper 3 Large","description":"Open Whisper checkpoint for robust multilingual transcription and captioning","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":4096},"cost":{"input":0.0023,"output":0.0023,"output_audio":2.3},"canonical_model_id":"openai/whisper-large-v3"},"evroc/roc":{"id":"evroc/roc","name":"roc","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-06-06","last_updated":"2026-06-06","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":2.875,"output":11.516}},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":131072},"cost":{"input":1.4375,"output":5.75},"canonical_model_id":"zhipuai/glm-5.2"},"nvidia/Llama-3.3-70B-Instruct-FP8":{"id":"nvidia/Llama-3.3-70B-Instruct-FP8","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":1.15,"output":1.15},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"Qwen/Qwen3.6-35B-A3B":{"id":"Qwen/Qwen3.6-35B-A3B","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.345,"output":1.38},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen/Qwen3-Reranker-4B":{"id":"Qwen/Qwen3-Reranker-4B","type":"reranking","name":"Qwen3 Reranker 4B","description":"Reranking model for improving retrieval quality in search and recommendation systems","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.0575,"output":0}},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8-27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.87,"output":3.5},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3-Embedding-8B":{"id":"Qwen/Qwen3-Embedding-8B","type":"embedding","name":"Qwen3 Embedding 8B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":4096},"cost":{"input":0.115,"output":0.115},"canonical_model_id":"alibaba/qwen3-embedding-8b"},"KBLab/kb-whisper-large":{"id":"KBLab/kb-whisper-large","type":"transcription","name":"KB Whisper","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"release_date":"2024-10-01","last_updated":"2024-10-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":448}},"google/gemma-4-26B-A4B-it":{"id":"google/gemma-4-26B-A4B-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.144,"output":0.575},"canonical_model_id":"google/gemma-4-26b-a4b-it"}}},"qvac":{"id":"qvac","env":["QVAC_API_KEY"],"npm":"@qvac/ai-sdk-provider","name":"QVAC","doc":"https://www.npmjs.com/package/@qvac/ai-sdk-provider","models":{"qwen3.5-0.8b":{"id":"qwen3.5-0.8b","name":"Qwen3.5 0.8B","description":"Qwen instruction model for multilingual chat and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-02","last_updated":"2026-03-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.5-0.8b"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-oss-20b"},"qwen3.5-4b":{"id":"qwen3.5-4b","name":"Qwen3.5 4B","description":"Qwen instruction model for multilingual chat and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-02","last_updated":"2026-03-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.5-4b"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-oss-120b"},"qwen3.5-2b":{"id":"qwen3.5-2b","name":"Qwen3.5 2B","description":"Qwen instruction model for multilingual chat and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-02","last_updated":"2026-03-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.5-2b"},"gemma4-31b":{"id":"gemma4-31b","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-31b-it"},"qwen3.5-9b":{"id":"qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.5-9b"},"qwen3.6-27b":{"id":"qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.6-27b"}}},"volcengine-coding-plan":{"id":"volcengine-coding-plan","env":["ARK_CODING_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://ark.cn-beijing.volces.com/api/coding/v3","name":"Volcengine Ark Coding Plan","doc":"https://www.volcengine.com/docs/82379/1928261","models":{"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-pro"},"doubao-seed-2.0-lite":{"id":"doubao-seed-2.0-lite","name":"Seed 2.0 Lite","description":"Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"doubao-seed-evolving":{"id":"doubao-seed-evolving","name":"Seed Evolving","description":"Rolling ByteDance Seed model for rapidly updated reasoning, coding, and agent capabilities","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"bytedance-seed/seed-evolving"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k3"},"doubao-seed-2.1-turbo":{"id":"doubao-seed-2.1-turbo","name":"Seed 2.1 Turbo","description":"Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"minimax/MiniMax-M3"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.3"}}},"pendra":{"id":"pendra","env":["PENDRA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.pendra.ai/api/v1","name":"Pendra","doc":"https://pendra.ai/docs/integrations/opencode","models":{"qwen3-coder:30b":{"id":"qwen3-coder:30b","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"qwen3.6:27b":{"id":"qwen3.6:27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.6-27b"},"glm-4.7-flash":{"id":"glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"gpt-oss:120b":{"id":"gpt-oss:120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-oss-120b"},"llama3.3:70b":{"id":"llama3.3:70b","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-3.3-70b-instruct"}}},"fireworks-ai":{"id":"fireworks-ai","env":["FIREWORKS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.fireworks.ai/inference/v1/","name":"Fireworks AI","doc":"https://fireworks.ai/docs/","models":{"accounts/fireworks/routers/minimax-latest":{"id":"accounts/fireworks/routers/minimax-latest","name":"MiniMax Latest","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":512000},"experimental":{"modes":{"priority":{"cost":{"input":0.45,"output":1.8,"cache_read":0.09},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"accounts/fireworks/routers/qwen-max-latest":{"id":"accounts/fireworks/routers/qwen-max-latest","name":"Qwen Max Latest (Qwen3.8 Max)","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"experimental":{"modes":{"priority":{"cost":{"input":3,"output":9,"cache_read":0.375},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max"},"accounts/fireworks/routers/glm-latest":{"id":"accounts/fireworks/routers/glm-latest","name":"GLM Latest","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048573,"output":262144},"experimental":{"modes":{"priority":{"cost":{"input":1.75,"output":5.5,"cache_read":0.325},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"accounts/fireworks/routers/kimi-latest":{"id":"accounts/fireworks/routers/kimi-latest","name":"Kimi Latest","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"experimental":{"modes":{"priority":{"cost":{"input":3.75,"output":18.75,"cache_read":0.375},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"accounts/fireworks/routers/glm-5p3-fast":{"id":"accounts/fireworks/routers/glm-5p3-fast","name":"GLM 5.3 Fast","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048572,"output":262144},"cost":{"input":2.1,"output":6.6,"cache_read":0.39},"canonical_model_id":"zhipuai/glm-5.3"},"accounts/fireworks/routers/kimi-fast-latest":{"id":"accounts/fireworks/routers/kimi-fast-latest","name":"Kimi Fast Latest","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.5,"output":22.5,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"accounts/fireworks/routers/deepseek-flash-latest":{"id":"accounts/fireworks/routers/deepseek-flash-latest","name":"DeepSeek Flash Latest","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"experimental":{"modes":{"priority":{"cost":{"input":0.375,"output":1.5,"cache_read":0.0075},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"accounts/fireworks/routers/kimi-k3-fast":{"id":"accounts/fireworks/routers/kimi-k3-fast","name":"Kimi K3 Fast","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.5,"output":22.5,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"accounts/fireworks/routers/glm-fast-latest":{"id":"accounts/fireworks/routers/glm-fast-latest","name":"GLM 5.3 Fast (Latest)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048572,"output":262144},"cost":{"input":2.1,"output":6.6,"cache_read":0.39},"canonical_model_id":"zhipuai/glm-5.3"},"accounts/fireworks/routers/glm-flash-latest":{"id":"accounts/fireworks/routers/glm-flash-latest","name":"GLM Flash Latest (GLM 5.3 Flash)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048573,"output":131072},"experimental":{"modes":{"priority":{"cost":{"input":0.1875,"output":0.625,"cache_read":0.0375},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"accounts/fireworks/models/qwen3p8-max":{"id":"accounts/fireworks/models/qwen3p8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"experimental":{"modes":{"priority":{"cost":{"input":3,"output":9,"cache_read":0.375},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max"},"accounts/fireworks/models/deepseek-v4p1-flash":{"id":"accounts/fireworks/models/deepseek-v4p1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"experimental":{"modes":{"priority":{"cost":{"input":0.375,"output":1.5,"cache_read":0.0075},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"accounts/fireworks/models/deepseek-v4p1-flash-1017":{"id":"accounts/fireworks/models/deepseek-v4p1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"experimental":{"modes":{"priority":{"cost":{"input":0.375,"output":1.5,"cache_read":0.0075},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"accounts/fireworks/models/gpt-oss-120b":{"id":"accounts/fireworks/models/gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","experimental":{"modes":{"priority":{"cost":{"input":0.18,"output":0.72,"cache_read":0.018},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"openai/gpt-oss-120b"},"accounts/fireworks/models/glm-5p3":{"id":"accounts/fireworks/models/glm-5p3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048573,"output":262144},"experimental":{"modes":{"priority":{"cost":{"input":1.75,"output":5.5,"cache_read":0.325},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"accounts/fireworks/models/kimi-k3":{"id":"accounts/fireworks/models/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"experimental":{"modes":{"priority":{"cost":{"input":3.75,"output":18.75,"cache_read":0.375},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"accounts/fireworks/models/minimax-m3":{"id":"accounts/fireworks/models/minimax-m3","name":"MiniMax-M3","description":"Fireworks text-only MiniMax coding model for long-context reasoning and agent tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":512000},"experimental":{"modes":{"priority":{"cost":{"input":0.45,"output":1.8,"cache_read":0.09},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"accounts/fireworks/models/qwen3p8-2p4t-a95b":{"id":"accounts/fireworks/models/qwen3p8-2p4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"accounts/fireworks/models/nemotron-lightning-3p5-30b-a3b":{"id":"accounts/fireworks/models/nemotron-lightning-3p5-30b-a3b","name":"Nemotron 3.5 Lightning 30B A3B","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","experimental":{"modes":{"priority":{"cost":{"input":0.0625,"output":0.25,"cache_read":0.0125},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.05,"output":0.2,"cache_read":0.01},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"accounts/fireworks/models/glm-5p3-flash":{"id":"accounts/fireworks/models/glm-5p3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048573,"output":131072},"experimental":{"modes":{"priority":{"cost":{"input":0.1875,"output":0.625,"cache_read":0.0375},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"accounts/fireworks/models/nemotron-3-ultra-nvfp4":{"id":"accounts/fireworks/models/nemotron-3-ultra-nvfp4","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"status":"deprecated","cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"accounts/fireworks/models/inkling":{"id":"accounts/fireworks/models/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"status":"deprecated","cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"accounts/fireworks/models/ember-1":{"id":"accounts/fireworks/models/ember-1","name":"Ember-1","description":"Specialized model from Fireworks built on Kimi K3, producing shorter reasoning traces with approximately 40% fewer tokens while maintaining comparable quality","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"fireworks/ember-1"},"accounts/fireworks/models/muse-spark-open":{"id":"accounts/fireworks/models/muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"}}},"tokengo":{"id":"tokengo","env":["TOKENGO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.tokengo.com/v1","name":"TokenGo","doc":"https://www.tokengo.com/docs","models":{"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.89,"output":3.2647,"cache_read":0.2226},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"deepseek/deepseek-v3.1":{"id":"deepseek/deepseek-v3.1","name":"DeepSeek-V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.19,"output":0.71,"cache_read":0.06},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.098,"output":0.196,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":64000},"cost":{"input":0.2174,"output":0.326,"cache_read":0.06},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.435,"output":0.87},"canonical_model_id":"deepseek/deepseek-v4-pro"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":2.65,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"}}},"amd":{"id":"amd","env":["AMD_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://developer.amd.com.cn/radeon/api/v1","name":"AMD","doc":"https://developer.amd.com.cn/radeon/tokenfactory","models":{"MiniCPM-V-4.7-1B":{"id":"MiniCPM-V-4.7-1B","name":"MiniCPM-V 4.7 1B","description":"Compact MiniCPM-V model for single-image, multi-image and video understanding","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"openbmb/minicpm-v-4.7-1b"},"Qwen3.8-Flash-Next":{"id":"Qwen3.8-Flash-Next","name":"Qwen3.8 Flash Next","description":"Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"alibaba/qwen3.8-flash-next"},"DeepSeek-V4-Flash-Vision-Exp":{"id":"DeepSeek-V4-Flash-Vision-Exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"MiniCPM-V-4.7-35B-A3B":{"id":"MiniCPM-V-4.7-35B-A3B","name":"MiniCPM-V 4.7 35B-A3B","description":"First MoE MiniCPM-V model for single-image, multi-image and video understanding","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"openbmb/minicpm-v-4.7-35b-a3b"},"Qwen3.8-27B":{"id":"Qwen3.8-27B","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"alibaba/qwen3.8-27b"},"MiniCPM5-2B":{"id":"MiniCPM5-2B","name":"MiniCPM5-2B","description":"Dense 2B-class open-source model for on-device and resource-constrained use, with native long-context support, tool calling, and agentic tasks","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-09-06","last_updated":"2026-09-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"openbmb/minicpm5-2b"},"DeepSeek-V4-Flash":{"id":"DeepSeek-V4-Flash","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"DeepSeek-V4.1-Flash":{"id":"DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"Gemma-4-31B-IT":{"id":"Gemma-4-31B-IT","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"google/gemma-4-31b-it"},"MiMo-V2.6-Flash":{"id":"MiMo-V2.6-Flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"GLM-5.3-Flash":{"id":"GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.3-flash"}}},"infer":{"id":"infer","env":["INFER_API_KEY"],"npm":"@ai-sdk/openai","api":"https://infer.flow7.org/v1","name":"Infer by Flow7","doc":"https://infer.flow7.org/opencode","models":{"infer/gpt-5.6-sol:official":{"id":"infer/gpt-5.6-sol:official","name":"GPT-5.6 Sol (Official API)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":271999,"output":128000},"provider":{"shape":"responses"},"cost":{"input":2.5,"output":12.5,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"openai/gpt-5.6-sol"},"infer/gpt-6-astra:official":{"id":"infer/gpt-6-astra:official","name":"GPT-6 Astra (Official API)","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":271999,"output":128000},"provider":{"shape":"responses"},"cost":{"input":12.5,"output":62.5,"cache_read":1.25,"cache_write":15.625},"canonical_model_id":"openai/gpt-6-astra"}}},"ovhcloud":{"id":"ovhcloud","env":["OVHCLOUD_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://oai.endpoints.kepler.ai.cloud.ovh.net/v1","name":"OVHcloud AI Endpoints","doc":"https://www.ovhcloud.com/en/public-cloud/ai-endpoints/catalog//","models":{"qwen3guard-gen-0.6b":{"id":"qwen3guard-gen-0.6b","name":"Qwen3Guard-Gen-0.6B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-22","last_updated":"2026-01-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0,"output":0}},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"gpt-oss-20b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.18},"canonical_model_id":"openai/gpt-oss-20b"},"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8-27B","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.47,"output":3.19},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen3guard-gen-8b":{"id":"qwen3guard-gen-8b","name":"Qwen3Guard-Gen-8B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-22","last_updated":"2026-01-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0,"output":0}},"qwen2.5-vl-72b-instruct":{"id":"qwen2.5-vl-72b-instruct","name":"Qwen2.5-VL-72B-Instruct","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-01-26","last_updated":"2025-01-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":1.01,"output":1.01},"canonical_model_id":"alibaba/qwen2-5-vl-72b-instruct"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"gpt-oss-120b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.09,"output":0.47},"canonical_model_id":"openai/gpt-oss-120b"},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5-397B-A17B","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.71,"output":4.25},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"mistral-small-3.2-24b-instruct-2506":{"id":"mistral-small-3.2-24b-instruct-2506","name":"Mistral-Small-3.2-24B-Instruct-2506","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.1,"output":0.31},"canonical_model_id":"mistral/mistral-small-3.2-24b-instruct"},"mistral-nemo-instruct-2407":{"id":"mistral-nemo-instruct-2407","name":"Mistral-Nemo-Instruct-2407","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.14,"output":0.14},"canonical_model_id":"mistral/mistral-nemo"},"qwen3.5-9b":{"id":"qwen3.5-9b","name":"Qwen3.5-9B","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.12,"output":0.18},"canonical_model_id":"alibaba/qwen3.5-9b"},"meta-llama-3_3-70b-instruct":{"id":"meta-llama-3_3-70b-instruct","name":"Meta-Llama-3_3-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.74,"output":0.74},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"qwen3.6-27b":{"id":"qwen3.6-27b","name":"Qwen3.6-27B","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.47,"output":3.19},"canonical_model_id":"alibaba/qwen3.6-27b"}}},"empiriolabs":{"id":"empiriolabs","env":["EMPIRIOLABS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.empiriolabs.ai/v1","name":"EmpirioLabs AI","doc":"https://docs.empiriolabs.ai","models":{"seed-2-0-pro":{"id":"seed-2-0-pro","name":"Seed 2.0 Pro","description":"Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.63,"output":3.79,"cache_read":0.63,"tiers":[{"input":1.26,"output":7.58,"cache_read":1.26,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-pro"},"qwen3-8-flash":{"id":"qwen3-8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.16,"output":0.47,"cache_read":0.16},"canonical_model_id":"alibaba/qwen3.8-flash"},"seed-2-1-turbo":{"id":"seed-2-1-turbo","name":"Seed 2.1 Turbo","description":"Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"cost":{"input":0.63,"output":3.13,"cache_read":0.63},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"qwen3-5-plus":{"id":"qwen3-5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"status":"deprecated","cost":{"input":0.36,"output":2.21,"cache_read":0.36,"tiers":[{"input":1.08,"output":6.62,"cache_read":1.08,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.08,"output":6.62,"cache_read":1.08}},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen3-7-flash":{"id":"qwen3-7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":131072}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":65536},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"tiers":[{"input":0.1,"output":0.4,"cache_read":0.02,"tier":{"type":"context","size":32000}},{"input":0.2,"output":0.8,"cache_read":0.04,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"glm-4-6v-flash":{"id":"glm-4-6v-flash","name":"GLM 4.6V Flash","description":"Lightweight GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.6v-flash"},"gemma-3-27b":{"id":"gemma-3-27b","name":"Gemma 3 27B","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-3-27b-it"},"glm-4-7-flash":{"id":"glm-4-7-flash","name":"GLM 4.7 Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"qwen3-6-27b":{"id":"qwen3-6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":80000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.412564,"output":2.475384,"cache_read":0.412564},"canonical_model_id":"alibaba/qwen3.6-27b"},"minimax-m2-7":{"id":"minimax-m2-7","name":"MiniMax M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32768},"status":"deprecated","cost":{"input":0.15,"output":0.6,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax-m2-7-highspeed":{"id":"minimax-m2-7-highspeed","name":"MiniMax M2.7 Highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32768},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"seed-2-0-code":{"id":"seed-2-0-code","name":"Seed 2.0 Code","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.4,"output":2.4,"cache_read":0.4,"tiers":[{"input":0.8,"output":4.8,"cache_read":0.8,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"mistral-small-4":{"id":"mistral-small-4","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.15,"output":0.6,"cache_read":0.15},"canonical_model_id":"mistral/mistral-small-2603"},"step-3-7-flash":{"id":"step-3-7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":131072},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"qwen3-6-35b-a3b":{"id":"qwen3-6-35b-a3b","name":"Qwen3.6 35B A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.07,"output":0.42,"cache_read":0.035},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen3-7-max":{"id":"qwen3-7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":64000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":2.5,"output":7.5,"cache_read":2.5},"canonical_model_id":"alibaba/qwen3.7-max"},"seed-2-0-mini":{"id":"seed-2-0-mini","name":"Seed 2.0 Mini","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.12,"output":0.5,"cache_read":0.12,"tiers":[{"input":0.24,"output":1,"cache_read":0.24,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"fugu-ultra-v2-0":{"id":"fugu-ultra-v2-0","name":"Fugu Ultra v2.0","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-08-28","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra-v2"},"qwen3-6-plus":{"id":"qwen3-6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.5,"tiers":[{"input":2,"output":6,"cache_read":2,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":2}},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen3-7-plus":{"id":"qwen3-7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":256000}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.4,"output":1.6,"cache_read":0.4,"tiers":[{"input":1.2,"output":4.8,"cache_read":1.2,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":1.2}},"canonical_model_id":"alibaba/qwen3.7-plus"},"glm-4-5-flash":{"id":"glm-4-5-flash","name":"GLM 4.5 Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":98304},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.5-flash"},"qwen3-8-27b":{"id":"qwen3-8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.17,"output":0.5,"cache_read":0.08},"canonical_model_id":"alibaba/qwen3.8-27b"},"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":131072},"cost":{"input":1,"output":2.7,"cache_read":0.05},"canonical_model_id":"stepfun/step-5-preview"},"deepseek-v4-1-flash-1017":{"id":"deepseek-v4-1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":393216}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.3},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"muse-spark-1-2":{"id":"muse-spark-1-2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":1},"canonical_model_id":"meta/muse-spark-1.2"},"qwen3-6-flash":{"id":"qwen3-6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":64000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.25,"tiers":[{"input":1,"output":4,"cache_read":1,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1,"output":4,"cache_read":1}},"canonical_model_id":"alibaba/qwen3.6-flash"},"mimo-v2-6-flash":{"id":"mimo-v2-6-flash","name":"MiMo V2.6 Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.14},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":393216}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":1.32,"output":3.96,"cache_read":1.32},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"mimo-v2-6-pro":{"id":"mimo-v2-6-pro","name":"MiMo V2.6 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.435},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":393216}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":1.65,"output":3.3,"cache_read":1.65},"canonical_model_id":"deepseek/deepseek-v4-pro"},"kimi-k2-6":{"id":"kimi-k2-6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16000},"cost":{"input":0.8939,"output":3.7131,"cache_read":0.1788},"canonical_model_id":"moonshotai/kimi-k2.6"},"step-3-5-flash-2603":{"id":"step-3-5-flash-2603","name":"Step 3.5 Flash 2603","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":131072},"cost":{"input":0.1,"output":0.3,"cache_read":0.02},"canonical_model_id":"stepfun/step-3.5-flash-2603"},"kimi-k2-7-code":{"id":"kimi-k2-7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":131072},"cost":{"input":0.95,"output":4,"cache_read":0.95},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"muse-glimmer-30b":{"id":"muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.2,"output":0.8,"cache_read":0.05},"canonical_model_id":"meta/muse-glimmer-30b"},"qwen3-5-27b":{"id":"qwen3-5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":80000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.086,"output":0.688,"cache_read":0.086,"tiers":[{"input":0.258,"output":2.064,"cache_read":0.258,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen3-8-max-0902":{"id":"qwen3-8-max-0902","name":"Qwen3.8 Max 0902","description":"2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":2},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3,"output":15,"cache_read":3},"canonical_model_id":"moonshotai/kimi-k3"},"qwen3-5-35b-a3b":{"id":"qwen3-5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":80000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.057,"output":0.459,"cache_read":0.057,"tiers":[{"input":0.229,"output":1.835,"cache_read":0.229,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"glm-5-3-flash":{"id":"glm-5-3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.075,"output":0.25,"cache_read":0.075},"canonical_model_id":"zhipuai/glm-5.3-flash"},"qwen3-8-omni-flash":{"id":"qwen3-8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.3,"output":0.94,"cache_read":0.3},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"kimi-k2-7-code-highspeed":{"id":"kimi-k2-7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":131072},"cost":{"input":1.9,"output":8,"cache_read":1.9},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"deepseek-v4-1-flash":{"id":"deepseek-v4-1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":393216}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.3},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"fugu-ultra-v1-1":{"id":"fugu-ultra-v1-1","name":"Fugu Ultra v1.1","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra"},"qwen3-5-4b":{"id":"qwen3-5-4b","name":"Qwen3.5 4B","description":"Qwen3.5 4B is a low-cost multimodal reasoning model with 256K context, image and video input, function tools, and structured output.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1024,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-02","last_updated":"2026-03-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.04,"output":0.07,"cache_read":0.02},"canonical_model_id":"alibaba/qwen3.5-4b"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":524288},"cost":{"input":0.225,"output":0.9,"cache_read":0.045,"tiers":[{"input":0.45,"output":1.8,"cache_read":0.09,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":0.45,"output":1.8,"cache_read":0.09}},"canonical_model_id":"minimax/MiniMax-M3"},"gemma-4-26b-a4b":{"id":"gemma-4-26b-a4b","name":"Gemma 4 26B-A4B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.05,"output":0.29,"cache_read":0.025},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"glm-5-3":{"id":"glm-5-3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":1.4},"canonical_model_id":"zhipuai/glm-5.3"},"qwen3-8-max":{"id":"qwen3-8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":2},"canonical_model_id":"alibaba/qwen3.8-max"},"fugu-max":{"id":"fugu-max","name":"Fugu Max","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"sakana/fugu-max"},"qwen3-5-9b":{"id":"qwen3-5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1024,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.09,"output":0.13,"cache_read":0.045},"canonical_model_id":"alibaba/qwen3.5-9b"},"muse-spark-1-1":{"id":"muse-spark-1-1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":1},"canonical_model_id":"meta/muse-spark-1.1"},"glm-5-2":{"id":"glm-5-2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":1.4},"canonical_model_id":"zhipuai/glm-5.2"},"seed-2-0-lite":{"id":"seed-2-0-lite","name":"Seed 2.0 Lite","description":"Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.31,"output":2.5,"cache_read":0.31,"tiers":[{"input":0.62,"output":5,"cache_read":0.62,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"minimax-m3-1-flash":{"id":"minimax-m3-1-flash","name":"MiniMax M3.1 Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0.1125,"output":0.45,"cache_read":0.01125},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":393216}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.424,"output":1.272,"cache_read":0.424},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"qwen3-5-122b-a10b":{"id":"qwen3-5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":80000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.115,"output":0.917,"cache_read":0.115,"tiers":[{"input":0.287,"output":2.294,"cache_read":0.287,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen3-5-397b-a17b":{"id":"qwen3-5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":80000}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0.172,"output":1.032,"cache_read":0.172,"tiers":[{"input":0.43,"output":2.58,"cache_read":0.43,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"step-3-5-flash":{"id":"step-3-5-flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","family":"step","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":131072},"cost":{"input":0.1,"output":0.3,"cache_read":0.02},"canonical_model_id":"stepfun/step-3.5-flash"},"muse-spark-1-3":{"id":"muse-spark-1-3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":1},"canonical_model_id":"meta/muse-spark-1.3"},"fugu-ultra-v1-0":{"id":"fugu-ultra-v1-0","name":"Fugu Ultra v1.0","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":7.5,"output":45,"cache_read":1.5,"tiers":[{"input":15,"output":67.5,"cache_read":3,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":15,"output":67.5,"cache_read":3}},"canonical_model_id":"sakana/fugu-ultra"},"glm-5-1":{"id":"glm-5-1","name":"GLM 5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens","min":1,"max":38912}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202000,"output":128000},"cost":{"input":0.825,"output":3.301,"cache_read":0.165,"tiers":[{"input":1.1,"output":3.851,"cache_read":0.22,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5.1"},"mimo-v2-6-pro-ultraspeed":{"id":"mimo-v2-6-pro-ultraspeed","name":"MiMo V2.6 Pro UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":4.35,"output":8.7,"cache_read":4.35},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"qwen3-5-flash":{"id":"qwen3-5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.09,"output":0.368,"cache_read":0.09},"canonical_model_id":"alibaba/qwen3.5-flash"}}},"stepfun-ai-step-plan":{"id":"stepfun-ai-step-plan","env":["STEPFUN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.stepfun.ai/step_plan/v1","name":"StepFun Step Plan (Global)","doc":"https://platform.stepfun.ai/docs/en/step-plan/integrations/reasoning-api","models":{"step-3.7-flash":{"id":"step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"canonical_model_id":"stepfun/step-3.7-flash"},"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"canonical_model_id":"stepfun/step-5-preview"},"step-3.5-flash":{"id":"step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"canonical_model_id":"stepfun/step-3.5-flash"},"step-3.5-flash-2603":{"id":"step-3.5-flash-2603","name":"Step 3.5 Flash 2603","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"canonical_model_id":"stepfun/step-3.5-flash-2603"}}},"sambanova":{"id":"sambanova","env":["SAMBANOVA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.sambanova.ai/v1","name":"SambaNova","doc":"https://docs.sambanova.ai/cloud/docs/get-started/supported-models","models":{"gemma-4-31B-it":{"id":"gemma-4-31B-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.38,"output":1.15},"canonical_model_id":"google/gemma-4-31b-it"},"Meta-Llama-3.3-70B-Instruct":{"id":"Meta-Llama-3.3-70B-Instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":3072},"cost":{"input":0.6,"output":1.2},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"DeepSeek-V3.2":{"id":"DeepSeek-V3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":7168},"cost":{"input":3,"output":4.5},"canonical_model_id":"deepseek/deepseek-v3.2"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.22,"output":0.59},"canonical_model_id":"openai/gpt-oss-120b"},"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.6,"output":2.4,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"DeepSeek-V3.1":{"id":"DeepSeek-V3.1","name":"DeepSeek-V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":7168},"cost":{"input":3,"output":4.5},"canonical_model_id":"deepseek/deepseek-v3.1"}}},"frogbot":{"id":"frogbot","env":["FROGBOT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://app.frogbot.ai/api/v1","name":"FrogBot","doc":"https://docs.frogbot.ai","models":{"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.075},"canonical_model_id":"google/gemini-2.5-flash"},"minimax-m2-7":{"id":"minimax-m2-7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":192000,"output":8192},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.07,"output":0.2},"canonical_model_id":"openai/gpt-oss-20b"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"qwen-3-6-plus":{"id":"qwen-3-6-plus","name":"Qwen 3.6 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","cost":{"input":0.5,"output":3,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3.6-plus"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek v4 Pro","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro"},"zai-glm-5-1":{"id":"zai-glm-5-1","name":"Z.AI GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":8192},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"grok-4-1-fast-reasoning":{"id":"grok-4-1-fast-reasoning","name":"Grok 4.1 Fast (Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":128000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast-reasoning"},"kimi-k2-6":{"id":"kimi-k2-6","name":"Kimi-K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":128000},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"grok-code-fast-1":{"id":"grok-code-fast-1","name":"Grok 4.1 Fast (Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2023-10","release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.2,"output":1.5,"cache_read":0.02},"canonical_model_id":"xai/grok-code-fast-1"},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi-K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":128000},"status":"deprecated","cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"minimax-m2-5":{"id":"minimax-m2-5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":192000,"output":8192},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"gpt-5-4-mini":{"id":"gpt-5-4-mini","name":"GPT-5.4 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"gpt-5-4-nano":{"id":"gpt-5-4-nano","name":"GPT-5.4 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3-flash-preview"},"grok-4-1-fast-non-reasoning":{"id":"grok-4-1-fast-non-reasoning","name":"Grok 4.1 Fast (Non-Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":128000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast"},"grok-4-3":{"id":"grok-4-3","name":"Grok 4.3","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"gemini-3-1-pro-preview":{"id":"gemini-3-1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"google/gemini-3.1-pro-preview"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.31},"canonical_model_id":"google/gemini-2.5-pro"},"gpt-5-5":{"id":"gpt-5-5","name":"GPT-5.5","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"gpt-5-3-codex":{"id":"gpt-5-3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"}}},"tencent-token-plan":{"id":"tencent-token-plan","env":["TENCENT_TOKEN_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.lkeap.cloud.tencent.com/plan/v3","name":"Tencent Token Plan","doc":"https://cloud.tencent.com/document/product/1823/130060","models":{"hy4-preview":{"id":"hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"hy3":{"id":"hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"tencent/hy3"}}},"longcat":{"id":"longcat","env":["LONGCAT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.longcat.chat/openai","name":"LongCat","doc":"https://longcat.chat/platform/docs/","models":{"LongCat-2.0":{"id":"LongCat-2.0","name":"LongCat-2.0","description":"Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"meituan/longcat-2.0"},"LongCat-2.5-Preview":{"id":"LongCat-2.5-Preview","name":"LongCat-2.5-Preview","description":"Meituan's multimodal reasoning model for coding and agent tasks, with image understanding and a 1M-token context window","family":"longcat","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":262144},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"meituan/longcat-2.5-preview"}}},"huggingface":{"id":"huggingface","env":["HF_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"https://router.huggingface.co/v1","name":"Hugging Face","doc":"https://huggingface.co/docs/inference-providers","models":{"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3,"output":15},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K2.5":{"id":"moonshotai/Kimi-K2.5","name":"Kimi-K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/Kimi-K2-Thinking":{"id":"moonshotai/Kimi-K2-Thinking","name":"Kimi-K2-Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":2.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"moonshotai/Kimi-K2-Instruct-0905":{"id":"moonshotai/Kimi-K2-Instruct-0905","name":"Kimi-K2-Instruct-0905","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":1,"output":3},"canonical_model_id":"moonshotai/kimi-k2-0905"},"moonshotai/Kimi-K2-Instruct":{"id":"moonshotai/Kimi-K2-Instruct","name":"Kimi-K2-Instruct","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":1,"output":3},"canonical_model_id":"moonshotai/kimi-k2-instruct"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi-K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"MiniMaxAI/MiniMax-M2":{"id":"MiniMaxAI/MiniMax-M2","name":"MiniMax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":512000},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMaxAI/MiniMax-M2.7":{"id":"MiniMaxAI/MiniMax-M2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"MiniMaxAI/MiniMax-M1-80k":{"id":"MiniMaxAI/MiniMax-M1-80k","name":"MiniMax M1 80K","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":80000},"cost":{"input":0.55,"output":2.2},"canonical_model_id":"minimax/minimax-m1-80k"},"MiniMaxAI/MiniMax-M2.1":{"id":"MiniMaxAI/MiniMax-M2.1","name":"MiniMax-M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-10","release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.1"},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"thinkingmachines/Inkling":{"id":"thinkingmachines/Inkling","name":"Inkling","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1,"output":4.05},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/Inkling-Small":{"id":"thinkingmachines/Inkling-Small","name":"Inkling Small","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":1048576},"cost":{"input":0.5,"output":1.2},"canonical_model_id":"thinkingmachines/inkling-small"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.1,"output":0.5},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.25,"output":0.69},"canonical_model_id":"openai/gpt-oss-120b"},"XiaomiMiMo/MiMo-V2.5-Pro":{"id":"XiaomiMiMo/MiMo-V2.5-Pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1,"output":3},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"XiaomiMiMo/MiMo-V2-Flash":{"id":"XiaomiMiMo/MiMo-V2-Flash","name":"MiMo-V2-Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12-01","release_date":"2025-12-16","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":4096},"status":"deprecated","cost":{"input":0.1,"output":0.3},"canonical_model_id":"xiaomi/mimo-v2-flash"},"XiaomiMiMo/MiMo-V2.5":{"id":"XiaomiMiMo/MiMo-V2.5","name":"MiMo-V2.5","description":"MiMo model for long-context reasoning, perception, and agentic tasks","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.4,"output":2},"canonical_model_id":"xiaomi/mimo-v2.5"},"tencent/Hy4-preview":{"id":"tencent/Hy4-preview","name":"Hy4 preview","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":64000},"cost":{"input":0.834,"output":2.501},"canonical_model_id":"tencent/hy4-preview"},"tencent/Hy3":{"id":"tencent/Hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":128000},"cost":{"input":0.14,"output":0.58},"canonical_model_id":"tencent/hy3"},"meta-llama/Llama-3.3-70B-Instruct":{"id":"meta-llama/Llama-3.3-70B-Instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.59,"output":0.79},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/Llama-3.1-8B-Instruct":{"id":"meta-llama/Llama-3.1-8B-Instruct","name":"Llama-3.1-8B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.06,"output":0.06},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta-llama/Llama-Guard-4-12B":{"id":"meta-llama/Llama-Guard-4-12B","name":"Llama Guard 4 12B","description":"Multimodal Llama 4-based safety classifier for moderating prompts and responses","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.18,"output":0.18},"canonical_model_id":"meta/llama-guard-4-12b"},"zai-org/GLM-4.5":{"id":"zai-org/GLM-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"status":"deprecated","cost":{"input":0.6,"output":2.2},"canonical_model_id":"zhipuai/glm-4.5"},"zai-org/GLM-5":{"id":"zai-org/GLM-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"zai-org/GLM-4.6V-Flash":{"id":"zai-org/GLM-4.6V-Flash","name":"GLM-4.6V-Flash","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":0.9},"canonical_model_id":"zhipuai/glm-4.6v-flash"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":1.4,"output":4.4},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-4.7-Flash":{"id":"zai-org/GLM-4.7-Flash","name":"GLM-4.7-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"zai-org/AutoGLM-Phone-9B-Multilingual":{"id":"zai-org/AutoGLM-Phone-9B-Multilingual","name":"AutoGLM Phone 9B Multilingual","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-10","last_updated":"2025-12-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.035,"output":0.138},"canonical_model_id":"zhipuai/autoglm-phone-9b-multilingual"},"zai-org/GLM-4.5V":{"id":"zai-org/GLM-4.5V","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":0.6,"output":1.8},"canonical_model_id":"zhipuai/glm-4.5v"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-4.5-Air":{"id":"zai-org/GLM-4.5-Air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.13,"output":0.85},"canonical_model_id":"zhipuai/glm-4.5-air"},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-4.7":{"id":"zai-org/GLM-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"zai-org/GLM-4.6":{"id":"zai-org/GLM-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.55,"output":2.2},"canonical_model_id":"zhipuai/glm-4.6"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5},"canonical_model_id":"zhipuai/glm-5.3-flash"},"stepfun-ai/Step-3.5-Flash":{"id":"stepfun-ai/Step-3.5-Flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","family":"step","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"stepfun/step-3.5-flash"},"stepfun-ai/Step-3.7-Flash":{"id":"stepfun-ai/Step-3.7-Flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.2,"output":1.15},"canonical_model_id":"stepfun/step-3.7-flash"},"deepseek-ai/DeepSeek-V4.1-Flash-1017":{"id":"deepseek-ai/DeepSeek-V4.1-Flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-ai/DeepSeek-V4-Flash-Vision-Exp":{"id":"deepseek-ai/DeepSeek-V4-Flash-Vision-Exp","name":"DeepSeek V4 Flash Vision Exp","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.44,"output":1.32},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.14,"output":0.28},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-R1-0528":{"id":"deepseek-ai/DeepSeek-R1-0528","name":"DeepSeek-R1-0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":3,"output":5},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek-ai/DeepSeek-V3.2":{"id":"deepseek-ai/DeepSeek-V3.2","name":"DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":65536},"cost":{"input":0.28,"output":0.4},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek-ai/DeepSeek-R1":{"id":"deepseek-ai/DeepSeek-R1","name":"DeepSeek-R1","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":32768},"cost":{"input":0.7,"output":2.5},"canonical_model_id":"deepseek/deepseek-r1"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.435,"output":0.87,"cache_read":0.003625},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/DeepSeek-V3":{"id":"deepseek-ai/DeepSeek-V3","name":"DeepSeek-V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":8192},"cost":{"input":0.4,"output":1.3},"canonical_model_id":"deepseek/deepseek-v3"},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek-V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.27,"output":1},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek-ai/DeepSeek-V4-Flash":{"id":"deepseek-ai/DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.14,"output":0.28},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.96},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek-ai/DeepSeek-V4.1-Flash":{"id":"deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-ai/DeepSeek-V3-0324":{"id":"deepseek-ai/DeepSeek-V3-0324","name":"DeepSeek V3 0324","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.27,"output":1.12},"canonical_model_id":"deepseek/deepseek-v3-0324"},"Qwen/Qwen3.6-27B":{"id":"Qwen/Qwen3.6-27B","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.47,"output":3.19},"canonical_model_id":"alibaba/qwen3.6-27b"},"Qwen/Qwen3.6-35B-A3B":{"id":"Qwen/Qwen3.6-35B-A3B","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.15,"output":0.95},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen/Qwen3-Next-80B-A3B-Thinking":{"id":"Qwen/Qwen3-Next-80B-A3B-Thinking","name":"Qwen3-Next-80B-A3B-Thinking","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.3,"output":2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"Qwen/Qwen3.5-27B":{"id":"Qwen/Qwen3.5-27B","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":2.4},"canonical_model_id":"alibaba/qwen3.5-27b"},"Qwen/Qwen3-235B-A22B-Thinking-2507":{"id":"Qwen/Qwen3-235B-A22B-Thinking-2507","name":"Qwen3-235B-A22B-Thinking-2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":81920},"cost":{"input":0.3,"output":3},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"Qwen/Qwen3-Next-80B-A3B-Instruct":{"id":"Qwen/Qwen3-Next-80B-A3B-Instruct","name":"Qwen3-Next-80B-A3B-Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.25,"output":1},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"Qwen/Qwen3-235B-A22B":{"id":"Qwen/Qwen3-235B-A22B","name":"Qwen3 235B-A22B","description":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.2,"output":0.8},"canonical_model_id":"alibaba/qwen3-235b-a22b"},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B-A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.855,"output":2.565},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"Qwen/Qwen3-VL-30B-A3B-Instruct":{"id":"Qwen/Qwen3-VL-30B-A3B-Instruct","name":"Qwen3 VL 30B A3B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.2,"output":0.7},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"Qwen/Qwen3.5-35B-A3B":{"id":"Qwen/Qwen3.5-35B-A3B","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.4,"output":3},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3-VL-235B-A22B-Instruct":{"id":"Qwen/Qwen3-VL-235B-A22B-Instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":1.5},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"Qwen/Qwen3-Coder-480B-A35B-Instruct":{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct","name":"Qwen3-Coder-480B-A35B-Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":2,"output":2},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"Qwen/Qwen2.5-Coder-32B-Instruct":{"id":"Qwen/Qwen2.5-Coder-32B-Instruct","name":"Qwen2.5-Coder-32B-Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.06,"output":0.2},"canonical_model_id":"alibaba/qwen2.5-coder-32b-instruct"},"Qwen/Qwen3.5-397B-A17B":{"id":"Qwen/Qwen3.5-397B-A17B","name":"Qwen3.5-397B-A17B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"Qwen/Qwen3-Embedding-8B":{"id":"Qwen/Qwen3-Embedding-8B","type":"embedding","name":"Qwen 3 Embedding 8B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-12","release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":0},"cost":{"input":0.01,"output":0},"canonical_model_id":"alibaba/qwen3-embedding-8b"},"Qwen/Qwen3-Coder-30B-A3B-Instruct":{"id":"Qwen/Qwen3-Coder-30B-A3B-Instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.07,"output":0.26},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"Qwen/Qwen3-30B-A3B":{"id":"Qwen/Qwen3-30B-A3B","name":"Qwen3 30B A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.5},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"Qwen/Qwen3.5-122B-A10B":{"id":"Qwen/Qwen3.5-122B-A10B","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":3.2},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"Qwen/Qwen3-Embedding-4B":{"id":"Qwen/Qwen3-Embedding-4B","type":"embedding","name":"Qwen 3 Embedding 4B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-12","release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":0},"cost":{"input":0.01,"output":0},"canonical_model_id":"alibaba/qwen3-embedding-4b"},"Qwen/Qwen3.5-9B":{"id":"Qwen/Qwen3.5-9B","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.17,"output":0.25},"canonical_model_id":"alibaba/qwen3.5-9b"},"Qwen/Qwen3-Coder-Next":{"id":"Qwen/Qwen3-Coder-Next","name":"Qwen3-Coder-Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.2,"output":1.5},"canonical_model_id":"alibaba/qwen3-coder-next"},"Qwen/Qwen3-8B":{"id":"Qwen/Qwen3-8B","name":"Qwen3 8B","description":"Dense 8B Qwen3 model with switchable thinking for chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":32768},"cost":{"input":0.07,"output":0.18},"canonical_model_id":"alibaba/qwen3-8b"},"Qwen/Qwen3-VL-235B-A22B-Thinking":{"id":"Qwen/Qwen3-VL-235B-A22B-Thinking","name":"Qwen3 VL 235B A22B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.98,"output":3.95},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-thinking"},"Qwen/Qwen3.8-2.4T-A95B":{"id":"Qwen/Qwen3.8-2.4T-A95B","name":"Qwen3.8 2.4T A95B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":2.5,"output":6.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"Qwen/Qwen3-32B":{"id":"Qwen/Qwen3-32B","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.29,"output":0.59},"canonical_model_id":"alibaba/qwen3-32b"},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemma-4-26B-A4B-it":{"id":"google/gemma-4-26B-A4B-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.13,"output":0.4},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.08,"output":0.16},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemma-3-4b-it":{"id":"google/gemma-3-4b-it","name":"Gemma 3 4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.1},"canonical_model_id":"google/gemma-3-4b-it"},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.15},"canonical_model_id":"google/gemma-3-12b-it"}}},"vancine":{"id":"vancine","env":["VANCINE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://vancine.com/v1","name":"Vancine","doc":"https://vancine.com/docs","models":{"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"status":"deprecated","cost":{"input":0.24,"output":0.96,"cache_read":0.048},"canonical_model_id":"minimax/MiniMax-M3"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.6,"output":4.8,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.8-max"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":2.4,"output":12,"cache_read":0.24},"canonical_model_id":"moonshotai/kimi-k3"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.12,"output":0.38,"cache_read":0.013},"canonical_model_id":"alibaba/qwen3.8-flash"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.12,"output":0.4,"cache_read":0.024},"canonical_model_id":"zhipuai/glm-5.3-flash"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.24,"output":0.96,"cache_read":0.0048},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"hy4-preview":{"id":"hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":64000},"cost":{"input":0.67,"output":2,"cache_read":0.034},"canonical_model_id":"tencent/hy4-preview"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.12,"output":3.52,"cache_read":0.208},"canonical_model_id":"zhipuai/glm-5.3"}}},"nova":{"id":"nova","env":["NOVA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.nova.amazon.com/v1","name":"Nova","doc":"https://nova.amazon.com/dev/documentation","models":{"nova-2-lite-v1":{"id":"nova-2-lite-v1","name":"Nova 2 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"reasoning":0},"canonical_model_id":"amazon/nova-2-lite-v1"},"nova-2-pro-v1":{"id":"nova-2-pro-v1","name":"Nova 2 Pro","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-03","last_updated":"2026-01-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"reasoning":0}}}},"agnes":{"id":"agnes","env":["AGNES_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://apihub.agnes-ai.com/v1","name":"Agnes AI","doc":"https://agnes-ai.com/doc","models":{"agnes-2.5-pro-alpha":{"id":"agnes-2.5-pro-alpha","name":"Agnes 2.5 Pro Alpha","description":"Paid reasoning model for advanced coding, scientific reasoning, long-context analysis, agentic workflows, and multimodal understanding.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0.45,"output":0.9,"cache_read":0.0038}},"agnes-3.0-flash":{"id":"agnes-3.0-flash","name":"Agnes 3.0 Flash","description":"A new-generation Agnes AI text model with stronger Agnes Code task execution, tool orchestration, and trustworthy delivery.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09","last_updated":"2026-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":512000,"output":65536},"cost":{"input":0,"output":0}},"agnes-2.0-flash":{"id":"agnes-2.0-flash","name":"Agnes 2.0 Flash","description":"Fast and efficient model for agent workflows, tool calling, coding, and image understanding.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-25","last_updated":"2026-05-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":512000,"output":65536},"status":"deprecated","cost":{"input":0,"output":0}},"agnes-2.5-flash":{"id":"agnes-2.5-flash","name":"Agnes 2.5 Flash","description":"Upgraded model with improved coding, agent workflows, tool calling, and multimodal understanding.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07","last_updated":"2026-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":512000,"output":65536},"cost":{"input":0,"output":0}}}},"hetzner":{"id":"hetzner","env":["HETZNER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.hetzner.com/api/v1","name":"Hetzner","doc":"https://experiments.hetzner.com/docs/inference","models":{"Qwen3.8-27B":{"id":"Qwen3.8-27B","name":"Qwen3.8-27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"beta","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3.6-35B-A3B-FP8":{"id":"Qwen/Qwen3.6-35B-A3B-FP8","name":"Qwen3.6 35B A3B FP8","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"beta","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"}}},"moark":{"id":"moark","env":["MOARK_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://moark.com/v1","name":"Moark","doc":"https://moark.com/docs/openapi/v1#tag/%E6%96%87%E6%9C%AC%E7%94%9F%E6%88%90","models":{"DeepSeek-V4-Pro":{"id":"DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.8,"output":3.6},"canonical_model_id":"deepseek/deepseek-v4-pro"},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.29,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.7"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15},"canonical_model_id":"moonshotai/kimi-k3"},"DeepSeek-V4-Flash":{"id":"DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.5,"output":0.5},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.3},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"MiniMax-M2.1":{"id":"MiniMax-M2.1","name":"MiniMax-M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.29,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.1"},"Kimi-K2.6":{"id":"Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.98,"output":4},"canonical_model_id":"moonshotai/kimi-k2.6"},"GLM-5.3":{"id":"GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.2,"output":4.18},"canonical_model_id":"zhipuai/glm-5.3"},"GLM-4.7":{"id":"GLM-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.42,"output":2.2},"canonical_model_id":"zhipuai/glm-4.7"},"Qwen3.6-Flash":{"id":"Qwen3.6-Flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.18,"output":1.03},"canonical_model_id":"alibaba/qwen3.6-flash"}}},"prime":{"id":"prime","env":["PRIME_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.pinference.ai/api/v1","name":"Prime Inference","doc":"https://docs.primeintellect.ai/inference/overview","models":{"meta/muse-spark-open":{"id":"meta/muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25},"canonical_model_id":"meta/muse-spark-open"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03,"cache_write":0.15},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":1.54,"output":4.84},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":1.4},"canonical_model_id":"zhipuai/glm-5.3"},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 0905","description":"September 2025 Kimi K2 instruction update with a 256K context and improved agentic coding","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.6,"output":2.5},"canonical_model_id":"moonshotai/kimi-k2-0905"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":3.45,"output":17.25},"canonical_model_id":"moonshotai/kimi-k3"},"mistralai/mixtral-8x22b-instruct":{"id":"mistralai/mixtral-8x22b-instruct","name":"Mixtral 8x22B Instruct","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-01-31","release_date":"2024-04-17","last_updated":"2024-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":52428},"cost":{"input":2.2,"output":6.6},"canonical_model_id":"mistral/mixtral-8x22b-instruct"},"mistralai/mistral-small-2603":{"id":"mistralai/mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.165,"output":0.66},"canonical_model_id":"mistral/mistral-small-2603"},"mistralai/mistral-nemo":{"id":"mistralai/mistral-nemo","name":"Mistral Nemo","description":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.165,"output":0.17},"canonical_model_id":"mistral/mistral-nemo"},"mistralai/mistral-large-2512":{"id":"mistralai/mistral-large-2512","name":"Mistral Large 3 2512","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/mistral-large-2512"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.25,"output":10},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 Mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":10,"output":50},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"gpt-oss-20b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.07,"output":0.3},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2.5,"output":15},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.25,"output":2},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.75,"output":14},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.05,"output":0.4},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.75,"output":14},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":5,"output":30},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2.5,"output":15},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0.1,"output":0.5},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"gpt-oss-120b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.35,"output":0.75},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":5,"output":30},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0.2,"output":1.2},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-5.6-luna-pro":{"id":"openai/gpt-5.6-luna-pro","name":"GPT-5.6 Luna Pro","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":1,"output":6},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 Nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o-mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-5.6-sol-pro":{"id":"openai/gpt-5.6-sol-pro","name":"GPT-5.6 Sol Pro","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":5,"output":30},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":1.25,"output":10},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.75,"output":4.5},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.6-terra-pro":{"id":"openai/gpt-5.6-terra-pro","name":"GPT-5.6 Terra Pro","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2.5,"output":15},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.2,"output":1.25},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash 0423","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943717},"cost":{"input":0.44,"output":1.32},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.006,"cache_write":0.3},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro 0423","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.91,"output":3.83},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.3},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.44,"output":1.32},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50},"canonical_model_id":"anthropic/claude-fable-5-1"},"meta-llama/llama-3.3-70b-instruct":{"id":"meta-llama/llama-3.3-70b-instruct","name":"Llama 3.3 70B Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":1.04,"output":2.253},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"qwen/qwen3-vl-30b-a3b-instruct":{"id":"qwen/qwen3-vl-30b-a3b-instruct","name":"Qwen3 VL 30B A3B Instruct","description":"Open vision-language MoE model for image and video understanding, documents, and visual agents","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.2,"output":0.7},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0.99,"output":3.2},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3-coder-next":{"id":"qwen/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.3,"output":1.5},"canonical_model_id":"alibaba/qwen3-coder-next"},"qwen/qwen3-30b-a3b-instruct-2507":{"id":"qwen/qwen3-30b-a3b-instruct-2507","name":"Qwen3 30B A3B Instruct 2507","description":"Updated compact Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.13,"output":0.52},"canonical_model_id":"alibaba/qwen3-30b-a3b-instruct-2507"},"qwen/qwen3-vl-235b-a22b-instruct":{"id":"qwen/qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.3,"output":1.5},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["audio","image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":2,"output":10},"canonical_model_id":"google/gemini-4-argon"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.119,"output":0.3},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["audio","image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.35,"output":6.75},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3-flash-preview"},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6},"canonical_model_id":"xai/grok-4.6"},"x-ai/grok-4.7":{"id":"x-ai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":1.6,"output":4.8},"canonical_model_id":"xai/grok-4.7"}}},"trustedrouter":{"id":"trustedrouter","env":["TRUSTEDROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.trustedrouter.com/v1","name":"TrustedRouter","doc":"https://trustedrouter.com/docs","models":{"trustedrouter/zdr":{"id":"trustedrouter/zdr","name":"Zero Data Retention","description":"TrustedRouter privacy routing alias that prefers zero data retention model endpoints.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072}},"trustedrouter/cheap":{"id":"trustedrouter/cheap","name":"Cheap","description":"TrustedRouter low-cost routing alias that prefers inexpensive healthy model endpoints.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-01","last_updated":"2026-06-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072}},"trustedrouter/fast":{"id":"trustedrouter/fast","name":"Fast","description":"TrustedRouter speed routing alias that prefers low-latency healthy model endpoints.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072}},"trustedrouter/synth":{"id":"trustedrouter/synth","name":"Synth","description":"TrustedRouter synthesis orchestration alias that combines multiple model responses into one answer.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-20","last_updated":"2026-06-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072}},"trustedrouter/synth-code":{"id":"trustedrouter/synth-code","name":"Synth Code","description":"TrustedRouter code synthesis orchestration alias that combines multiple model responses into one answer.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-20","last_updated":"2026-06-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072}},"trustedrouter/e2e":{"id":"trustedrouter/e2e","name":"End-to-End Encrypted","description":"TrustedRouter privacy routing alias for end-to-end encrypted provider routes where available.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072}},"trustedrouter/auto":{"id":"trustedrouter/auto","name":"Auto","description":"TrustedRouter automatic routing alias that chooses a healthy supported model endpoint for the request.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-01","last_updated":"2026-06-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072}}}},"arcee":{"id":"arcee","env":["ARCEE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.arcee.ai/api/v1","name":"Arcee","doc":"https://docs.arcee.ai","models":{"trinity-large-thinking":{"id":"trinity-large-thinking","name":"Trinity Large Thinking","description":"Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use","family":"trinity","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-01","last_updated":"2026-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":80000},"status":"beta","cost":{"input":0.25,"output":0.8,"cache_read":0.06},"canonical_model_id":"arcee-ai/trinity-large-thinking"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"beta","cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.5,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"deepseek/deepseek-v4-flash-latest":{"id":"deepseek/deepseek-v4-flash-latest","name":"DeepSeek V4 Flash Latest","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"status":"beta","cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"beta","cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"status":"beta","cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":384000},"status":"deprecated","cost":{"input":1.74,"output":3.48,"cache_read":0.2},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"beta","cost":{"input":0.22,"output":0.66,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"zai-org/glm-5.3-flash":{"id":"zai-org/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"beta","cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai-org/glm-5.2":{"id":"zai-org/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"status":"beta","cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/glm-5.3":{"id":"zai-org/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"beta","cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"}}},"digitalocean":{"id":"digitalocean","env":["DIGITALOCEAN_ACCESS_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.do-ai.run/v1","name":"DigitalOcean","doc":"https://docs.digitalocean.com/products/gradient-ai-platform/details/models/","models":{"deepseek-3.2":{"id":"deepseek-3.2","name":"Deepseek 3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.5,"output":1.6,"cache_read":0.15},"canonical_model_id":"deepseek/deepseek-v3.2"},"gemma-4-31B-it":{"id":"gemma-4-31B-it","name":"Gemma 4","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.18,"output":0.5,"cache_read":0.036},"canonical_model_id":"google/gemma-4-31b-it"},"wan2-2-t2v-a14b":{"id":"wan2-2-t2v-a14b","type":"video","name":"Wan2.2-T2V-A14B","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-07-28","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["video"]},"open_weights":true,"limit":{"context":100,"output":1}},"anthropic-claude-opus-5":{"id":"anthropic-claude-opus-5","name":"Anthropic Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"e5-large-v2":{"id":"e5-large-v2","type":"embedding","name":"E5 Large v2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-05","last_updated":"2023-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":1024},"cost":{"input":0.02,"output":0},"canonical_model_id":"intfloat/e5-large-v2"},"openai-gpt-6-1-sol":{"id":"openai-gpt-6-1-sol","name":"OpenAI GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"tiers":[{"input":4,"output":15,"cache_read":0.2,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai-gpt-image-1.5":{"id":"openai-gpt-image-1.5","type":"image","name":"OpenAI GPT Image 1.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":0,"output":16384},"cost":{"input":5,"output":10,"cache_read":1},"canonical_model_id":"openai/gpt-image-1.5"},"anthropic-claude-4.6-sonnet":{"id":"anthropic-claude-4.6-sonnet","name":"Anthropic Claude Sonnet 4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"llama3-8b-instruct":{"id":"llama3-8b-instruct","name":"Llama 3.1 Instruct (8B)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.198,"output":0.198},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"anthropic-claude-4.1-opus":{"id":"anthropic-claude-4.1-opus","name":"Anthropic Claude 4.1 Opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"openai-gpt-5.2":{"id":"openai-gpt-5.2","name":"OpenAI GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"multi-qa-mpnet-base-dot-v1":{"id":"multi-qa-mpnet-base-dot-v1","type":"embedding","name":"Multi-QA-mpnet-base-dot-v1","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2021-08-23","last_updated":"2021-08-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":768},"cost":{"input":0.009,"output":0},"canonical_model_id":"sentence-transformers/multi-qa-mpnet-base-dot-v1"},"openai-o3-mini":{"id":"openai-o3-mini","name":"OpenAI o3 mini","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"qwen3-embedding-0.6b":{"id":"qwen3-embedding-0.6b","type":"embedding","name":"Qwen3 Embedding 0.6B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-03","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"output":1024},"status":"beta","cost":{"input":0.04,"output":0}},"openai-gpt-5":{"id":"openai-gpt-5","name":"OpenAI GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"mistral-7b-instruct-v0.3":{"id":"mistral-7b-instruct-v0.3","name":"Mistral 7B Instruct v0.3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-05-22","last_updated":"2024-05-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768}},"openai-gpt-4.1":{"id":"openai-gpt-4.1","name":"OpenAI GPT-4.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai-gpt-5.6-terra":{"id":"openai-gpt-5.6-terra","name":"OpenAI GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"openai/gpt-5.6-terra"},"anthropic-claude-3.5-sonnet":{"id":"anthropic-claude-3.5-sonnet","name":"Claude 3.5 Sonnet","description":"Legacy model retained for compatibility with older integrations","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-06-20","last_updated":"2024-10-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"anthropic-claude-sonnet-5.5":{"id":"anthropic-claude-sonnet-5.5","name":"Anthropic Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"openai-gpt-5.6-sol":{"id":"openai-gpt-5.6-sol","name":"OpenAI GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"tiers":[{"input":8,"output":30,"cache_read":0.8,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8}},"canonical_model_id":"openai/gpt-5.6-sol"},"ministral-3-8b-instruct-2512":{"id":"ministral-3-8b-instruct-2512","name":"Ministral 3 8B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"canonical_model_id":"mistral/ministral-3-8b-instruct-2512"},"openai-gpt-5.5":{"id":"openai-gpt-5.5","name":"OpenAI GPT-5.5","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo V2.5 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.8,"output":3,"cache_read":0.16},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"qwen-2.5-14b-instruct":{"id":"qwen-2.5-14b-instruct","name":"Qwen 2.5 14B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2024-09-19","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072}},"anthropic-claude-sonnet-4":{"id":"anthropic-claude-sonnet-4","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.3,"cache_write":3.75,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.3,"cache_write":3.75}},"canonical_model_id":"anthropic/claude-sonnet-4-0"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.6"},"openai-gpt-image-2":{"id":"openai-gpt-image-2","type":"image","name":"OpenAI GPT Image 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["image","text"]},"open_weights":false,"limit":{"context":0,"output":16384},"cost":{"input":8,"output":30},"canonical_model_id":"openai/gpt-image-2"},"nemotron-3-nano-30b":{"id":"nemotron-3-nano-30b","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"anthropic-claude-5-sonnet":{"id":"anthropic-claude-5-sonnet","name":"Anthropic Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"deepseek-v3":{"id":"deepseek-v3","name":"DeepSeek V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-12-26","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":131072}},"stable-diffusion-3.5-large":{"id":"stable-diffusion-3.5-large","type":"image","name":"Stable Diffusion 3.5 Large","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"stable-diffusion","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-10-22","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["image"]},"open_weights":true,"limit":{"context":256,"output":1}},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"openai-gpt-5.6-luna":{"id":"openai-gpt-5.6-luna","name":"OpenAI GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04}},"canonical_model_id":"openai/gpt-5.6-luna"},"alibaba-qwen3-32b":{"id":"alibaba-qwen3-32b","name":"Qwen3 32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.25,"output":0.55},"canonical_model_id":"alibaba/qwen3-32b"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"Deepseek V4 Pro","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.74,"output":3.48,"cache_read":0.348},"canonical_model_id":"deepseek/deepseek-v4-pro"},"openai-gpt-5.4-nano":{"id":"openai-gpt-5.4-nano","name":"OpenAI GPT-5.4 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"openai-gpt-5-mini":{"id":"openai-gpt-5-mini","name":"OpenAI GPT-5 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai-gpt-image-1":{"id":"openai-gpt-image-1","type":"image","name":"GPT Image 1","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-24","last_updated":"2025-04-24","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":40,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-1"},"anthropic-claude-3.5-haiku":{"id":"anthropic-claude-3.5-haiku","name":"Claude 3.5 Haiku","description":"Legacy model retained for compatibility with older integrations","family":"claude-haiku","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07-31","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"status":"deprecated","cost":{"input":0.8,"output":4,"cache_read":0.08,"cache_write":1},"canonical_model_id":"anthropic/claude-3-5-haiku-20241022"},"nemotron-3-nano-omni":{"id":"nemotron-3-nano-omni","name":"Nemotron 3 Nano Omni","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.5,"output":0.9,"cache_read":0.1},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"anthropic-claude-opus-5.5":{"id":"anthropic-claude-opus-5.5","name":"Anthropic Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.4,"cache_write":10,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.4,"cache_write":10}},"canonical_model_id":"anthropic/claude-opus-5-5"},"bge-m3":{"id":"bge-m3","type":"embedding","name":"BGE M3","description":"Multilingual embedding model combining dense, sparse, and multi-vector retrieval","family":"bge","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-01-30","last_updated":"2024-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":1024},"cost":{"input":0.02,"output":0},"canonical_model_id":"baai/bge-m3"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"status":"deprecated","cost":{"input":1.3,"output":4.3,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"openai-gpt-5-nano":{"id":"openai-gpt-5-nano","name":"OpenAI GPT-5 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai-gpt-5.4-mini":{"id":"openai-gpt-5.4-mini","name":"OpenAI GPT-5.4 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"deepseek-4-flash":{"id":"deepseek-4-flash","name":"Deepseek V4 Flash","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8-Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":262144},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.8-max"},"openai-gpt-oss-120b":{"id":"openai-gpt-oss-120b","name":"OpenAI GPT-oss-120b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.1,"output":0.7,"cache_read":0.02},"canonical_model_id":"openai/gpt-oss-120b"},"reflection-beam":{"id":"reflection-beam","name":"Beam","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-10-21","last_updated":"2026-10-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.4,"output":1.2,"cache_read":0.04}},"llama3.3-70b-instruct":{"id":"llama3.3-70b-instruct","name":"Llama 3.3 Instruct (70B)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.65,"output":0.65},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"gte-large-en-v1.5":{"id":"gte-large-en-v1.5","type":"embedding","name":"GTE Large (v1.5)","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-03-27","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":1024},"cost":{"input":0.09,"output":0}},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen 3.5 397B A17B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.55,"output":3.5,"cache_read":0.111},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"nemotron-3-ultra-550b":{"id":"nemotron-3-ultra-550b","name":"Nemotron 3 Ultra","description":"Flagship Nemotron model for high-throughput reasoning and complex agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.9,"output":1.7,"cache_read":0.18},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"anthropic-claude-fable-5":{"id":"anthropic-claude-fable-5","name":"Anthropic Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic-claude-3-opus":{"id":"anthropic-claude-3-opus","name":"Claude 3 Opus","description":"Legacy model retained for compatibility with older integrations","family":"claude-opus","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-08","release_date":"2024-02-29","last_updated":"2024-02-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":4096},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"nemotron-3.5-super-vl":{"id":"nemotron-3.5-super-vl","name":"NVIDIA Nemotron 3.5 Super VL","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.3,"output":0.65,"cache_read":0.06},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nemotron-nano-12b-v2-vl":{"id":"nemotron-nano-12b-v2-vl","name":"Nemotron-nano 12b v2-vl","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"temperature":true,"release_date":"2025-10-28","last_updated":"2025-10-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.2,"output":0.6,"cache_read":0.04},"canonical_model_id":"nvidia/nemotron-nano-12b-v2-vl"},"openai-gpt-5.2-pro":{"id":"openai-gpt-5.2-pro","name":"OpenAI GPT-5.2 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"openai-gpt-6-luna":{"id":"openai-gpt-6-luna","name":"OpenAI GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02}},"canonical_model_id":"openai/gpt-6-luna"},"openai-gpt-5.4-pro":{"id":"openai-gpt-5.4-pro","name":"OpenAI GPT-5.4 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"mistral-nemo-instruct-2407":{"id":"mistral-nemo-instruct-2407","name":"Mistral Nemo Instruct","description":"Legacy model retained for compatibility with older integrations","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":0.3,"output":0.3},"canonical_model_id":"mistral/mistral-nemo"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.5,"output":2.7,"cache_read":0.203},"canonical_model_id":"moonshotai/kimi-k2.5"},"all-mini-lm-l6-v2":{"id":"all-mini-lm-l6-v2","type":"embedding","name":"All-MiniLM-L6-v2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2021-08-30","last_updated":"2021-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256,"output":384},"cost":{"input":0.009,"output":0},"canonical_model_id":"sentence-transformers/all-minilm-l6-v2"},"anthropic-claude-fable-5.1":{"id":"anthropic-claude-fable-5.1","name":"Anthropic Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"openai-gpt-4o":{"id":"openai-gpt-4o","name":"OpenAI GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"nvidia-nemotron-3-super-120b":{"id":"nvidia-nemotron-3-super-120b","name":"NVIDIA Nemotron 3 Super 120B (Public Preview)","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"status":"deprecated","cost":{"input":0.3,"output":0.65,"cache_read":0.06},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"openai-gpt-oss-20b":{"id":"openai-gpt-oss-20b","name":"OpenAI GPT-oss-20b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.05,"output":0.45,"cache_read":0.01},"canonical_model_id":"openai/gpt-oss-20b"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"anthropic-claude-opus-4.7":{"id":"anthropic-claude-opus-4.7","name":"Anthropic Claude Opus 4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic-claude-opus-4":{"id":"anthropic-claude-opus-4","name":"Claude Opus 4","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-0"},"anthropic-claude-4.5-sonnet":{"id":"anthropic-claude-4.5-sonnet","name":"Anthropic Claude 4.5 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.21},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"anthropic-claude-haiku-4.5":{"id":"anthropic-claude-haiku-4.5","name":"Anthropic Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"openai-gpt-6-sol":{"id":"openai-gpt-6-sol","name":"OpenAI GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"tiers":[{"input":4,"output":15,"cache_read":0.4,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4}},"canonical_model_id":"openai/gpt-6-sol"},"deepseek-r1-distill-llama-70b":{"id":"deepseek-r1-distill-llama-70b","name":"DeepSeek R1 Distill Llama 70B","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-01-30","last_updated":"2025-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32678,"output":8192},"status":"deprecated","cost":{"input":0.99,"output":0.99}},"minimax-m2.5":{"id":"minimax-m2.5","name":"MiniMax M2.5 (Public Preview)","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax-m2.5","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08","release_date":"2026-02-12","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.3,"output":1.2,"cache_read":0.06}},"arcee-trinity-large-thinking":{"id":"arcee-trinity-large-thinking","name":"Arcee Trinity Large Thinking (Public Preview)","description":"Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use","family":"trinity","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.25,"output":0.9,"cache_read":0.06},"canonical_model_id":"arcee-ai/trinity-large-thinking"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"glm-5":{"id":"glm-5","name":"GLM 5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":64000},"status":"deprecated","cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"openai-gpt-5.4":{"id":"openai-gpt-5.4","name":"OpenAI GPT-5.4","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"anthropic-claude-opus-4.8":{"id":"anthropic-claude-opus-4.8","name":"Anthropic Claude Opus 4.8","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic-claude-3.7-sonnet":{"id":"anthropic-claude-3.7-sonnet","name":"Claude 3.7 Sonnet","description":"Legacy model retained for compatibility with older integrations","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-10-31","release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-3-7-sonnet-20250219"},"qwen3-tts-voicedesign":{"id":"qwen3-tts-voicedesign","type":"speech","name":"Qwen3 TTS VoiceDesign","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-30","modalities":{"input":["text"],"output":["audio"]},"open_weights":true,"limit":{"context":32768,"output":1}},"openai-gpt-5.3-codex":{"id":"openai-gpt-5.3-codex","name":"OpenAI GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"bge-reranker-v2-m3":{"id":"bge-reranker-v2-m3","type":"reranking","name":"BGE Reranker v2 M3","description":"Reranking model for improving retrieval quality in search and recommendation systems","family":"bge","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-03-15","last_updated":"2024-03-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":1},"cost":{"input":0.01,"output":0},"canonical_model_id":"baai/bge-reranker-v2-m3"},"openai-gpt-4o-mini":{"id":"openai-gpt-4o-mini","name":"OpenAI GPT-4o mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"glm-5.3":{"id":"glm-5.3","name":"GLM5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"qwen3-coder-flash":{"id":"qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.45,"output":1.7,"cache_read":0.09},"canonical_model_id":"alibaba/qwen3-coder-flash"},"mistral-3-14B":{"id":"mistral-3-14B","name":"Ministral 3 14B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"cost":{"input":0.2,"output":0.2,"cache_read":0.04},"canonical_model_id":"mistral/ministral-3-14b-instruct-2512"},"anthropic-claude-opus-4.5":{"id":"anthropic-claude-opus-4.5","name":"Anthropic Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic-claude-opus-4.6":{"id":"anthropic-claude-opus-4.6","name":"Anthropic Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"anthropic/claude-opus-4-6"},"openai-o1":{"id":"openai-o1","name":"OpenAI o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"openai-o3":{"id":"openai-o3","name":"OpenAI o3","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"anthropic-claude-5.5-haiku":{"id":"anthropic-claude-5.5-haiku","name":"Anthropic Claude 5.5 Haiku","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"anthropic/claude-haiku-5-5"},"openai-gpt-6-astra":{"id":"openai-gpt-6-astra","name":"OpenAI GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"tiers":[{"input":20,"output":75,"cache_read":2,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2}},"canonical_model_id":"openai/gpt-6-astra"},"llama-4-maverick":{"id":"llama-4-maverick","name":"Llama 4 Maverick","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.25,"output":0.87,"cache_read":0.05},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"openai-gpt-5.1-codex-max":{"id":"openai-gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex-max"},"fal-ai/fast-sdxl":{"id":"fal-ai/fast-sdxl","type":"image","name":"Fast SDXL","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"stable-diffusion","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-07-26","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["image"]},"open_weights":true,"limit":{"context":0,"output":0}},"fal-ai/elevenlabs/tts/multilingual-v2":{"id":"fal-ai/elevenlabs/tts/multilingual-v2","type":"speech","name":"ElevenLabs Multilingual TTS v2","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"elevenlabs","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-08-22","last_updated":"2023-08-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"canonical_model_id":"elevenlabs/eleven_multilingual_v2"},"fal-ai/stable-audio-25/text-to-audio":{"id":"fal-ai/stable-audio-25/text-to-audio","name":"Stable Audio 2.5 (Text-to-Audio)","description":"Speech generation model for controllable voice, narration, and audio delivery","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-08","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"fal-ai/flux/schnell":{"id":"fal-ai/flux/schnell","type":"image","name":"FLUX.1 [schnell]","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-08-01","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["image"]},"open_weights":true,"limit":{"context":0,"output":0}}}},"zhipuai-coding-plan":{"id":"zhipuai-coding-plan","env":["ZHIPU_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://open.bigmodel.cn/api/coding/paas/v4","name":"Zhipu AI Coding Plan","doc":"https://docs.bigmodel.cn/cn/coding-plan/overview","models":{"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"glm-4.6v":{"id":"glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6v"},"glm-5.3-highspeed":{"id":"glm-5.3-highspeed","name":"GLM-5.3 Highspeed","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"}}},"llmtech":{"id":"llmtech","env":["LLMTECH_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.llmtech.eu/v1","name":"LLM Tech","doc":"https://llmtech.eu/models/qwen3.8-27b","models":{"nvidia/Qwen3.8-27B-NVFP4":{"id":"nvidia/Qwen3.8-27B-NVFP4","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.25,"output":2.09,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.8-27b"}}},"opencode-go":{"id":"opencode-go","env":["OPENCODE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://opencode.ai/inference/go/openai/v1","name":"OpenCode Go","doc":"https://opencode.ai/v2/docs/console/go/","models":{"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen3.7-plus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/go/anthropic/v1"},"cost":{"input":0.4,"output":1.6,"cache_read":0.04,"cache_write":0.5,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.12,"cache_write":1.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.12,"cache_write":1.5}},"canonical_model_id":"alibaba/qwen3.7-plus"},"step-5-preview-free":{"id":"step-5-preview-free","name":"Step 5 Preview Free","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"stepfun/step-5-preview"},"gemini-4-argon-contributor":{"id":"gemini-4-argon-contributor","name":"Gemini 4 Argon Contributor","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/go/google/v1beta"},"cost":{"input":4,"output":20,"cache_read":0.2,"input_audio":4},"canonical_model_id":"google/gemini-4-argon"},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/go/anthropic/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo V2.5","description":"MiMo omni model for text, image, video, audio, and agents","family":"mimo-v2.5","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.5"},"minimax-m3.1-flash-free":{"id":"minimax-m3.1-flash-free","name":"MiniMax M3.1 Flash Free","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":512000},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/go/anthropic/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo V2.5 Pro","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo-v2.5-pro","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.435,"output":0.87,"cache_read":0.003625},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"muse-spark-1.3-contributor":{"id":"muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for coding and agentic workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.1,"output":0.2,"cache_read":0.002},"canonical_model_id":"meta/muse-spark-1.3"},"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":1,"output":2.7,"cache_read":0.1},"canonical_model_id":"stepfun/step-5-preview"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.435,"output":0.87,"cache_read":0.003625},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro (New)","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro"},"beam-free":{"id":"beam-free","name":"Beam Free","description":"Sparse mixture-of-experts reasoning model with 23B active parameters for coding and agentic workflows","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"space-bunny":{"id":"space-bunny","name":"Space Bunny","description":"Anonymous preview reasoning model for coding, agentic tasks, tool use, and multimodal input","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":524288,"output":524288},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.15,"output":0.6,"cache_read":0.03,"cache_write":0}},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter multimodal flagship for coding, professional work, and long-horizon agentic workflows","family":"qwen3.8-max","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/go/anthropic/v1"},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"longcat-2.0":{"id":"longcat-2.0","name":"LongCat-2.0","description":"Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"meituan/longcat-2.0"},"longcat-2.5-preview-free":{"id":"longcat-2.5-preview-free","name":"LongCat 2.5 Preview Free","description":"Meituan's multimodal reasoning model for coding and agent tasks, with image understanding and a 1M-token context window","family":"longcat","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"meituan/longcat-2.5-preview"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/go/anthropic/v1"},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax-m3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/go/anthropic/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"tiers":[{"input":0.6,"output":2.4,"cache_read":0.12,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":0.6,"output":2.4,"cache_read":0.12}},"canonical_model_id":"minimax/MiniMax-M3"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"muse-spark-1.2-contributor":{"id":"muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.1,"output":0.2,"cache_read":0.002},"canonical_model_id":"meta/muse-spark-1.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"hy4-preview":{"id":"hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":64000},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"hy3":{"id":"hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0.14,"output":0.58,"cache_read":0.035},"canonical_model_id":"tencent/hy3"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"cat-jam-free":{"id":"cat-jam-free","name":"Cat Jam Free","description":"A text-only reasoning model for coding and agentic workflows, with tool calling and structured JSON outputs.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/go/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}}}},"alibaba-token-plan":{"id":"alibaba-token-plan","env":["ALIBABA_TOKEN_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://token-plan.ap-southeast-1.maas.aliyuncs.com/compatible-mode/v1","name":"Alibaba Token Plan","doc":"https://www.alibabacloud.com/help/en/model-studio/token-plan-overview","models":{"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen-image-2.0-pro":{"id":"qwen-image-2.0-pro","type":"image","name":"Qwen Image 2.0 Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-03","last_updated":"2025-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"status":"deprecated","cost":{"input":0,"output":0}},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"qwen-image-2.0":{"id":"qwen-image-2.0","type":"image","name":"Qwen Image 2.0","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-pro"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"qwen3.8-max-preview":{"id":"qwen3.8-max-preview","name":"Qwen3.8 Max Preview","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-19","last_updated":"2026-07-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.8-max"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":131072}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-flash"},"happyhorse-1.1-r2v":{"id":"happyhorse-1.1-r2v","type":"video","name":"HappyHorse 1.1 Reference-to-Video","description":"Video model for reference-guided video generation","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-17","last_updated":"2026-07-17","modalities":{"input":["image","text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}},"happyhorse-1.1-i2v":{"id":"happyhorse-1.1-i2v","type":"video","name":"HappyHorse 1.1 Image-to-Video","description":"Video model for image-to-video generation","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-17","last_updated":"2026-07-17","modalities":{"input":["image","text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}},"wan2.7-image-pro":{"id":"wan2.7-image-pro","type":"image","name":"Wan2.7 Image Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.8-flash"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.5"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":131072}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-plus"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"glm-5":{"id":"glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":16384},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"wan2.7-image":{"id":"wan2.7-image","type":"image","name":"Wan2.7 Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"happyhorse-1.1-t2v":{"id":"happyhorse-1.1-t2v","type":"video","name":"HappyHorse 1.1 Text-to-Video","description":"Video model for prompt-driven text-to-video generation","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-17","last_updated":"2026-07-17","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"input":196601,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"minimax/MiniMax-M2.5"}}},"nan":{"id":"nan","env":["NAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.nan.builders/v1","name":"NaN","doc":"https://nan.builders/docs/models","models":{"qwen-image-2.1":{"id":"qwen-image-2.1","type":"image","name":"Qwen-Image-2.1","description":"Unified text-to-image generation and image editing model with native RGBA transparency and reference-image editing","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-14","last_updated":"2026-09-21","modalities":{"input":["text"],"output":["image"]},"open_weights":true,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen-image-2.1"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"gemma4":{"id":"gemma4","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"glm5.3":{"id":"glm5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen3.6":{"id":"qwen3.6","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"glm5.3-flash":{"id":"glm5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3-flash"}}},"synthetic":{"id":"synthetic","env":["SYNTHETIC_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.synthetic.new/openai/v1","name":"Synthetic","doc":"https://synthetic.new/pricing","models":{"hf:Qwen/Qwen3.6-27B":{"id":"hf:Qwen/Qwen3.6-27B","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.45,"output":3.6,"cache_read":0.45},"canonical_model_id":"alibaba/qwen3.6-27b"},"hf:zai-org/GLM-5.2":{"id":"hf:zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":1.4},"canonical_model_id":"zhipuai/glm-5.2"},"hf:zai-org/GLM-4.7-Flash":{"id":"hf:zai-org/GLM-4.7-Flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":65536},"cost":{"input":0.1,"output":0.5,"cache_read":0.02},"canonical_model_id":"zhipuai/glm-4.7-flash"},"hf:zai-org/GLM-5.3":{"id":"hf:zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"hf:zai-org/GLM-5.3-Flash":{"id":"hf:zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"cost":{"input":0.15,"output":0.5,"cache_read":0.04},"canonical_model_id":"zhipuai/glm-5.3-flash"},"hf:MiniMaxAI/MiniMax-M3":{"id":"hf:MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"status":"deprecated","cost":{"input":0.6,"output":1.2,"cache_read":0.6},"canonical_model_id":"minimax/MiniMax-M3"},"hf:moonshotai/Kimi-K3":{"id":"hf:moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"cost":{"input":3,"output":15,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"hf:moonshotai/Kimi-K2.7-Code":{"id":"hf:moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.95,"output":4,"cache_read":0.95},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"hf:nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4":{"id":"hf:nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":1,"cache_read":0.06},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"hf:deepseek-ai/DeepSeek-V4.1-Flash-1017":{"id":"hf:deepseek-ai/DeepSeek-V4.1-Flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"cost":{"input":0.6,"output":1.2,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"hf:deepseek-ai/DeepSeek-V4.1-Flash":{"id":"hf:deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"cost":{"input":0.6,"output":1.2,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"hf:openai/gpt-oss-120b":{"id":"hf:openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.1,"output":0.1,"cache_read":0.02},"canonical_model_id":"openai/gpt-oss-120b"}}},"oci":{"id":"oci","env":["OCI_GENAI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.generativeai.us-chicago-1.oci.oraclecloud.com/openai/v1","name":"OCI Generative AI","doc":"https://docs.oracle.com/en-us/iaas/Content/generative-ai/pretrained-models.htm","models":{"xai.grok-4.20-reasoning":{"id":"xai.grok-4.20-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"meta.llama-3.3-70b-instruct":{"id":"meta.llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"xai.grok-4.6":{"id":"xai.grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"openai.gpt-oss-20b":{"id":"openai.gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.07,"output":0.3},"canonical_model_id":"openai/gpt-oss-20b"},"openai.gpt-oss-120b":{"id":"openai.gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"xai.grok-4.20-non-reasoning":{"id":"xai.grok-4.20-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"meta.llama-4-maverick-17b-128e-instruct-fp8":{"id":"meta.llama-4-maverick-17b-128e-instruct-fp8","name":"Llama 4 Maverick 17B Instruct","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":16384},"cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"xai.grok-4.3":{"id":"xai.grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"meta.llama-4-scout-17b-16e-instruct":{"id":"meta.llama-4-scout-17b-16e-instruct","name":"Llama 4 Scout 17B Instruct","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":192000,"output":16384},"cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-4-scout-17b-instruct"}}},"minimax-cn-coding-plan":{"id":"minimax-cn-coding-plan","env":["MINIMAX_API_KEY"],"npm":"@ai-sdk/anthropic","api":"https://api.minimax.cn/anthropic/v1","name":"MiniMax Token Plan (minimax.cn)","doc":"https://platform.minimaxi.com/docs/token-plan/intro","models":{"MiniMax-M2":{"id":"MiniMax-M2","name":"MiniMax-M2","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0}},"MiniMax-M3.1-Flash-Preview":{"id":"MiniMax-M3.1-Flash-Preview","name":"MiniMax-M3.1-Flash-Preview","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-27","last_updated":"2026-09-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":512000},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash-Preview"},"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":524288},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.5-highspeed":{"id":"MiniMax-M2.5-highspeed","name":"MiniMax-M2.5-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M3.1-Flash":{"id":"MiniMax-M3.1-Flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"MiniMax-M2.1":{"id":"MiniMax-M2.1","name":"MiniMax-M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.7-highspeed":{"id":"MiniMax-M2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}}}},"zhipuai":{"id":"zhipuai","env":["ZHIPU_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://open.bigmodel.cn/api/paas/v4","name":"Zhipu AI","doc":"https://docs.bigmodel.cn/cn/guide/start/pricing","models":{"glm-4.5-air":{"id":"glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.11952,"output":0.89637,"cache_read":0.0239,"cache_write":0,"tiers":[{"input":0.17927,"output":1.19516,"cache_read":0.03585,"cache_write":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.5-air"},"glm-4.7-flashx":{"id":"glm-4.7-flashx","name":"GLM-4.7-FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.0747,"output":0.44819,"cache_read":0.01494,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"glm-4.6v-flash":{"id":"glm-4.6v-flash","name":"GLM-4.6V-Flash","description":"Lightweight GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6v-flash"},"glm-4.5v":{"id":"glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":16384},"cost":{"input":0.29879,"output":0.89637,"cache_read":0.05976,"cache_write":0,"tiers":[{"input":0.59758,"output":1.79274,"cache_read":0.11952,"cache_write":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.5v"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.89637,"output":3.58548,"cache_read":0.19421,"cache_write":0,"tiers":[{"input":1.19516,"output":4.18306,"cache_read":0.29879,"cache_write":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5.1"},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.44819,"output":2.09153,"cache_read":0.08964,"cache_write":0,"tiers":[{"input":0.59758,"output":2.39032,"cache_read":0.11952,"cache_write":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.7"},"glm-4.5":{"id":"glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.11952,"output":0.41831,"cache_read":0.03436,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.19516,"output":4.18306,"cache_read":0.29879,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"glm-4.7-flash":{"id":"glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"glm-4.5-flash":{"id":"glm-4.5-flash","name":"GLM-4.5-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":98304},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-flash"},"glm-5.3-flashx":{"id":"glm-5.3-flashx","name":"GLM-5.3-FlashX","description":"High-speed GLM-5.3-Flash serving option for coding and agent workflows","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.29879,"output":1.04577,"cache_read":0.08516,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5":{"id":"glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.59758,"output":2.68911,"cache_read":0.1494,"cache_write":0,"tiers":[{"input":0.89637,"output":3.28669,"cache_read":0.22409,"cache_write":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5"},"glm-4.6":{"id":"glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.19516,"output":4.18306,"cache_read":0.29879,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"glm-5v-turbo":{"id":"glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":0.74698,"output":3.28669,"cache_read":0.17927,"cache_write":0,"tiers":[{"input":1.04577,"output":3.88427,"cache_read":0.26891,"cache_write":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5v-turbo"},"glm-4.6v":{"id":"glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.1494,"output":0.44819,"cache_read":0.02988,"cache_write":0,"tiers":[{"input":0.29879,"output":0.89637,"cache_read":0.05976,"cache_write":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6v"}}},"tencent-coding-plan":{"id":"tencent-coding-plan","env":["TENCENT_CODING_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.lkeap.cloud.tencent.com/coding/v3","name":"Tencent Coding Plan (China)","doc":"https://cloud.tencent.com/document/product/1772/128947","models":{"hunyuan-t1":{"id":"hunyuan-t1","name":"Hunyuan-T1","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-03-08","last_updated":"2026-03-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"hunyuan-2.0-instruct":{"id":"hunyuan-2.0-instruct","name":"Tencent HY 2.0 Instruct","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-08","last_updated":"2026-03-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"hunyuan-turbos":{"id":"hunyuan-turbos","name":"Hunyuan-TurboS","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-08","last_updated":"2026-03-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi-K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.5"},"minimax-m2.5":{"id":"minimax-m2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":32768},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"minimax/MiniMax-M2.5"},"glm-5":{"id":"glm-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5"},"hunyuan-2.0-thinking":{"id":"hunyuan-2.0-thinking","name":"Tencent HY 2.0 Think","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-03-08","last_updated":"2026-03-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"tc-code-latest":{"id":"tc-code-latest","name":"Auto","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-08","last_updated":"2026-03-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}}}},"modelis":{"id":"modelis","env":["MODELIS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://modelishub.com/v1","name":"Modelis","doc":"https://modelishub.com/pricing","models":{"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5},"canonical_model_id":"google/gemini-2.5-flash"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.0983,"output":0.1966},"canonical_model_id":"deepseek/deepseek-v4-flash"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.435,"output":0.87},"canonical_model_id":"deepseek/deepseek-v4-pro"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50},"canonical_model_id":"anthropic/claude-fable-5"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10},"canonical_model_id":"google/gemini-2.5-pro"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":3,"output":9},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.768,"output":3.072},"canonical_model_id":"alibaba/qwen3.7-plus"}}},"sensenova":{"id":"sensenova","env":["SENSENOVA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://token.sensenova.cn/v1","name":"SenseNova (China)","doc":"https://platform.sensenova.cn/docs","models":{"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"sensenova-6.8-flash-lite":{"id":"sensenova-6.8-flash-lite","name":"SenseNova 6.8 Flash Lite","description":"SenseNova lightweight multimodal agent model for real-world complex tasks, data analysis, and complex information presentation","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-11","last_updated":"2026-08-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0,"cache_read":0}},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k3"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash is a high-performance conversational model by DeepSeek, supporting thinking/non-thinking modes, 1M context window, and tool calling.","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"}}},"inferx":{"id":"inferx","env":["INFERX_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://model.inferx.net/endpoints/v1","name":"InferX","doc":"https://model.inferx.net/endpoints","models":{"Agents-A1":{"id":"Agents-A1","name":"Agents-A1","description":"35B MoE agentic model built for long-horizon search, engineering, and scientific reasoning tasks","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"release_date":"2026-06-26","last_updated":"2026-06-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":100000},"cost":{"input":0,"output":0}},"Qwen3.6-35B-A3B-fp8-no-thinking":{"id":"Qwen3.6-35B-A3B-fp8-no-thinking","name":"Qwen3.6-35B-A3B-fp8-no-thinking","description":"Qwen3.6-35B-A3B-fp8 disable thinking","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"deepseek-v4-flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":100000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"Qwen3.6-27B-FP8":{"id":"Qwen3.6-27B-FP8","name":"Qwen3.6 27B FP8","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.6-27b"},"Qwen3.6-35B-A3B-FP8":{"id":"Qwen3.6-35B-A3B-FP8","name":"Qwen3.6 35B A3B FP8","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Devstral-2-123B-Instruct-2512-int4-AutoRound":{"id":"Devstral-2-123B-Instruct-2512-int4-AutoRound","name":"Devstral-2-123B-Instruct-2512-int4-AutoRound","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"mistral/devstral-2512"},"Qwen3-Embedding-8B":{"id":"Qwen3-Embedding-8B","type":"embedding","name":"Qwen3-Embedding-8B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-embedding-8b"},"Qwen3-Coder-Next-FP8":{"id":"Qwen3-Coder-Next-FP8","name":"Qwen3 Coder Next FP8","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-coder-next"},"Qwen3-Coder-Next-FP8-no-thinking":{"id":"Qwen3-Coder-Next-FP8-no-thinking","name":"Qwen3-Coder-Next-FP8-no-thinking","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":260000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-coder-next"},"Ornith-1.0-35B-FP8":{"id":"Ornith-1.0-35B-FP8","name":"Ornith-1.0-35B-FP8","description":"Large coding-reasoning model for agentic software tasks and RL search","family":"ornith","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-06-25","last_updated":"2026-06-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":100000},"cost":{"input":0,"output":0},"canonical_model_id":"deepreinforce/ornith-1.0-35b"},"mimo-v25":{"id":"mimo-v25","name":"mimo-v25","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":100000},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"gemma-4-31B-it-fp8":{"id":"gemma-4-31B-it-fp8","name":"Gemma 4 31B IT FP8","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-31b-it"}}},"abacus":{"id":"abacus","env":["ABACUS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://routellm.abacus.ai/v1","name":"Abacus","doc":"https://abacus.ai/help/api","models":{"route-llm":{"id":"route-llm","name":"RouteLLM","description":"RouteLLM routes prompts to an appropriate Abacus-backed text-generation model","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-01-01","last_updated":"2026-07-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"cost":{"input":2,"output":10,"cache_read":0.1}},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025},"canonical_model_id":"google/gemini-3.1-flash-lite"},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-2.5-flash"},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"grok-4-fast-non-reasoning":{"id":"grok-4-fast-non-reasoning","name":"Grok 4 Fast (Non-Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":16384},"status":"deprecated","cost":{"input":0.2,"output":0.5},"canonical_model_id":"xai/grok-4-fast-non-reasoning"},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2.5,"output":7.5},"canonical_model_id":"alibaba/qwen3.7-max"},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":32768},"cost":{"input":2,"output":6,"cache_read":0.3},"canonical_model_id":"xai/grok-4.5"},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":32768},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"qwen3-max":{"id":"qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":1.2,"output":6},"canonical_model_id":"alibaba/qwen3-max"},"o3-mini":{"id":"o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-7"},"gemini-3.1-flash-lite-preview":{"id":"gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":1},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"claude-opus-4-1-20250805":{"id":"claude-opus-4-1-20250805","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":15,"output":75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"gemini-3-pro-image-preview":{"id":"gemini-3-pro-image-preview","type":"image","name":"Nano Banana Pro Preview","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"status":"deprecated","cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"google/gemini-3-pro-image-preview"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gemini-2.5-flash-image":{"id":"gemini-2.5-flash-image","type":"image","name":"Nano Banana","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":30},"canonical_model_id":"google/gemini-2.5-flash-image"},"gpt-5-codex":{"id":"gpt-5-codex","name":"GPT-5-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5-codex"},"o3":{"id":"o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50},"canonical_model_id":"anthropic/claude-fable-5"},"claude-opus-4-20250514":{"id":"claude-opus-4-20250514","name":"Claude Opus 4","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":15,"output":75},"canonical_model_id":"anthropic/claude-opus-4-20250514"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4},"canonical_model_id":"openai/gpt-5.6-sol"},"gpt-5.3-chat-latest":{"id":"gpt-5.3-chat-latest","name":"GPT-5.3 Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"status":"deprecated","cost":{"input":1.75,"output":14},"canonical_model_id":"openai/gpt-5.3-chat-latest"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.6-luna"},"gemini-3.1-flash-image-preview":{"id":"gemini-3.1-flash-image-preview","type":"image","name":"Nano Banana 2 Preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":1048576,"output":32768},"status":"deprecated","cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image-preview"},"claude-opus-4-5-20251101":{"id":"claude-opus-4-5-20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-5"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"google/gemini-3.1-pro-preview"},"grok-code-fast-1":{"id":"grok-code-fast-1","name":"Grok Code Fast 1","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":16384},"status":"deprecated","cost":{"input":0.2,"output":1.5},"canonical_model_id":"xai/grok-code-fast-1"},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.6-flash"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-6"},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"gpt-4o-mini":{"id":"gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-4o-mini"},"muse-spark-1.1":{"id":"muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"muse-spark-1.2":{"id":"muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"grok-4-0709":{"id":"grok-4-0709","name":"Grok 4","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":16384},"status":"deprecated","cost":{"input":3,"output":15}},"gpt-4o-2024-11-20":{"id":"gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":2.5,"output":10},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.6,"output":3},"canonical_model_id":"moonshotai/kimi-k2.5"},"claude-sonnet-4-5-20250929":{"id":"claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4},"canonical_model_id":"openai/o4-mini"},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"google/gemini-3.5-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2,"output":10},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-5.3-codex-xhigh":{"id":"gpt-5.3-codex-xhigh","name":"GPT-5.3 Codex XHigh","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":1.75,"output":14},"canonical_model_id":"openai/gpt-5.3-codex"},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"kimi-k2-turbo-preview":{"id":"kimi-k2-turbo-preview","name":"Kimi K2 Turbo Preview","description":"Fast Kimi model for responsive chat, coding help, and agent loops","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-08","last_updated":"2025-07-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":8192},"status":"deprecated","cost":{"input":0.15,"output":8}},"qwen-2.5-coder-32b":{"id":"qwen-2.5-coder-32b","name":"Qwen 2.5 Coder 32B","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0.79,"output":0.79},"canonical_model_id":"alibaba/qwen2.5-coder-32b-instruct"},"gpt-5.1-codex-max":{"id":"gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":1.25,"output":10},"canonical_model_id":"openai/gpt-5.1-codex-max"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"o3-pro":{"id":"o3-pro","name":"o3-pro","description":"High-effort o3 tier for difficult technical reasoning and careful answers","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":20,"output":80},"canonical_model_id":"openai/o3-pro"},"llama-3.3-70b-versatile":{"id":"llama-3.3-70b-versatile","name":"Llama 3.3 70B Versatile","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"status":"deprecated","cost":{"input":0.59,"output":0.79},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"gemini-3.1-flash-image":{"id":"gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":1048576,"output":32768},"status":"deprecated","cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3-flash-preview"},"grok-4-1-fast-non-reasoning":{"id":"grok-4-1-fast-non-reasoning","name":"Grok 4.1 Fast (Non-Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-11-17","last_updated":"2025-11-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":16384},"status":"deprecated","cost":{"input":0.2,"output":0.5}},"mimo-v2-pro":{"id":"mimo-v2-pro","name":"MiMo-V2-Pro","description":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1,"output":3,"cache_read":0.2},"canonical_model_id":"xiaomi/mimo-v2-pro"},"claude-sonnet-4-20250514":{"id":"claude-sonnet-4-20250514","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"gpt-5.2-chat-latest":{"id":"gpt-5.2-chat-latest","name":"GPT-5.2 Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"status":"deprecated","cost":{"input":1.75,"output":14},"canonical_model_id":"openai/gpt-5.2-chat-latest"},"gemini-3-pro-image":{"id":"gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"google/gemini-3-pro-image"},"gpt-5.1-chat-latest":{"id":"gpt-5.1-chat-latest","name":"GPT-5.1 Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"status":"deprecated","cost":{"input":1.25,"output":10},"canonical_model_id":"openai/gpt-5.1-chat-latest"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"google/gemini-2.5-pro"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"claude-3-7-sonnet-20250219":{"id":"claude-3-7-sonnet-20250219","name":"Claude Sonnet 3.7","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-10-31","release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15},"canonical_model_id":"anthropic/claude-3-7-sonnet-20250219"},"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMaxAI/MiniMax-M2.7":{"id":"MiniMaxAI/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.7"},"thinkingmachines/Inkling":{"id":"thinkingmachines/Inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":3.74,"output":9.36,"cache_read":0.748},"canonical_model_id":"thinkingmachines/inkling"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.037,"output":0.17},"canonical_model_id":"openai/gpt-oss-120b"},"deepseek/deepseek-v3.1":{"id":"deepseek/deepseek-v3.1","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0.55,"output":1.66},"canonical_model_id":"deepseek/deepseek-v3.1"},"meta-llama/Meta-Llama-3.3-70B-Instruct":{"id":"meta-llama/Meta-Llama-3.3-70B-Instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.59,"output":0.79},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8":{"id":"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8","name":"Llama 4 Maverick 17B Instruct","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":8192},"status":"deprecated","cost":{"input":0.14,"output":0.59},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo":{"id":"meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo","name":"Llama 3.1 405B Instruct Turbo","description":"Compact Llama instruction model for fast chat and local deployment","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":3.5,"output":3.5}},"meta-llama/Meta-Llama-3.1-8B-Instruct":{"id":"meta-llama/Meta-Llama-3.1-8B-Instruct","name":"Llama 3.1 8B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.02,"output":0.05},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"zai-org/GLM-4.5":{"id":"zai-org/GLM-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":96000},"status":"deprecated","cost":{"input":0.6,"output":2.2},"canonical_model_id":"zhipuai/glm-4.5"},"zai-org/GLM-5":{"id":"zai-org/GLM-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1,"output":3.2},"canonical_model_id":"zhipuai/glm-5"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1.38,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-4.7":{"id":"zai-org/GLM-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2},"canonical_model_id":"zhipuai/glm-4.7"},"zai-org/GLM-4.6":{"id":"zai-org/GLM-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.55,"output":2.2},"canonical_model_id":"zhipuai/glm-4.6"},"deepseek-ai/DeepSeek-V3.1-Terminus":{"id":"deepseek-ai/DeepSeek-V3.1-Terminus","name":"DeepSeek V3.1 Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-06-01","last_updated":"2025-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0.27,"output":1}},"deepseek-ai/DeepSeek-V3.2":{"id":"deepseek-ai/DeepSeek-V3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0.27,"output":0.4},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek-ai/DeepSeek-R1":{"id":"deepseek-ai/DeepSeek-R1","name":"DeepSeek R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":3,"output":7},"canonical_model_id":"deepseek/deepseek-r1"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"status":"deprecated","cost":{"input":1.74,"output":3.48,"cache_read":0.15},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/DeepSeek-V4-Flash":{"id":"deepseek-ai/DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4-flash"},"Qwen/QwQ-32B":{"id":"Qwen/QwQ-32B","name":"QwQ 32B","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-03-05","last_updated":"2025-03-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"status":"deprecated","cost":{"input":0.4,"output":0.4},"canonical_model_id":"alibaba/qwq-32b"},"Qwen/Qwen3.6-27B":{"id":"Qwen/Qwen3.6-27B","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.32,"output":3.2},"canonical_model_id":"alibaba/qwen3.6-27b"},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B A22B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"status":"deprecated","cost":{"input":0.13,"output":0.6},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"Qwen/Qwen3-Coder-480B-A35B-Instruct":{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct","name":"Qwen3-Coder 480B-A35B Instruct","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.38,"output":1.55},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"Qwen/Qwen2.5-72B-Instruct":{"id":"Qwen/Qwen2.5-72B-Instruct","name":"Qwen 2.5 72B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-09-19","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0.11,"output":0.38}},"Qwen/Qwen3-32B":{"id":"Qwen/Qwen3-32B","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.08,"output":0.28},"canonical_model_id":"alibaba/qwen3-32b"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"}}},"boundless":{"id":"boundless","env":["BOUNDLESS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.inference.boundless.network/v1","name":"Boundless","doc":"https://inference.boundless.network/models","models":{"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2},"canonical_model_id":"openai/gpt-6-sol"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.04,"output":2.08,"cache_read":0.0352},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01},"canonical_model_id":"openai/gpt-6-luna"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":128000},"status":"deprecated","cost":{"input":2,"output":10,"cache_read":0.2},"canonical_model_id":"openai/gpt-5.6-sol"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":2.3,"output":11.4,"cache_read":0.23},"canonical_model_id":"moonshotai/kimi-k3"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":128000},"status":"deprecated","cost":{"input":0.2,"output":1.2,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.6-luna"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":128000},"status":"deprecated","cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-5"},"claude-opus-5.5":{"id":"claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"anthropic/claude-opus-5-5"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-flash"},"dsv4":{"id":"dsv4","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.04,"output":0.21,"cache_read":0.02},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.12,"output":0.4,"cache_read":0.024},"canonical_model_id":"zhipuai/glm-5.3-flash"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.28,"output":1.1,"cache_read":0.056},"canonical_model_id":"minimax/MiniMax-M3"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.34,"output":2.3,"cache_read":0.17},"canonical_model_id":"zhipuai/glm-5.2"},"qwen3.6":{"id":"qwen3.6","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.04,"output":0.35,"cache_read":0.02},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.2,"output":1,"cache_read":0.01},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"claude-fable-5.1":{"id":"claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25},"canonical_model_id":"anthropic/claude-fable-5-1"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.12,"output":3.52,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.3"},"nemotron3-super":{"id":"nemotron3-super","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.12,"output":0.52,"cache_read":0.06},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"}}},"bee":{"id":"bee","env":["BEE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.bee.heossi.com/bee","name":"Bee by HEOSSI","doc":"https://bee.heossi.com/docs/sdks","models":{"bee-brood":{"id":"bee-brood","name":"Bee Brood 1.0","description":"Production multimodal reasoning for image and video analysis, tool-enabled workflows, and multi-step planning at controlled cost.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-10","last_updated":"2026-09-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":16384},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.72}},"bee-cell":{"id":"bee-cell","name":"Bee Cell 1.0","description":"Fast hosted AI for chat, coding help, document question answering, image understanding, and lightweight technical reasoning.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-10","last_updated":"2026-09-28","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.36}},"bee-buzz":{"id":"bee-buzz","name":"Bee Buzz 1.0","description":"Mid-tier production model for tool-using agents, multi-step workflows, and builder/integration work. Bigger context than Comb, deeper planning, faster than Hive on routine tasks.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-10","last_updated":"2026-09-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":24576},"cost":{"input":2.5,"output":10,"cache_read":0.25,"cache_write":3}},"bee-hive":{"id":"bee-hive","name":"Bee Hive 1.0","description":"Scalable Bee serving with quantum-reasoning routing and multi-base domain-adapter routing for shared knowledge workflows, document intelligence, and lower-cost cloud inference. Domain adapters activate per customer domain after evaluation.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-10","last_updated":"2026-09-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":5,"output":20,"cache_read":0.5,"cache_write":6}},"bee-swarm":{"id":"bee-swarm","name":"Bee Swarm 1.0","description":"Premium reasoning, multi-model cloud execution, large-document analysis, and high-value technical work. Includes the evolution engine and stronger eval gates by default.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":false,"structured_output":false,"release_date":"2026-07-10","last_updated":"2026-09-28","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":25,"output":150,"cache_read":2.5,"cache_write":30}},"bee-comb":{"id":"bee-comb","name":"Bee Comb 2.0","description":"Stronger local/workstation AI for coding, architecture, security reasoning, larger RAG, and domain-specific workflows. The default workhorse for engineers and dev teams.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-10","last_updated":"2026-09-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":16384},"cost":{"input":1.25,"output":5,"cache_read":0.125,"cache_write":1.5}}}},"crusoe":{"id":"crusoe","env":["CRUSOE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.inference.crusoecloud.com/v1","name":"Crusoe","doc":"https://docs.crusoecloud.com/managed-inference/overview","models":{"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.7,"output":3.5,"cache_read":0.35},"canonical_model_id":"moonshotai/kimi-k2.6"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.05,"output":0.25,"cache_read":0.05},"canonical_model_id":"openai/gpt-oss-120b"},"meta-llama/Llama-3.3-70B-Instruct":{"id":"meta-llama/Llama-3.3-70B-Instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.25,"output":0.75,"cache_read":0.13},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"nvidia/NVIDIA-Nemotron-3-Super-120B-A12B":{"id":"nvidia/NVIDIA-Nemotron-3-Super-120B-A12B","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":2.4,"cache_read":0.15},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B":{"id":"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.05,"output":0.2,"cache_read":0.03},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/Nemotron-3-Nano-Omni-Reasoning-30B-A3B":{"id":"nvidia/Nemotron-3-Nano-Omni-Reasoning-30B-A3B","name":"Nemotron 3 Nano Omni 30B A3B Reasoning","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.3,"output":1.83,"cache_read":0.3,"input_audio":0.5},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"deepseek-ai/DeepSeek-V3-0324":{"id":"deepseek-ai/DeepSeek-V3-0324","name":"DeepSeek V3 0324","description":"March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"status":"deprecated","cost":{"input":0.5,"output":1.5,"cache_read":0.25},"canonical_model_id":"deepseek/deepseek-v3-0324"},"zai/GLM-5.2":{"id":"zai/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai/GLM-5.1":{"id":"zai/GLM-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"status":"deprecated","cost":{"input":1.2,"output":4.4,"cache_read":0.25},"canonical_model_id":"zhipuai/glm-5.1"},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B-A22B Instruct 2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"status":"deprecated","cost":{"input":0.22,"output":0.8,"cache_read":0.11},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.14,"output":0.4,"cache_read":0.14},"canonical_model_id":"google/gemma-4-31b-it"}}},"routing-run":{"id":"routing-run","env":["ROUTING_RUN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.routing.run/v1","name":"routing.run","doc":"https://docs.routing.run/api-reference/models","models":{"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":5,"output":25},"canonical_model_id":"anthropic/claude-opus-4-8"},"glm-5.2-nitro":{"id":"glm-5.2-nitro","name":"GLM 5.2 Nitro","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"cost":{"input":0.8,"output":2.4},"canonical_model_id":"zhipuai/glm-5.2"},"nemotron-3-ultra":{"id":"nemotron-3-ultra","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32000},"cost":{"input":0.1,"output":0.1},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":1.5,"output":9},"canonical_model_id":"openai/gpt-5.6-terra"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":64000},"cost":{"input":0.112,"output":0.224},"canonical_model_id":"deepseek/deepseek-v4-flash"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"cost":{"input":0.275,"output":1.1},"canonical_model_id":"moonshotai/kimi-k2.6"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"kimi-k2.7-code-nitro":{"id":"kimi-k2.7-code-nitro","name":"Kimi K2.7 Code Nitro","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"cost":{"input":0.275,"output":1.1},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":64000},"cost":{"input":0.348,"output":0.696},"canonical_model_id":"deepseek/deepseek-v4-pro"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15},"canonical_model_id":"openai/gpt-5.6-sol"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"cost":{"input":0.275,"output":1.1},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":0.7,"output":4.2},"canonical_model_id":"openai/gpt-5.6-luna"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"cost":{"input":0.8,"output":2.4},"canonical_model_id":"zhipuai/glm-5.2"},"qwen3.5-9b":{"id":"qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32000},"cost":{"input":0.16,"output":0.48},"canonical_model_id":"alibaba/qwen3.5-9b"},"kimi-k2.6-nitro":{"id":"kimi-k2.6-nitro","name":"Kimi K2.6 Nitro","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"cost":{"input":0.275,"output":1.1},"canonical_model_id":"moonshotai/kimi-k2.6"}}},"zenifra":{"id":"zenifra","env":["ZENIFRA_AI_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://ai.zenifra.com/v1","name":"Zenifra","doc":"https://docs.zenifra.com","models":{"alibaba/qwen3.6-35b-a3b":{"id":"alibaba/qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"provider":{"shape":"completions"},"cost":{"input":0.19,"output":0.48},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"}}},"xiaomi":{"id":"xiaomi","env":["XIAOMI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.xiaomimimo.com/v1","name":"Xiaomi","doc":"https://platform.xiaomimimo.com/#/docs","models":{"mimo-v2.5-pro-ultraspeed":{"id":"mimo-v2.5-pro-ultraspeed","name":"MiMo-V2.5-Pro-UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-06-08","last_updated":"2026-06-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":1.305,"output":2.61,"cache_read":0.0108},"canonical_model_id":"xiaomi/mimo-v2.5-pro-ultraspeed"},"mimo-v2.6-pro-ultraspeed":{"id":"mimo-v2.6-pro-ultraspeed","name":"MiMo-V2.6-Pro-UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.35,"output":8.7,"cache_read":0.036},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.5"},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"MiMo Pro model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"mimo-v2-flash":{"id":"mimo-v2-flash","name":"MiMo-V2-Flash","description":"Legacy model retained for compatibility with older integrations","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12-01","release_date":"2025-12-16","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2-flash"},"mimo-v2-omni":{"id":"mimo-v2-omni","name":"MiMo-V2-Omni","description":"Legacy model retained for compatibility with older integrations","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2-omni"},"mimo-v2-pro":{"id":"mimo-v2-pro","name":"MiMo-V2-Pro","description":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2-pro"}}},"tokenrouter":{"id":"tokenrouter","env":["TOKENROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.tokenrouter.com/v1","name":"TokenRouter","doc":"https://www.tokenrouter.com/docs/tokenrouter-feature-guide/","models":{"z-ai/glm-5.3-free":{"id":"z-ai/glm-5.3-free","name":"GLM-5.3 (free)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3"}}},"ollama-cloud":{"id":"ollama-cloud","env":["OLLAMA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://ollama.com/v1","name":"Ollama Cloud","doc":"https://docs.ollama.com/cloud","models":{"nemotron-3-ultra":{"id":"nemotron-3-ultra","name":"nemotron-3-ultra","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"cost":{"input":0.1,"output":3,"cache_read":0.1},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"minimax-m2.7":{"id":"minimax-m2.7","name":"minimax-m2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":196608},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"deepseek-v4-flash","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"status":"deprecated","cost":{"input":0.22,"output":0.66,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4-flash"},"kimi-k2.6":{"id":"kimi-k2.6","name":"kimi-k2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"mistral-large-4":{"id":"mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":262144},"cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"deepseek-v4-pro:0813":{"id":"deepseek-v4-pro:0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"mistral-large-3:675b":{"id":"mistral-large-3:675b","name":"mistral-large-3:675b","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/mistral-large-2512"},"gemma4:31b":{"id":"gemma4:31b","name":"gemma4:31b","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.14,"output":0.4,"cache_read":0.05},"canonical_model_id":"google/gemma-4-31b-it"},"nemotron-3-super":{"id":"nemotron-3-super","name":"nemotron-3-super","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.015,"output":0.6,"cache_read":0.015},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"deepseek-v4-pro","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro"},"gpt-oss:20b":{"id":"gpt-oss:20b","name":"gpt-oss:20b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.07,"output":0.3,"cache_read":0.035},"canonical_model_id":"openai/gpt-oss-20b"},"glm-5.1":{"id":"glm-5.1","name":"glm-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"status":"deprecated","cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5.1"},"deepseek-v4.1-flash:1017":{"id":"deepseek-v4.1-flash:1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"kimi-k2.7-code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"kimi-k3":{"id":"kimi-k3","name":"kimi-k3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"kimi-k2.5":{"id":"kimi-k2.5","name":"kimi-k2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","canonical_model_id":"moonshotai/kimi-k2.5"},"deepseek-v4-flash:0731":{"id":"deepseek-v4-flash:0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"status":"deprecated","cost":{"input":0.22,"output":0.66,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"nemotron-3-nano:30b":{"id":"nemotron-3-nano:30b","name":"nemotron-3-nano:30b","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.06,"output":0.24},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"minimax-m3":{"id":"minimax-m3","name":"minimax-m3","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"minimax/MiniMax-M3"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":976000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"gpt-oss:120b":{"id":"gpt-oss:120b","name":"gpt-oss:120b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.014},"canonical_model_id":"openai/gpt-oss-120b"},"minimax-m2.5":{"id":"minimax-m2.5","name":"minimax-m2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","canonical_model_id":"minimax/MiniMax-M2.5"},"minimax-m3.1-flash":{"id":"minimax-m3.1-flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":512000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"beam":{"id":"beam","name":"Beam 1.0 501B A23B","description":"Sparse mixture-of-experts reasoning model with 23B active parameters for coding and agentic workflows","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"qwen3.5:397b":{"id":"qwen3.5:397b","name":"qwen3.5:397b","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":false,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"muse-spark-open":{"id":"muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"}}},"lilac":{"id":"lilac","env":["LILAC_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.getlilac.com/v1","name":"Lilac","doc":"https://docs.getlilac.com/inference/models","models":{"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.7,"output":3.5,"cache_read":0.2},"canonical_model_id":"moonshotai/kimi-k2.6"},"zai-org/glm-5.2":{"id":"zai-org/glm-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":524288},"status":"deprecated","cost":{"input":0.9,"output":3,"cache_read":0.27},"canonical_model_id":"zhipuai/glm-5.2"},"minimaxai/minimax-m3":{"id":"minimaxai/minimax-m3","name":"MiniMax M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"status":"deprecated","cost":{"input":0.28,"output":1.1,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M3"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262100,"output":262100},"status":"deprecated","cost":{"input":0.11,"output":0.35},"canonical_model_id":"google/gemma-4-31b-it"}}},"anyapi":{"id":"anyapi","env":["ANYAPI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.anyapi.ai/v1","name":"AnyAPI","doc":"https://docs.anyapi.ai","models":{"mistralai/devstral-2512":{"id":"mistralai/devstral-2512","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","canonical_model_id":"mistral/devstral-2512"},"mistralai/mistral-large-2512":{"id":"mistralai/mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"canonical_model_id":"mistral/mistral-large-2512"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"canonical_model_id":"openai/o3-mini"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"canonical_model_id":"openai/gpt-5.2"},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"canonical_model_id":"openai/o3"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"canonical_model_id":"openai/gpt-5.4"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"canonical_model_id":"openai/o4-mini"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"canonical_model_id":"openai/gpt-5"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-r1":{"id":"deepseek/deepseek-r1","name":"DeepSeek Reasoner","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"canonical_model_id":"deepseek/deepseek-reasoner"},"deepseek/deepseek-chat":{"id":"deepseek/deepseek-chat","name":"DeepSeek Chat","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"canonical_model_id":"deepseek/deepseek-chat"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-haiku-4-5":{"id":"anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-sonnet-4-5":{"id":"anthropic/claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024,"max":127999}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"canonical_model_id":"anthropic/claude-opus-4-6"},"perplexity/sonar-reasoning-pro":{"id":"perplexity/sonar-reasoning-pro","name":"Sonar Reasoning Pro","description":"Web-grounded Sonar for multi-step research questions that need cited reasoning","family":"sonar-reasoning","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"canonical_model_id":"perplexity/sonar-reasoning-pro"},"perplexity/sonar-pro":{"id":"perplexity/sonar-pro","name":"Sonar Pro","description":"Deeper Sonar search model with broader retrieval and stronger synthesis","family":"sonar-pro","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"canonical_model_id":"perplexity/sonar-pro"},"xai/grok-4.3":{"id":"xai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"canonical_model_id":"xai/grok-4.3"},"cohere/command-r-plus-08-2024":{"id":"cohere/command-r-plus-08-2024","name":"Command R+","description":"Cohere's RAG workhorse for long-context enterprise search and tool use","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"canonical_model_id":"cohere/command-r-plus-08-2024"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-2.5-pro"},"google/gemini-3-pro-preview":{"id":"google/gemini-3-pro-preview","name":"Gemini 3 Pro Preview","description":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-3-pro-preview"}}},"lmstudio":{"id":"lmstudio","env":["LMSTUDIO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"http://127.0.0.1:1234/v1","name":"LMStudio","doc":"https://lmstudio.ai/models","models":{"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-oss-20b"},"qwen/qwen3-30b-a3b-2507":{"id":"qwen/qwen3-30b-a3b-2507","name":"Qwen3 30B A3B 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-30b-a3b-instruct-2507"},"qwen/qwen3-coder-30b":{"id":"qwen/qwen3-coder-30b","name":"Qwen3 Coder 30B","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"}}},"modelscope":{"id":"modelscope","env":["MODELSCOPE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api-inference.modelscope.cn/v1","name":"ModelScope","doc":"https://modelscope.cn/docs/model-service/API-Inference/intro","models":{"ZhipuAI/GLM-4.5":{"id":"ZhipuAI/GLM-4.5","name":"GLM-4.5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.5"},"ZhipuAI/GLM-4.6":{"id":"ZhipuAI/GLM-4.6","name":"GLM-4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":98304},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.6"},"Qwen/Qwen3-235B-A22B-Thinking-2507":{"id":"Qwen/Qwen3-235B-A22B-Thinking-2507","name":"Qwen3-235B-A22B-Thinking-2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"Qwen/Qwen3-30B-A3B-Instruct-2507":{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507","name":"Qwen3 30B A3B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-30b-a3b-instruct-2507"},"Qwen/Qwen3-30B-A3B-Thinking-2507":{"id":"Qwen/Qwen3-30B-A3B-Thinking-2507","name":"Qwen3 30B A3B Thinking 2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0}},"Qwen/Qwen3-Coder-30B-A3B-Instruct":{"id":"Qwen/Qwen3-Coder-30B-A3B-Instruct","name":"Qwen3 Coder 30B A3B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"}}},"clarifai":{"id":"clarifai","env":["CLARIFAI_PAT"],"npm":"@ai-sdk/openai-compatible","api":"https://api.clarifai.com/v2/ext/openai/v1","name":"Clarifai","doc":"https://docs.clarifai.com/compute/inference/","models":{"moonshotai/chat-completion/models/Kimi-K2_6":{"id":"moonshotai/chat-completion/models/Kimi-K2_6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4},"canonical_model_id":"moonshotai/kimi-k2.6"},"mistralai/completion/models/Ministral-3-3B-Reasoning-2512":{"id":"mistralai/completion/models/Ministral-3-3B-Reasoning-2512","name":"Ministral 3 3B Reasoning 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12","last_updated":"2026-02-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.039,"output":0.54825}},"mistralai/completion/models/Ministral-3-14B-Reasoning-2512":{"id":"mistralai/completion/models/Ministral-3-14B-Reasoning-2512","name":"Ministral 3 14B Reasoning 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-01","last_updated":"2025-12-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":2.5,"output":1.7}},"openai/chat-completion/models/gpt-oss-20b":{"id":"openai/chat-completion/models/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.045,"output":0.18},"canonical_model_id":"openai/gpt-oss-20b"},"openai/chat-completion/models/gpt-oss-120b-high-throughput":{"id":"openai/chat-completion/models/gpt-oss-120b-high-throughput","name":"GPT OSS 120B High Throughput","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.09,"output":0.36},"canonical_model_id":"openai/gpt-oss-120b"},"arcee_ai/AFM/models/trinity-mini":{"id":"arcee_ai/AFM/models/trinity-mini","name":"Trinity Mini","description":"Reasoning-tuned 26B MoE model with 3B active parameters for agents, tools, and multi-step workloads","family":"trinity","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-12-01","last_updated":"2026-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.045,"output":0.15},"canonical_model_id":"arcee-ai/trinity-mini"},"clarifai/main/models/mm-poly-8b":{"id":"clarifai/main/models/mm-poly-8b","name":"MM Poly 8B","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"mm-poly","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-06","last_updated":"2026-02-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":4096},"cost":{"input":0.658,"output":1.11}},"deepseek-ai/deepseek-ocr/models/DeepSeek-OCR":{"id":"deepseek-ai/deepseek-ocr/models/DeepSeek-OCR","name":"DeepSeek OCR","description":"OCR model for extracting structured text from documents and screenshots","family":"deepseek","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-10-20","last_updated":"2026-02-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"cost":{"input":0.2,"output":0.7}},"qwen/qwenLM/models/Qwen3-30B-A3B-Instruct-2507":{"id":"qwen/qwenLM/models/Qwen3-30B-A3B-Instruct-2507","name":"Qwen3 30B A3B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":0.5},"canonical_model_id":"alibaba/qwen3-30b-a3b-instruct-2507"},"qwen/qwenLM/models/Qwen3-30B-A3B-Thinking-2507":{"id":"qwen/qwenLM/models/Qwen3-30B-A3B-Thinking-2507","name":"Qwen3 30B A3B Thinking 2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-31","last_updated":"2026-02-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.36,"output":1.3}},"qwen/qwenCoder/models/Qwen3-Coder-30B-A3B-Instruct":{"id":"qwen/qwenCoder/models/Qwen3-Coder-30B-A3B-Instruct","name":"Qwen3 Coder 30B A3B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.11458,"output":0.74812},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"minimaxai/chat-completion/models/MiniMax-M2_5-high-throughput":{"id":"minimaxai/chat-completion/models/MiniMax-M2_5-high-throughput","name":"MiniMax-M2.5 High Throughput","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.5"}}},"watsonx":{"id":"watsonx","env":["WATSONX_AI_APIKEY","WATSONX_AI_PROJECT_ID"],"npm":"watsonx-ai-provider","name":"watsonx.ai","doc":"https://www.ibm.com/docs/en/watsonx/saas?topic=solutions-supported-foundation-models","models":{"mistralai/mistral-small-3-1-24b-instruct-2503":{"id":"mistralai/mistral-small-3-1-24b-instruct-2503","name":"Mistral Small 3.1 24B","description":"Efficient multimodal model for instruction following, coding, reasoning, and function calling","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.106,"output":0.318},"canonical_model_id":"mistral/mistral-small-3-1-24b-instruct-2503"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.159,"output":0.636},"canonical_model_id":"openai/gpt-oss-120b"},"meta-llama/llama-4-maverick-17b-128e-instruct-fp8":{"id":"meta-llama/llama-4-maverick-17b-128e-instruct-fp8","name":"Llama 4 Maverick 17B 128E Instruct FP8","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.371,"output":1.484},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"meta-llama/llama-3-3-70b-instruct":{"id":"meta-llama/llama-3-3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.7526,"output":0.7526},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"ibm/granite-4-h-small":{"id":"ibm/granite-4-h-small","name":"Granite-4.0-H-Small","description":"Open-weight hybrid model for enterprise chat, coding, retrieval-augmented generation, and tool-calling workloads","family":"granite","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-02","last_updated":"2025-10-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.0636,"output":0.265},"canonical_model_id":"ibm/granite-4-h-small"}}},"chutes":{"id":"chutes","env":["CHUTES_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://llm.chutes.ai/v1","name":"Chutes","doc":"https://llm.chutes.ai/v1/models","models":{"Nemotron-3-Nano-Omni-30B-TEE":{"id":"Nemotron-3-Nano-Omni-30B-TEE","name":"Nemotron 3 Nano Omni 30B TEE","description":"Omni-modal model for text, vision, audio, and multimodal agent tasks","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":0},"cost":{"input":0.0245,"output":0.0978,"cache_read":0.0024499999999999995}},"moonshotai/Kimi-K2.6-TEE":{"id":"moonshotai/Kimi-K2.6-TEE","name":"Kimi K2.6 TEE","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65535},"cost":{"input":0.45,"output":1.95,"cache_read":0.04499999999999999},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/Kimi-K3-TEE":{"id":"moonshotai/Kimi-K3-TEE","name":"Kimi K3 TEE","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65535},"cost":{"input":1.85,"output":9.5,"cache_read":0.18499999999999997},"canonical_model_id":"moonshotai/kimi-k3"},"unsloth/Mistral-Nemo-Instruct-2407-TEE":{"id":"unsloth/Mistral-Nemo-Instruct-2407-TEE","name":"Mistral Nemo Instruct 2407 TEE","description":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.0245,"output":0.0978,"cache_read":0.0024499999999999995},"canonical_model_id":"mistral/mistral-nemo"},"zai-org/GLM-5.1-TEE":{"id":"zai-org/GLM-5.1-TEE","name":"GLM 5.1 TEE","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":65535},"cost":{"input":0.98,"output":3.08,"cache_read":0.09799999999999998},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-5.2-TEE":{"id":"zai-org/GLM-5.2-TEE","name":"GLM 5.2 TEE","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65535},"cost":{"input":1.25,"output":3.95,"cache_read":0.12499999999999997},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-ai/DeepSeek-V4-Flash-0731-TEE":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731-TEE","name":"DeepSeek V4 Flash 0731 TEE","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.44,"output":1.32,"cache_read":0.04399999999999999},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-V3.2-TEE":{"id":"deepseek-ai/DeepSeek-V3.2-TEE","name":"DeepSeek V3.2 TEE","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":1,"output":1,"cache_read":0.09999999999999998},"canonical_model_id":"deepseek/deepseek-v3.2"},"Qwen/Qwen3.6-27B-TEE":{"id":"Qwen/Qwen3.6-27B-TEE","name":"Qwen3.6 27B TEE","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2,"cache_read":0.024999999999999994},"canonical_model_id":"alibaba/qwen3.6-27b"},"Qwen/Qwen3-235B-A22B-Thinking-2507-TEE":{"id":"Qwen/Qwen3-235B-A22B-Thinking-2507-TEE","name":"Qwen3 235B A22B Thinking 2507 TEE","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.2989,"output":1.1957,"cache_read":0.029889999999999993},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"Qwen/Qwen3.5-397B-A17B-TEE":{"id":"Qwen/Qwen3.5-397B-A17B-TEE","name":"Qwen3.5 397B A17B TEE","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.45,"output":3,"cache_read":0.04499999999999999},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"Qwen/Qwen3-32B-TEE":{"id":"Qwen/Qwen3-32B-TEE","name":"Qwen3 32B TEE","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":40960},"cost":{"input":0.104,"output":0.416,"cache_read":0.010399999999999998},"canonical_model_id":"alibaba/qwen3-32b"},"Qwen/Qwen3.8-27B-TEE":{"id":"Qwen/Qwen3.8-27B-TEE","name":"Qwen3.8 27B TEE","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.039,"output":1.25,"cache_read":0.003899999999999999},"canonical_model_id":"alibaba/qwen3.8-27b"},"google/gemma-4-31B-turbo-TEE":{"id":"google/gemma-4-31B-turbo-TEE","name":"gemma 4 31B turbo TEE","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.12,"output":0.37,"cache_read":0.011999999999999997},"canonical_model_id":"google/gemma-4-31b-it"}}},"inco":{"id":"inco","env":["INCO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.inco.ai/v1","name":"Inco","doc":"https://platform.inco.ai/docs","models":{"glm-5.3:fast":{"id":"glm-5.3:fast","name":"GLM-5.3 Fast","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2.8,"output":8.8,"cache_read":0.52},"canonical_model_id":"zhipuai/glm-5.3"},"kimi-k3:fast":{"id":"kimi-k3:fast","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":6,"output":30,"cache_read":0.6},"canonical_model_id":"moonshotai/kimi-k3"},"glm-5.3-flash:fast":{"id":"glm-5.3-flash:fast","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.3,"output":1,"cache_read":0.06},"canonical_model_id":"zhipuai/glm-5.3-flash"},"deepseek-v4.1-flash:fast":{"id":"deepseek-v4.1-flash:fast","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.6,"output":2.4,"cache_read":0.012},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"minimax-m3:fast":{"id":"minimax-m3:fast","name":"MiniMax M3 Fast","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"minimax/MiniMax-M3"}}},"lynkr":{"id":"lynkr","env":["LYNKR_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"http://127.0.0.1:8081/v1","name":"Lynkr","doc":"https://github.com/Fast-Editor/Lynkr","models":{"lynkr-auto":{"id":"lynkr-auto","name":"Lynkr Auto (complexity routing)","description":"Virtual model: Lynkr scores each request on complexity and routes it to the tier model the user configured (local Ollama/llama.cpp for simple requests, configured cloud providers for complex ones).","family":"auto","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-12-03","last_updated":"2026-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0}}}},"neon":{"id":"neon","env":["NEON_AI_GATEWAY_BASE_URL","NEON_AI_GATEWAY_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"${NEON_AI_GATEWAY_BASE_URL}/v1","name":"Neon","doc":"https://neon.com/docs","models":{"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gpt-5-6-luna":{"id":"gpt-5-6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"gpt-5-6-terra":{"id":"gpt-5-6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":25000},"cost":{"input":0.07,"output":0.3},"canonical_model_id":"openai/gpt-oss-20b"},"grok-4-6":{"id":"grok-4-6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":524288},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"grok-4-7":{"id":"grok-4-7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":2.5,"output":7.5,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"gpt-5-2":{"id":"gpt-5-2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"gpt-5-6-sol":{"id":"gpt-5-6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":45,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"openai/gpt-5.6-sol"},"gemini-3-flash":{"id":"gemini-3-flash","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"claude-opus-4-1":{"id":"claude-opus-4-1","name":"Claude Opus 4.1 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"meta-llama-3-3-70b-instruct":{"id":"meta-llama-3-3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"gemini-3-5-flash-lite":{"id":"gemini-3-5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"gemini-3-1-pro":{"id":"gemini-3-1-pro","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":25000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"gpt-6-1-sol":{"id":"gpt-6-1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"gpt-5-1":{"id":"gpt-5-1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"gemma-3-12b":{"id":"gemma-3-12b","name":"Gemma 3 12B","description":"Google's open-weight Gemma 3 vision-language model for text and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.15,"output":0.5},"canonical_model_id":"google/gemma-3-12b-it"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"glm-5-3-flash":{"id":"glm-5-3-flash","name":"GLM-5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"gpt-5-5-pro":{"id":"gpt-5-5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":127999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"gemini-3-6-flash":{"id":"gemini-3-6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.6-flash"},"deepseek-v4-1-flash":{"id":"deepseek-v4-1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"gpt-5-4":{"id":"gpt-5-4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"gemini-3-8-flash":{"id":"gemini-3-8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.8-flash"},"qwen3-next-80b-a3b-instruct":{"id":"qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":10000},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"qwen35-122b-a10b":{"id":"qwen35-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":25000},"cost":{"input":0.22,"output":2.2},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"meta-llama-3-1-8b-instruct":{"id":"meta-llama-3-1-8b-instruct","name":"Llama 3.1 8B Instruct","description":"Meta's compact open-weight Llama 3.1 model for fast, low-cost text generation","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.15,"output":0.45},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"gemini-3-5-flash":{"id":"gemini-3-5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"glm-5-3":{"id":"glm-5-3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-5-4-mini":{"id":"gpt-5-4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"glm-5-2":{"id":"glm-5-2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"gpt-5-4-nano":{"id":"gpt-5-4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"gemini-3-7-flash":{"id":"gemini-3-7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-3.7-flash"},"gpt-5-5":{"id":"gpt-5-5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"inkling":{"id":"inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"gemini-3-1-flash-lite":{"id":"gemini-3-1-flash-lite","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"gpt-5-3-codex":{"id":"gpt-5-3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"${NEON_AI_GATEWAY_BASE_URL}/openai/v1","shape":"responses"},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"llama-4-maverick":{"id":"llama-4-maverick","name":"Llama 4 Maverick 17B Instruct","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":8192},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"}}},"vercel":{"id":"vercel","env":["AI_GATEWAY_API_KEY"],"npm":"@ai-sdk/gateway","name":"Vercel AI Gateway","doc":"https://github.com/vercel/ai/tree/5eb85cc45a259553501f535b8ac79a77d0e79223/packages/gateway","models":{"meta/muse-wonder-1.0":{"id":"meta/muse-wonder-1.0","name":"Muse Wonder","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-image-1.0":{"id":"meta/muse-image-1.0","type":"image","name":"Muse Image 1.0","description":"Agentic image generation and editing model that plans, searches and refines before rendering","family":"muse","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-07-07","last_updated":"2026-07-07","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"meta/muse-image-1.0"},"meta/muse-spark-1.3-contributor":{"id":"meta/muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/muse-glimmer-30b":{"id":"meta/muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.35,"output":1.5,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"meta/llama-3.1-8b":{"id":"meta/llama-3.1-8b","name":"Llama 3.1 8B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.22,"output":0.22},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"meta/muse-spark-1.2-contributor":{"id":"meta/muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/llama-4-scout":{"id":"meta/llama-4-scout","name":"Llama-4-Scout-17B-16E-Instruct-FP8","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.17,"output":0.66},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"meta/llama-3.1-70b":{"id":"meta/llama-3.1-70b","name":"Llama 3.1 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-3.1-70b-instruct"},"meta/llama-3.3-70b":{"id":"meta/llama-3.3-70b","name":"Llama-3.3-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta/llama-4-maverick":{"id":"meta/llama-4-maverick","name":"Llama-4-Maverick-17B-128E-Instruct-FP8","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.24,"output":0.97},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"meta/muse-spark-open":{"id":"meta/muse-spark-open","name":"Muse Spark Open","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"sakana/namazu":{"id":"sakana/namazu","name":"Sakana Namazu","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"sakana-namazu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.95,"output":4,"cache_read":0.15},"canonical_model_id":"sakana/sakana-namazu"},"sakana/fugu-ultra-v2":{"id":"sakana/fugu-ultra-v2","name":"Fugu Ultra v2","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-08-28","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra-v2"},"sakana/fugu-max":{"id":"sakana/fugu-max","name":"Fugu Max","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"sakana/fugu-max"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra"},"fish-audio/drama-3":{"id":"fish-audio/drama-3","type":"speech","name":"Drama 3","description":"Speech generation model for controllable voice, narration, and audio delivery","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"fish-audio/s2.1-pro":{"id":"fish-audio/s2.1-pro","type":"speech","name":"S2.1 Pro","description":"Speech generation model for controllable voice, narration, and audio delivery","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-07-28","last_updated":"2026-07-28","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"fish-audio/s1":{"id":"fish-audio/s1","type":"speech","name":"S1","description":"Speech generation model for controllable voice, narration, and audio delivery","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-20","last_updated":"2025-10-20","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"fish-audio/s2-pro":{"id":"fish-audio/s2-pro","type":"speech","name":"S2 Pro","description":"Speech generation model for controllable voice, narration, and audio delivery","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"fish-audio/transcribe-1":{"id":"fish-audio/transcribe-1","type":"transcription","name":"Transcribe-1","description":"Speech transcription model for accurate audio-to-text and captioning workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-01","last_updated":"2026-03-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k2.7-code-highspeed":{"id":"moonshotai/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code High Speed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.6,"output":3},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/kimi-k3-fast":{"id":"moonshotai/kimi-k3-fast","name":"Kimi K3 Fast","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":4.5,"output":22.5,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2":{"id":"moonshotai/kimi-k2","name":"Kimi K2 Instruct","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-12-31","release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.57,"output":2.3},"canonical_model_id":"moonshotai/kimi-k2"},"mixedbread/toast-1":{"id":"mixedbread/toast-1","name":"Toast 1","description":"Specialized search model for knowledge-intensive questions, multi-step retrieval, and evidence synthesis","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":4000},"cost":{"input":0.3,"output":0.72,"cache_read":0.036},"canonical_model_id":"mixedbread/toast-1"},"topaz/wonder-3.5":{"id":"topaz/wonder-3.5","type":"image","name":"Wonder 3.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"topaz/starlight-precise-2.6":{"id":"topaz/starlight-precise-2.6","type":"video","name":"Starlight Precise 2.6","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"topaz/proteus":{"id":"topaz/proteus","type":"video","name":"Proteus","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"mistral/voxtral-mini-transcribe-3":{"id":"mistral/voxtral-mini-transcribe-3","type":"transcription","name":"Voxtral Mini Transcribe 3","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"mistral","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-20","last_updated":"2026-10-20","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"mistral/mistral-small":{"id":"mistral/mistral-small","name":"Mistral Small (latest)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":4000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"mistral/mistral-small-latest"},"mistral/mistral-large-4":{"id":"mistral/mistral-large-4","name":"Mistral Large 4","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":262144},"cost":{"input":1.36,"output":4.18,"cache_read":0.14},"canonical_model_id":"mistral/mistral-large-4"},"mistral/voxtral-mini-tts-3":{"id":"mistral/voxtral-mini-tts-3","type":"speech","name":"Voxtral TTS 3","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-20","last_updated":"2026-10-20","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"mistral/ministral-8b":{"id":"mistral/ministral-8b","name":"Ministral 8B (latest)","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-10-16","last_updated":"2024-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":4000},"cost":{"input":0.15,"output":0.15,"cache_read":0.015}},"mistral/codestral-embed":{"id":"mistral/codestral-embed","type":"embedding","name":"Codestral Embed","description":"Embedding model for code search, retrieval, and code-aware RAG pipelines","family":"codestral-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":3072},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"mistral/codestral-embed-2505"},"mistral/ministral-3b":{"id":"mistral/ministral-3b","name":"Ministral 3B (latest)","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-10-16","last_updated":"2024-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4000},"cost":{"input":0.1,"output":0.1,"cache_read":0.01}},"mistral/ministral-14b":{"id":"mistral/ministral-14b","name":"Ministral 14B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.2,"output":0.2,"cache_read":0.02},"canonical_model_id":"mistral/ministral-14b"},"mistral/mistral-nemo":{"id":"mistral/mistral-nemo","name":"Mistral Nemo","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.029,"output":0.03},"canonical_model_id":"mistral/mistral-nemo"},"mistral/mistral-large-3":{"id":"mistral/mistral-large-3","name":"Mistral Large 3","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"mistral/mistral-large-2512"},"mistral/mistral-embed":{"id":"mistral/mistral-embed","type":"embedding","name":"Mistral Embed","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"mistral-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-12-11","last_updated":"2023-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":3072},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"mistral/mistral-embed"},"mistral/mistral-medium-3.5":{"id":"mistral/mistral-medium-3.5","name":"Mistral Medium Latest","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":256000},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"mistral/mistral-medium-3-5"},"mistral/voxtral-small-transcribe-3":{"id":"mistral/voxtral-small-transcribe-3","type":"transcription","name":"Voxtral Small Transcribe 3","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"mistral","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-20","last_updated":"2026-10-20","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"mistral/codestral":{"id":"mistral/codestral","name":"Codestral (latest)","description":"Mistral code model for completions, refactors, and developer IDE workflows","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-05-29","last_updated":"2025-01-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.3,"output":0.9,"cache_read":0.03},"canonical_model_id":"mistral/codestral-latest"},"bfl/flux-pro-1.1":{"id":"bfl/flux-pro-1.1","type":"image","name":"FLUX1.1 [pro]","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-10-02","last_updated":"2024-10","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bfl/flux-2-max":{"id":"bfl/flux-2-max","type":"image","name":"FLUX.2 [max]","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-16","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":67300,"output":67300},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bfl/flux-2-klein-4b":{"id":"bfl/flux-2-klein-4b","type":"image","name":"FLUX.2 [klein] 4B","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-15","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bfl/flux-3-video":{"id":"bfl/flux-3-video","type":"video","name":"Flux 3","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"flux","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-08-04","last_updated":"2026-08-04","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bfl/flux-kontext-max":{"id":"bfl/flux-kontext-max","type":"image","name":"FLUX.1 Kontext Max","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-05-29","last_updated":"2025-06","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bfl/flux-3-image-klein":{"id":"bfl/flux-3-image-klein","type":"image","name":"FLUX 3 Image Klein","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bfl/flux-2-flex":{"id":"bfl/flux-2-flex","type":"image","name":"FLUX.2 [flex]","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-11-25","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bfl/flux-3-image":{"id":"bfl/flux-3-image","type":"image","name":"FLUX 3 Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bfl/flux-pro-1.1-ultra":{"id":"bfl/flux-pro-1.1-ultra","type":"image","name":"FLUX1.1 [pro] Ultra","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-11-01","last_updated":"2024-11","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bfl/flux-2-klein-9b":{"id":"bfl/flux-2-klein-9b","type":"image","name":"FLUX.2 [klein] 9B","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-15","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bfl/flux-kontext-pro":{"id":"bfl/flux-kontext-pro","type":"image","name":"FLUX.1 Kontext Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-05-29","last_updated":"2025-06","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bfl/flux-2-pro":{"id":"bfl/flux-2-pro","type":"image","name":"FLUX.2 [pro]","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-11-25","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":67300,"output":67300},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bfl/flux-pro-1.0-fill":{"id":"bfl/flux-pro-1.0-fill","type":"image","name":"FLUX.1 Fill [pro]","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-10-01","last_updated":"2024-10","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"inclusionai/ling-3.0-flash":{"id":"inclusionai/ling-3.0-flash","name":"Ling 3.0 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"cost":{"input":0.021,"output":0.063,"cache_read":0.0042},"canonical_model_id":"inclusionai/ling-3.0-flash"},"inclusionai/ling-3.0-flash-vl":{"id":"inclusionai/ling-3.0-flash-vl","name":"Ling 3.0 Flash VL","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash-vl"},"inclusionai/ling-3.1-flash":{"id":"inclusionai/ling-3.1-flash","name":"Ling 3.1 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.06},"canonical_model_id":"inclusionai/ling-3.1-flash"},"inclusionai/ling-3.0-flash-sante":{"id":"inclusionai/ling-3.0-flash-sante","name":"Ling 3.0 Flash Sante","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash-sante"},"inclusionai/ling-3.0-flash-fin":{"id":"inclusionai/ling-3.0-flash-fin","name":"Ling 3.0 Flash Fin","description":"Finance-enhanced model for financial research, multi-step investment workflows, and long-horizon planning and execution","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-09-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash-fin"},"voyage/voyage-3-large":{"id":"voyage/voyage-3-large","type":"embedding","name":"voyage-3-large","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-01-07","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"voyage/voyage-code-2":{"id":"voyage/voyage-code-2","type":"embedding","name":"voyage-code-2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-01-01","last_updated":"2024-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"voyage/rerank-2.5-lite":{"id":"voyage/rerank-2.5-lite","type":"reranking","name":"Voyage Rerank 2.5 Lite","description":"Instruction-following reranker optimized for latency and quality","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"voyageai/rerank-2.5-lite"},"voyage/rerank-3-lite":{"id":"voyage/rerank-3-lite","type":"reranking","name":"Voyage Rerank 3 Lite","description":"Reranker optimized for latency and cost, successor to rerank-2.5-lite","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-30","last_updated":"2026-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"voyageai/rerank-3-lite"},"voyage/voyage-3.5":{"id":"voyage/voyage-3.5","type":"embedding","name":"voyage-3.5","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"voyage/rerank-3":{"id":"voyage/rerank-3","type":"reranking","name":"Voyage Rerank 3","description":"Reranker optimized for retrieval quality, successor to rerank-2.5","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-30","last_updated":"2026-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"voyageai/rerank-3"},"voyage/voyage-law-2":{"id":"voyage/voyage-law-2","type":"embedding","name":"voyage-law-2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-04-15","last_updated":"2024-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"voyage/voyage-finance-2":{"id":"voyage/voyage-finance-2","type":"embedding","name":"voyage-finance-2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-06-03","last_updated":"2024-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"voyage/voyage-4":{"id":"voyage/voyage-4","type":"embedding","name":"voyage-4","description":"General-purpose and multilingual embedding model balancing retrieval quality and cost","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-15","last_updated":"2026-01-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"voyageai/voyage-4"},"voyage/voyage-3.5-lite":{"id":"voyage/voyage-3.5-lite","type":"embedding","name":"voyage-3.5-lite","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"voyage/voyage-code-3":{"id":"voyage/voyage-code-3","type":"embedding","name":"voyage-code-3","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-12-04","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"voyage/voyage-4-large":{"id":"voyage/voyage-4-large","type":"embedding","name":"voyage-4-large","description":"General-purpose and multilingual embedding model optimized for retrieval quality","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-15","last_updated":"2026-01-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"voyageai/voyage-4-large"},"voyage/voyage-4-lite":{"id":"voyage/voyage-4-lite","type":"embedding","name":"voyage-4-lite","description":"Lightweight general-purpose embedding model optimized for latency and cost","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-15","last_updated":"2026-01-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"voyageai/voyage-4-lite"},"voyage/rerank-2.5":{"id":"voyage/rerank-2.5","type":"reranking","name":"Voyage Rerank 2.5","description":"Instruction-following reranker optimized for retrieval quality","family":"voyage","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"voyageai/rerank-2.5"},"liquid/lfm2.5-24b-a2b":{"id":"liquid/lfm2.5-24b-a2b","name":"LFM2.5-24B-A2B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"liquid","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8200},"cost":{"input":0.05,"output":0.7,"cache_read":0.025}},"poolside/laguna-s-2.1":{"id":"poolside/laguna-s-2.1","name":"Laguna S 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.09,"output":0.18,"cache_read":0.009},"canonical_model_id":"poolside/laguna-s-2.1"},"poolside/laguna-s-2.1-free":{"id":"poolside/laguna-s-2.1-free","name":"Laguna S 2.1 Free","description":"Free provider route for experiments, demos, and cost-sensitive chat workloads","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-s-2.1"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":1000000},"cost":{"input":0.45,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"quiverai/arrow-2":{"id":"quiverai/arrow-2","name":"Arrow 2","description":"Fast SVG generation model for creation, vectorization, editing, and animation","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"release_date":"2026-09-16","last_updated":"2026-09-16","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5},"canonical_model_id":"quiverai/arrow-2"},"quiverai/arrow-2-telos":{"id":"quiverai/arrow-2-telos","name":"Arrow 2 Telos","description":"High-fidelity SVG generation model for complex vector work and long-context refinement","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"temperature":false,"release_date":"2026-09-16","last_updated":"2026-09-16","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":6,"output":30,"cache_read":0.6,"cache_write":7.5},"canonical_model_id":"quiverai/arrow-2-telos"},"callstack/apex":{"id":"callstack/apex","name":"Apex","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"release_date":"2026-10-08","last_updated":"2026-10-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.5,"output":3,"cache_read":0.2}},"recraft/recraft-v4-pro":{"id":"recraft/recraft-v4-pro","type":"image","name":"Recraft V4 Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"recraft/recraft-v4.1-utility":{"id":"recraft/recraft-v4.1-utility","type":"image","name":"Recraft V4.1 Utility","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-05-14","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"recraft/recraft-v3":{"id":"recraft/recraft-v3","type":"image","name":"Recraft V3","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-10-30","last_updated":"2024-10","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"recraft/recraft-v2":{"id":"recraft/recraft-v2","type":"image","name":"Recraft V2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-03-13","last_updated":"2024-03","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"recraft/recraft-v4":{"id":"recraft/recraft-v4","type":"image","name":"Recraft V4","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"recraft/recraft-v4.1-pro":{"id":"recraft/recraft-v4.1-pro","type":"image","name":"Recraft V4.1 Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-05-14","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"recraft/recraft-v4.1-utility-pro":{"id":"recraft/recraft-v4.1-utility-pro","type":"image","name":"Recraft V4.1 Utility Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-05-14","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"recraft/recraft-v4.1":{"id":"recraft/recraft-v4.1","type":"image","name":"Recraft V4.1","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-05-14","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"recraft/recraft-v4.1-flash":{"id":"recraft/recraft-v4.1-flash","type":"image","name":"Recraft V4.1 Flash","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"recraft","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"morph/morph-v3-fast":{"id":"morph/morph-v3-fast","name":"Morph v3 Fast","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-08-15","last_updated":"2024-08-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":16000},"cost":{"input":0.8,"output":1.2}},"morph/morph-v3-large":{"id":"morph/morph-v3-large","name":"Morph v3 Large","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-08-15","last_updated":"2024-08-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"cost":{"input":0.9,"output":1.9}},"inception/mercury-2.5":{"id":"inception/mercury-2.5","name":"Mercury 2.5","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":260000,"output":65536},"cost":{"input":0.04,"output":0.15,"cache_read":0.004},"canonical_model_id":"inception/mercury-2.5"},"inception/mercury-3":{"id":"inception/mercury-3","name":"Mercury 3","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1,"output":0.35,"cache_read":0.01}},"inception/mercury-coder-small":{"id":"inception/mercury-coder-small","name":"Mercury Coder Small Beta","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"mercury","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-02-26","last_updated":"2025-02-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":16384},"cost":{"input":0.25,"output":1}},"inception/mercury-2":{"id":"inception/mercury-2","name":"Mercury 2","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":50000},"cost":{"input":0.25,"output":0.75,"cache_read":0.025},"canonical_model_id":"inception/mercury-2"},"openai/gpt-5.6-terra-fast":{"id":"openai/gpt-5.6-terra-fast","name":"GPT 5.6 Terra (Fast)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":24,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":36,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":36,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":128000,"output":272000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-5-mini-fast":{"id":"openai/gpt-5-mini-fast","name":"GPT-5 mini (Fast)","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.45,"output":3.6,"cache_read":0.045},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5-fast":{"id":"openai/gpt-5-fast","name":"GPT-5 (Fast)","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":2.5,"output":20,"cache_read":0.25},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT 5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":872000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-oss-safeguard-120b":{"id":"openai/gpt-oss-safeguard-120b","name":"GPT OSS Safeguard 120B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":112000,"output":16000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-safeguard-120b"},"openai/gpt-4o-fast":{"id":"openai/gpt-4o-fast","name":"GPT-4o (Fast)","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":111616,"output":16384},"cost":{"input":4.25,"output":17,"cache_read":2.125},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":122880,"output":8192},"cost":{"input":0.03,"output":0.14},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-4o-mini-fast":{"id":"openai/gpt-4o-mini-fast","name":"GPT-4o mini (Fast)","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":111616,"output":16384},"cost":{"input":0.25,"output":1,"cache_read":0.125},"canonical_model_id":"openai/gpt-4o-mini"},"openai/tts-1":{"id":"openai/tts-1","type":"speech","name":"TTS-1","description":"Text-to-speech model optimized for low latency","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"openai/tts-1"},"openai/gpt-4.1-nano-fast":{"id":"openai/gpt-4.1-nano-fast","name":"GPT-4.1 nano (Fast)","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"input":1014808,"output":32768},"cost":{"input":0.2,"output":0.8,"cache_read":0.05},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT 5.6 Terra","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT 5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2-Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/text-embedding-3-large":{"id":"openai/text-embedding-3-large","type":"embedding","name":"text-embedding-3-large","description":"Most capable OpenAI embedding model for English and non-English text","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"openai/text-embedding-3-large"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT 5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5.4-fast":{"id":"openai/gpt-5.4-fast","name":"GPT 5.4 (Fast)","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5.2-fast":{"id":"openai/gpt-5.2-fast","name":"GPT 5.2 (Fast)","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":3.5,"output":28,"cache_read":0.35},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-oss-safeguard-20b":{"id":"openai/gpt-oss-safeguard-20b","name":"gpt-oss-safeguard-20b","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":112000,"output":16000},"cost":{"input":0.07,"output":0.2},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT 5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":872000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/o4-mini-fast":{"id":"openai/o4-mini-fast","name":"o4-mini (Fast)","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":100000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o4-mini"},"openai/o3-fast":{"id":"openai/o3-fast","name":"o3 (Fast)","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":100000,"output":100000},"cost":{"input":3.5,"output":14,"cache_read":0.875},"canonical_model_id":"openai/o3"},"openai/gpt-6.1-sol-fast":{"id":"openai/gpt-6.1-sol-fast","name":"GPT-6.1 Sol (Fast)","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.4,"cache_write":10,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.4,"cache_write":10}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-4o-transcribe":{"id":"openai/gpt-4o-transcribe","type":"transcription","name":"GPT-4o Transcribe","description":"Speech-to-text model powered by GPT-4o for accurate transcription","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-06-01","release_date":"2025-03-20","last_updated":"2025-03-20","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":2000},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":2.5,"output":10,"input_audio":2.5},"canonical_model_id":"openai/gpt-4o-transcribe"},"openai/gpt-5.6-luna-fast":{"id":"openai/gpt-5.6-luna-fast","name":"GPT 5.6 Luna (Fast)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.4,"output":2.4,"cache_read":0.04,"cache_write":0.5,"tiers":[{"input":0.8,"output":3.6,"cache_read":0.08,"cache_write":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.8,"output":3.6,"cache_read":0.08,"cache_write":1}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5-codex":{"id":"openai/gpt-5-codex","name":"GPT-5-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5-codex"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT 5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-4o-mini-transcribe":{"id":"openai/gpt-4o-mini-transcribe","type":"transcription","name":"GPT-4o mini Transcribe","description":"Speech-to-text model powered by GPT-4o Mini for fast, accurate transcription","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-06-01","release_date":"2025-03-20","last_updated":"2025-03-20","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":2000},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":1.25,"output":5,"input_audio":1.25},"canonical_model_id":"openai/gpt-4o-mini-transcribe"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-realtime-2":{"id":"openai/gpt-realtime-2","type":"realtime","name":"gpt-realtime-2","description":"Realtime speech-to-speech model with configurable reasoning and reliable tool use for voice agents","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":128000,"output":32000},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":4,"output":24,"cache_read":0.4,"input_audio":32,"output_audio":64},"canonical_model_id":"openai/gpt-realtime-2"},"openai/gpt-5.1-thinking":{"id":"openai/gpt-5.1-thinking","name":"GPT 5.1 Thinking","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-11-12","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125}},"openai/gpt-image-2":{"id":"openai/gpt-image-2","type":"image","name":"GPT Image 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2"},"openai/gpt-4.1-mini-fast":{"id":"openai/gpt-4.1-mini-fast","name":"GPT-4.1 mini (Fast)","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"input":1014808,"output":32768},"cost":{"input":0.7,"output":2.8,"cache_read":0.175},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.1,"output":0.5,"cache_read":0.1},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT 5.6 Sol","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6-sol-fast":{"id":"openai/gpt-6-sol-fast","name":"GPT-6 Sol (Fast)","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.6-sol-fast":{"id":"openai/gpt-5.6-sol-fast","name":"GPT 5.6 Sol (Fast)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":8,"output":40,"cache_read":0.8,"cache_write":10,"tiers":[{"input":16,"output":60,"cache_read":1.6,"cache_write":20,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":16,"output":60,"cache_read":1.6,"cache_write":20}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/text-embedding-3-small":{"id":"openai/text-embedding-3-small","type":"embedding","name":"text-embedding-3-small","description":"Small, efficient OpenAI text embedding model","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"openai/text-embedding-3-small"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT 5.6 Luna","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-image-1-mini":{"id":"openai/gpt-image-1-mini","type":"image","name":"GPT Image 1 Mini","description":"Cost-efficient GPT Image 1 variant for image generation and editing","family":"gpt-image","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":2,"output":8,"cache_read":0.2},"canonical_model_id":"openai/gpt-image-1-mini"},"openai/gpt-live-1":{"id":"openai/gpt-live-1","type":"realtime","name":"GPT-Live 1","description":"Full-duplex voice model for natural conversations that delegates reasoning and tools to a backend agent","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"openai/gpt-live-1"},"openai/gpt-image-1":{"id":"openai/gpt-image-1","type":"image","name":"GPT Image 1","description":"OpenAI image model for production generation, edits, and brand-safe visual workflows","family":"gpt-image","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-24","last_updated":"2025-04-24","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":5,"output":40,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-1"},"openai/gpt-realtime-2.1":{"id":"openai/gpt-realtime-2.1","type":"realtime","name":"gpt-realtime-2.1","description":"Realtime speech-to-speech model with configurable reasoning, tool use, and robust voice-agent behavior","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":128000,"input":96000,"output":32000},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":4,"output":24,"cache_read":0.4,"input_audio":32,"output_audio":64},"canonical_model_id":"openai/gpt-realtime-2.1"},"openai/tts-1-hd":{"id":"openai/tts-1-hd","type":"speech","name":"TTS-1 HD","description":"Text-to-speech model optimized for audio quality","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"openai/tts-1-hd"},"openai/whisper-1":{"id":"openai/whisper-1","type":"transcription","name":"Whisper","description":"General-purpose speech recognition model for transcription and translation","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2023-03-01","last_updated":"2023-03-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"openai/whisper-1"},"openai/gpt-image-2.5-sunburst":{"id":"openai/gpt-image-2.5-sunburst","type":"image","name":"GPT Image 2.5 Sunburst","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-sunburst"},"openai/gpt-5.3-codex-fast":{"id":"openai/gpt-5.3-codex-fast","name":"GPT 5.3 Codex (Fast)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":3.5,"output":28,"cache_read":0.35},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/gpt-image-1.5":{"id":"openai/gpt-image-1.5","type":"image","name":"GPT Image 1.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":5,"output":32,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-1.5"},"openai/gpt-image-2.5-flare":{"id":"openai/gpt-image-2.5-flare","type":"image","name":"GPT Image 2.5 Flare","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-flare"},"openai/gpt-5.5-fast":{"id":"openai/gpt-5.5-fast","name":"GPT 5.5 (Fast)","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":872000,"output":128000},"cost":{"input":12.5,"output":75,"cache_read":1.25},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"input":12289,"output":4096},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-5.4-mini-fast":{"id":"openai/gpt-5.4-mini-fast","name":"GPT 5.4 Mini (Fast)","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.1-thinking-fast":{"id":"openai/gpt-5.1-thinking-fast","name":"GPT 5.1 Thinking (Fast)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"release_date":"2025-11-12","last_updated":"2025-11-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":2.5,"output":20,"cache_read":0.25}},"openai/gpt-6-luna-fast":{"id":"openai/gpt-6-luna-fast","name":"GPT-6 Luna (Fast)","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.5,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":0.4,"output":1.5,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-realtime-whisper":{"id":"openai/gpt-realtime-whisper","type":"transcription","name":"gpt-realtime-whisper","description":"Streaming speech-to-text model for low-latency transcript deltas from live audio","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":2000},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"openai/gpt-realtime-whisper"},"openai/gpt-5.1-codex-max":{"id":"openai/gpt-5.1-codex-max","name":"GPT 5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex-max"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.03},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/gpt-4.1-fast":{"id":"openai/gpt-4.1-fast","name":"GPT-4.1 (Fast)","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"input":1014808,"output":32768},"cost":{"input":3.5,"output":14,"cache_read":0.875},"canonical_model_id":"openai/gpt-4.1"},"openai/o3-pro":{"id":"openai/o3-pro","name":"o3 Pro","description":"High-effort o3 tier for difficult technical reasoning and careful answers","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":100000,"output":100000},"cost":{"input":20,"output":80},"canonical_model_id":"openai/o3-pro"},"openai/gpt-realtime-1.5":{"id":"openai/gpt-realtime-1.5","type":"realtime","name":"GPT-Realtime-1.5","description":"Realtime speech-to-speech model for voice agents and customer support","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":32000,"output":4096},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":4,"output":16,"cache_read":0.4,"input_audio":32,"output_audio":64},"canonical_model_id":"openai/gpt-realtime-1.5"},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT-5.1-Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5.1-codex"},"openai/gpt-6-astra-fast":{"id":"openai/gpt-6-astra-fast","name":"GPT-6 Astra (Fast)","description":"Fast variant of GPT-6 Astra for low-latency assistance and high-volume workloads.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":20,"output":100,"cache_read":2,"cache_write":25,"tiers":[{"input":40,"output":150,"cache_read":4,"cache_write":50,"tier":{"type":"context","size":272001}}],"context_over_200k":{"input":40,"output":150,"cache_read":4,"cache_write":50}},"canonical_model_id":"openai/gpt-6-astra-fast"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT 5.2 ","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/gpt-realtime-mini":{"id":"openai/gpt-realtime-mini","type":"realtime","name":"GPT-Realtime mini","description":"Cost-efficient realtime speech-to-speech model for voice agents","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-10-01","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":32000,"output":4096},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"input_audio":10,"output_audio":20},"canonical_model_id":"openai/gpt-realtime-mini"},"openai/text-embedding-ada-002":{"id":"openai/text-embedding-ada-002","type":"embedding","name":"text-embedding-ada-002","description":"Legacy OpenAI text embedding model","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2022-12-15","last_updated":"2022-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"openai/text-embedding-ada-002"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT 5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT 5.4 Nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":2.5,"output":20,"cache_read":0.25},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.625,"output":5,"cache_read":0.0625},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":15,"output":60,"cache_read":7.5}},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":14,"cache_read":0.875},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2,"output":8,"cache_read":0.5}},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","experimental":{"modes":{"fast":{"cost":{"input":2,"output":8,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.55,"output":2.2,"cache_read":0.1375},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":1.1,"output":4.4,"cache_read":0.275}},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":10,"output":30}},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"experimental":{"modes":{"fast":{"cost":{"input":0.25,"output":1,"cache_read":0.125},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.15,"output":0.6,"cache_read":0.075}},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","experimental":{"modes":{"fast":{"cost":{"input":0.2,"output":0.8,"cache_read":0.05},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.1,"output":0.4,"cache_read":0.025}},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"experimental":{"modes":{"fast":{"cost":{"input":4.25,"output":17,"cache_read":2.125},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":10,"cache_read":1.25}},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":14,"cache_read":0.875},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":1,"output":4,"cache_read":0.25},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":2,"output":8,"cache_read":0.5}},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"flex":{"cost":{"input":0.025,"output":0.2,"cache_read":0.0025},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":0.05,"output":0.4,"cache_read":0.005}},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":0.45,"output":3.6,"cache_read":0.045},"provider":{"body":{"service_tier":"priority"}}},"flex":{"cost":{"input":0.125,"output":1,"cache_read":0.0125},"provider":{"body":{"service_tier":"flex"}}}}},"cost":{"input":0.25,"output":2,"cache_read":0.025}},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.55}},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"experimental":{"modes":{"fast":{"cost":{"input":0.7,"output":2.8,"cache_read":0.175},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.4,"output":1.6,"cache_read":0.1}},"deepseek/deepseek-v3.1":{"id":"deepseek/deepseek-v3.1","name":"DeepSeek-V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":128000},"cost":{"input":0.25,"output":0.95,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek/deepseek-v4.1-flash-1017-fast":{"id":"deepseek/deepseek-v4.1-flash-1017-fast","name":"DeepSeek V4.1 Flash 1017 Fast","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0.6,"output":2.4,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.13,"output":0.26,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8000},"cost":{"input":0.62,"output":1.85},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.075,"output":0.3,"cache_read":0.0015},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.066},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.2156,"output":0.6468,"cache_read":0.0068},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4.1-flash-fast":{"id":"deepseek/deepseek-v4.1-flash-fast","name":"DeepSeek V4.1 Flash Fast","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.6,"output":2.4,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.075,"output":0.3,"cache_read":0.0015},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.076,"output":0.153,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-v3.2-thinking":{"id":"deepseek/deepseek-v3.2-thinking","name":"DeepSeek V3.2 Thinking","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8000},"cost":{"input":0.62,"output":1.85},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-r1":{"id":"deepseek/deepseek-r1","name":"DeepSeek-R1","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"status":"deprecated","cost":{"input":1.35,"output":5.4},"canonical_model_id":"deepseek/deepseek-r1"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-5.5-fast":{"id":"anthropic/claude-opus-5.5-fast","name":"Claude Opus 5.5 (Fast)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":8,"output":40,"cache_read":0.4,"cache_write":10},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-5-fast":{"id":"anthropic/claude-opus-5-fast","name":"Claude Opus 5 (Fast)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-opus-4.8-fast":{"id":"anthropic/claude-opus-4.8-fast","name":"Claude Opus 4.8 (Fast)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100001}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]},{"type":"budget_tokens"}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"prodia/flux-fast-schnell":{"id":"prodia/flux-fast-schnell","type":"image","name":"Flux Schnell","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-08-02","last_updated":"2026-06-08","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":512,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"klingai/kling-v3.0-t2v":{"id":"klingai/kling-v3.0-t2v","type":"video","name":"Kling v3.0 Text-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"ling","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"klingai/kling-v2.6-motion-control":{"id":"klingai/kling-v2.6-motion-control","type":"video","name":"Kling v2.6 Motion Control","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"ling","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-18","last_updated":"2025-12-21","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"klingai/kling-v2.6-i2v":{"id":"klingai/kling-v2.6-i2v","type":"video","name":"Kling v2.6 Image-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"ling","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-03","last_updated":"2025-12-21","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"klingai/kling-v2.5-turbo-t2v":{"id":"klingai/kling-v2.5-turbo-t2v","type":"video","name":"Kling v2.5 Turbo Text-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"ling","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"klingai/kling-v3.0-motion-control":{"id":"klingai/kling-v3.0-motion-control","type":"video","name":"Kling v3.0 Motion Control","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"ling","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-04","last_updated":"2026-03-04","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"klingai/kling-v2.6-t2v":{"id":"klingai/kling-v2.6-t2v","type":"video","name":"Kling v2.6 Text-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"ling","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-03","last_updated":"2025-12-21","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"klingai/kling-v2.5-turbo-i2v":{"id":"klingai/kling-v2.5-turbo-i2v","type":"video","name":"Kling v2.5 Turbo Image-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"ling","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"klingai/kling-v3.0-i2v":{"id":"klingai/kling-v3.0-i2v","type":"video","name":"Kling v3.0 Image-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"ling","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"tencent/hy-mt2-pro":{"id":"tencent/hy-mt2-pro","name":"Tencent Hy-MT2-Pro","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":4000},"cost":{"input":0.074,"output":0.295}},"tencent/hy-mt2-lite":{"id":"tencent/hy-mt2-lite","name":"Tencent Hy-MT2-Lite","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":4000},"cost":{"input":0.044,"output":0.177}},"tencent/hy-mt2-plus":{"id":"tencent/hy-mt2-plus","name":"Tencent Hy-MT2-Plus","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":4000},"cost":{"input":0.074,"output":0.295}},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Tencent Hy4 Preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.14,"output":0.58,"cache_read":0.035},"canonical_model_id":"tencent/hy3"},"stealth/glyph-cluster":{"id":"stealth/glyph-cluster","name":"Glyph Cluster","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0,"output":0}},"xiaomi/mimo-v2.6-pro-ultraspeed":{"id":"xiaomi/mimo-v2.6-pro-ultraspeed","name":"MiMo V2.6 Pro UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.35,"output":8.7,"cache_read":0.036},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo M2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":131100},"cost":{"input":0.119,"output":0.238,"cache_read":0.0025},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2.5-pro":{"id":"xiaomi/mimo-v2.5-pro","name":"MiMo V2.5 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":131000},"cost":{"input":0.3045,"output":0.609,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo V2.6 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"xiaomi/mimo-v2.6-flash":{"id":"xiaomi/mimo-v2.6-flash","name":"MiMo V2.6 Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"stepfun/step-3.7-flash":{"id":"stepfun/step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"stepfun/step-5-preview":{"id":"stepfun/step-5-preview","name":"Step 5 Preview","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1,"output":2.7,"cache_read":0.1},"canonical_model_id":"stepfun/step-5-preview"},"reflection/beam-free":{"id":"reflection/beam-free","name":"Beam 1.0 (Free)","description":"Free provider route for experiments, demos, and cost-sensitive chat workloads","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"reflection/beam":{"id":"reflection/beam","name":"Beam 1.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"elorian/glimpse":{"id":"elorian/glimpse","name":"Glimpse","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":false,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":220000,"output":4096},"cost":{"input":0.6,"output":3,"cache_read":0.06}},"fireworks/ember-1":{"id":"fireworks/ember-1","name":"Ember-1","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"fireworks/ember-1"},"amazon/titan-embed-text-v2":{"id":"amazon/titan-embed-text-v2","type":"embedding","name":"Titan Text Embeddings V2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"titan-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-04-30","last_updated":"2024-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"amazon/nova-2-lite":{"id":"amazon/nova-2-lite","name":"Nova 2 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.3,"output":2.5,"cache_read":0.075},"canonical_model_id":"amazon/nova-2-lite"},"amazon/nova-micro":{"id":"amazon/nova-micro","name":"Nova Micro","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.035,"output":0.14,"cache_read":0.00875,"cache_write":0.035},"canonical_model_id":"amazon/nova-micro"},"amazon/nova-pro":{"id":"amazon/nova-pro","name":"Nova Pro","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.8,"output":3.2,"cache_read":0.2,"cache_write":0.8},"canonical_model_id":"amazon/nova-pro"},"amazon/nova-lite":{"id":"amazon/nova-lite","name":"Nova Lite","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.06,"output":0.24,"cache_read":0.015,"cache_write":0.06},"canonical_model_id":"amazon/nova-lite"},"perplexity/pplx-embed-v1-0.6b":{"id":"perplexity/pplx-embed-v1-0.6b","type":"embedding","name":"Embed v1 0.6b","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"v0","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"perplexity/pplx-embed-v1-4b":{"id":"perplexity/pplx-embed-v1-4b","type":"embedding","name":"Embed v1 4b","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"perplexity/sonar":{"id":"perplexity/sonar","name":"Sonar","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-02","release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":127000,"output":8000},"cost":{"input":0.25,"output":2.5,"cache_write":0.0625}},"spacexai/grok-imagine-image-2.0":{"id":"spacexai/grok-imagine-image-2.0","type":"image","name":"Grok Imagine Image 2.0","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-08-07","last_updated":"2026-08-07","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":64000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"xai/grok-imagine-image-2.0"},"spacexai/grok-4.5":{"id":"spacexai/grok-4.5","name":"Grok 4.5","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"spacexai/grok-4.6":{"id":"spacexai/grok-4.6","name":"Grok 4.6","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"spacexai/grok-imagine-image":{"id":"spacexai/grok-imagine-image","type":"image","name":"Grok Imagine Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-28","last_updated":"2026-01-28","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":16000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"xai/grok-imagine-image"},"spacexai/grok-4.7":{"id":"spacexai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"spacexai/grok-4.20-multi-agent-beta":{"id":"spacexai/grok-4.20-multi-agent-beta","name":"Grok 4.20 Multi Agent Beta","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"spacexai/grok-4.1-fast-reasoning":{"id":"spacexai/grok-4.1-fast-reasoning","name":"Grok 4.1 Fast Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast-reasoning"},"spacexai/grok-stt":{"id":"spacexai/grok-stt","type":"transcription","name":"Grok STT","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"spacexai/grok-4.20-multi-agent":{"id":"spacexai/grok-4.20-multi-agent","name":"Grok 4.20 Multi-Agent","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"release_date":"2026-03-10","last_updated":"2026-03-10","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"spacexai/grok-voice-think-fast-2.0":{"id":"spacexai/grok-voice-think-fast-2.0","type":"realtime","name":"Grok Voice Think Fast 2.0","description":"Realtime voice model for low-latency spoken conversations","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"spacexai/grok-4.20-reasoning":{"id":"spacexai/grok-4.20-reasoning","name":"Grok 4.20 Reasoning","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"spacexai/grok-4.20-non-reasoning":{"id":"spacexai/grok-4.20-non-reasoning","name":"Grok 4.20 Non-Reasoning","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"spacexai/grok-tts":{"id":"spacexai/grok-tts","type":"speech","name":"Grok TTS","description":"Convert text into spoken audio with a single API call. The API supports a rich set of expressive voices, inline speech tags for fine-grained delivery control, and output formats from high-fidelity MP3 to telephony-optimized μ-law.","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":15000,"output":15000},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"xai/grok-tts"},"spacexai/grok-4.1-fast-non-reasoning":{"id":"spacexai/grok-4.1-fast-non-reasoning","name":"Grok 4.1 Fast Non-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast"},"spacexai/grok-4.20-reasoning-beta":{"id":"spacexai/grok-4.20-reasoning-beta","name":"Grok 4.20 Beta Reasoning","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"spacexai/grok-imagine-video":{"id":"spacexai/grok-imagine-video","type":"video","name":"Grok Imagine","description":"Video model for text-to-video, image-to-video, video editing and extension with native synchronized audio","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-28","last_updated":"2026-01-28","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"xai/grok-imagine-video"},"spacexai/grok-imagine-video-1.5-lite":{"id":"spacexai/grok-imagine-video-1.5-lite","type":"video","name":"Grok Imagine Video 1.5 Lite","description":"Cost-efficient video model for text-to-video and image-to-video generation","family":"grok","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":1024,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"xai/grok-imagine-video-1.5-lite"},"spacexai/grok-4.20-non-reasoning-beta":{"id":"spacexai/grok-4.20-non-reasoning-beta","name":"Grok 4.20 Beta Non-Reasoning","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.4,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"spacexai/grok-4.3":{"id":"spacexai/grok-4.3","name":"Grok 4.3","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"spacexai/grok-imagine-video-1.5":{"id":"spacexai/grok-imagine-video-1.5","type":"video","name":"Grok Imagine Video 1.5","description":"Video model for text-to-video, image-to-video, and reference-to-video generation with optional voice references","family":"grok","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-05-30","last_updated":"2026-05-30","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":1024,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"xai/grok-imagine-video-1.5"},"spacexai/grok-build-0.1":{"id":"spacexai/grok-build-0.1","name":"Grok Build 0.1","description":"Grok coding model for agentic engineering, edits, and codebase workflows","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}},"canonical_model_id":"xai/grok-build-0.1"},"minimax/minimax-m2.1-lightning":{"id":"minimax/minimax-m2.1-lightning","name":"MiniMax M2.1 Lightning","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-12-23","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":2.4,"cache_read":0.03,"cache_write":0.375}},"minimax/minimax-m2.5-highspeed":{"id":"minimax/minimax-m2.5-highspeed","name":"MiniMax M2.5 High Speed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131000},"cost":{"input":0.6,"output":2.4,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5-highspeed"},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"Minimax M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-h3":{"id":"minimax/minimax-h3","type":"video","name":"MiniMax H3","description":"MiniMax general-purpose video model with multimodal references and synchronized stereo audio. Open H3-Base checkpoints; Context-IR and 2K regeneration are not included in the initial weights release.","family":"minimax","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-07-31","last_updated":"2026-10-04","modalities":{"input":["text","image"],"output":["video"]},"open_weights":true,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"minimax/MiniMax-H3"},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":205000,"output":196608},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131000},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5"},"minimax/minimax-m3.1-flash":{"id":"minimax/minimax-m3.1-flash","name":"MiniMax-M3.1-Flash","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0,"output":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"minimax/minimax-h3-max":{"id":"minimax/minimax-h3-max","type":"video","name":"MiniMax H3 Max","description":"Fast video model post-trained by fal.ai on MiniMax H3, with synchronized audio at 480p and 768p","family":"minimax","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"minimax/MiniMax-H3-Max"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax M2.7 High Speed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"inference-net/schematron-v2-small":{"id":"inference-net/schematron-v2-small","name":"Schematron V2 Small","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":0.05,"output":0.23,"cache_read":0.05},"canonical_model_id":"inference-net/schematron-v2-small"},"inference-net/schematron-v2-turbo":{"id":"inference-net/schematron-v2-turbo","name":"Schematron V2 Turbo","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.03,"output":0.15,"cache_read":0.03},"canonical_model_id":"inference-net/schematron-v2-turbo"},"scaledown/scaledown-extract-1":{"id":"scaledown/scaledown-extract-1","name":"Extract-1","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-10-20","last_updated":"2026-10-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":0.05,"output":0}},"scaledown/scaledown-summarize-1":{"id":"scaledown/scaledown-summarize-1","name":"Summarize-1","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"release_date":"2026-10-20","last_updated":"2026-10-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000},"cost":{"input":0.05,"output":0}},"microsoft/mai-voice-2.1":{"id":"microsoft/mai-voice-2.1","type":"speech","name":"MAI-Voice-2.1","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mai","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"microsoft/mai-voice-2":{"id":"microsoft/mai-voice-2","type":"speech","name":"MAI-Voice-2","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mai","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"microsoft/mai-voice-2.1-flash":{"id":"microsoft/mai-voice-2.1-flash","type":"speech","name":"MAI-Voice-2.1-Flash","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mai","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"microsoft/mai-transcribe-2":{"id":"microsoft/mai-transcribe-2","type":"transcription","name":"MAI-Transcribe 2","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"mai","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-03","last_updated":"2026-09-03","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"microsoft/mai-transcribe-2-streaming":{"id":"microsoft/mai-transcribe-2-streaming","type":"transcription","name":"MAI-Transcribe-2-Streaming","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"mai","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"microsoft/mai-image-2.6":{"id":"microsoft/mai-image-2.6","type":"image","name":"MAI-Image-2.6","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"mai","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":5,"output":38}},"microsoft/mai-image-2.6-flash":{"id":"microsoft/mai-image-2.6-flash","type":"image","name":"MAI-Image-2.6-Flash","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"mai","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":1.75,"output":19}},"microsoft/mai-voice-2-flash":{"id":"microsoft/mai-voice-2-flash","type":"speech","name":"MAI-Voice-2-Flash","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mai","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"alibaba/qwen3.7-max":{"id":"alibaba/qwen3.7-max","name":"Qwen 3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":64000},"cost":{"input":2.5,"output":7.5,"cache_read":0.5,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.7-max"},"alibaba/qwen3.7-plus":{"id":"alibaba/qwen3.7-plus","name":"Qwen 3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1,"max":262144}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.24,"cache_write":1.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.24,"cache_write":1.5}},"canonical_model_id":"alibaba/qwen3.7-plus"},"alibaba/qwen3.8-27b":{"id":"alibaba/qwen3.8-27b","name":"Qwen3.8 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.5,"output":3,"cache_read":0.1,"cache_write":0.625},"canonical_model_id":"alibaba/qwen3.8-27b"},"alibaba/qwen3-embedding-0.6b":{"id":"alibaba/qwen3-embedding-0.6b","type":"embedding","name":"Qwen3 Embedding 0.6B","description":"Compact multilingual text embedding model for retrieval, code search, classification, and clustering","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"alibaba/qwen3-embedding-0.6b"},"alibaba/qwen3.8-max-prime":{"id":"alibaba/qwen3.8-max-prime","name":"Qwen 3.8 Max Prime","description":"High-throughput edition of Qwen3.8 Max for coding, professional work, multimodal understanding, and long-running agent workflows","family":"qwen3.8-max","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":4,"output":12,"cache_read":0.5,"cache_write":5},"canonical_model_id":"alibaba/qwen3.8-max-prime"},"alibaba/qwen-3-32b":{"id":"alibaba/qwen-3-32b","name":"Qwen 3.32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-32b"},"alibaba/qwen-3-235b":{"id":"alibaba/qwen-3-235b","name":"Qwen3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.22,"output":0.88},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"alibaba/wan-v2.7-t2v":{"id":"alibaba/wan-v2.7-t2v","type":"video","name":"Wan v2.7 Text-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text","audio"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"alibaba/wan-v3.0-video":{"id":"alibaba/wan-v3.0-video","type":"video","name":"Wan v3.0 Video","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-08-23","last_updated":"2026-08-23","modalities":{"input":["text","image","video","audio"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"alibaba/qwen3.8-2.4t-a95b":{"id":"alibaba/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"alibaba/qwen3-coder-next":{"id":"alibaba/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.5,"output":1.2},"canonical_model_id":"alibaba/qwen3-coder-next"},"alibaba/qwen3.5-flash":{"id":"alibaba/qwen3.5-flash","name":"Qwen 3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1,"max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"alibaba/qwen3.5-flash"},"alibaba/wan-v3.0-video-prime":{"id":"alibaba/wan-v3.0-video-prime","type":"video","name":"Wan v3.0 Video Prime","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text","image","video","audio"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"alibaba/qwen3.8-max-0902":{"id":"alibaba/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"alibaba/qwen3-embedding-8b":{"id":"alibaba/qwen3-embedding-8b","type":"embedding","name":"Qwen3 Embedding 8B","description":"Multilingual text embedding model for retrieval, code search, classification, and clustering","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"alibaba/qwen3-embedding-8b"},"alibaba/wan-v2.7-r2v":{"id":"alibaba/wan-v2.7-r2v","type":"video","name":"Wan v2.7 Reference-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text","image","video","audio"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"alibaba/wan-v2.6-t2v":{"id":"alibaba/wan-v2.6-t2v","type":"video","name":"Wan v2.6 Text-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text","audio"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"alibaba/qwen3.7-flash":{"id":"alibaba/qwen3.7-flash","name":"Qwen 3.7 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":64000},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.038,"tiers":[{"input":0.1,"output":0.4,"cache_read":0.02,"cache_write":0.125,"tier":{"type":"context","size":32000}},{"input":0.2,"output":0.8,"cache_read":0.04,"cache_write":0.25,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"alibaba/qwen3.8-max":{"id":"alibaba/qwen3.8-max","name":"Qwen 3.8 Max","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"temperature":true,"release_date":"2026-07-19","last_updated":"2026-07-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"alibaba/qwen3-embedding-4b":{"id":"alibaba/qwen3-embedding-4b","type":"embedding","name":"Qwen3 Embedding 4B","description":"Multilingual text embedding model for retrieval, code search, classification, and clustering","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"alibaba/qwen3-embedding-4b"},"alibaba/qwen3-next-80b-a3b-thinking":{"id":"alibaba/qwen3-next-80b-a3b-thinking","name":"Qwen3 Next 80B A3B Thinking","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"alibaba/qwen3.8-omni-flash":{"id":"alibaba/qwen3.8-omni-flash","name":"Qwen 3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"alibaba/qwen3.8-flash":{"id":"alibaba/qwen3.8-flash","name":"Qwen 3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":128000},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"alibaba/wan-v2.6-i2v-flash":{"id":"alibaba/wan-v2.6-i2v-flash","type":"video","name":"Wan v2.6 Image-to-Video Flash","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text","image","audio"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"alibaba/wan-v2.6-r2v":{"id":"alibaba/wan-v2.6-r2v","type":"video","name":"Wan v2.6 Reference-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"alibaba/qwen3-next-80b-a3b-instruct":{"id":"alibaba/qwen3-next-80b-a3b-instruct","name":"Qwen3 Next 80B A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262114,"output":262114},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"alibaba/qwen-3-30b":{"id":"alibaba/qwen-3-30b","name":"Qwen3-30B-A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.5},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"alibaba/qwen3-coder-30b-a3b":{"id":"alibaba/qwen3-coder-30b-a3b","name":"Qwen 3 Coder 30B A3B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"alibaba/qwen3.5-plus":{"id":"alibaba/qwen3.5-plus","name":"Qwen 3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1,"max":81920}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":2.4,"cache_read":0.04,"cache_write":0.5,"tiers":[{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":256001}}],"context_over_200k":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625}},"canonical_model_id":"alibaba/qwen3.5-plus"},"alibaba/qwen3.6-plus":{"id":"alibaba/qwen3.6-plus","name":"Qwen 3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1,"max":131072}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625,"tiers":[{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5}},"canonical_model_id":"alibaba/qwen3.6-plus"},"alibaba/qwen3-vl-instruct":{"id":"alibaba/qwen3-vl-instruct","name":"Qwen3 VL Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.2,"output":0.88,"cache_read":0.11}},"alibaba/wan-v2.6-i2v":{"id":"alibaba/wan-v2.6-i2v","type":"video","name":"Wan v2.6 Image-to-Video","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text","image","audio"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"alibaba/qwen-3-14b":{"id":"alibaba/qwen-3-14b","name":"Qwen3-14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.24}},"alibaba/wan-v2.6-r2v-flash":{"id":"alibaba/wan-v2.6-r2v-flash","type":"video","name":"Wan v2.6 Reference-to-Video Flash","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"alibaba/qwen3-max-thinking":{"id":"alibaba/qwen3-max-thinking","name":"Qwen 3 Max Thinking","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1,"max":81920}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-23","last_updated":"2025-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":1.2,"output":6,"cache_read":0.24,"tiers":[{"input":2.4,"output":12,"cache_read":0.48,"tier":{"type":"context","size":32001}},{"input":3,"output":15,"cache_read":0.6,"tier":{"type":"context","size":128001}}]}},"alibaba/wan-v2.5-t2v-preview":{"id":"alibaba/wan-v2.5-t2v-preview","type":"video","name":"Wan v2.5 Text-to-Video Preview","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-09-24","last_updated":"2025-09-24","modalities":{"input":["text","audio"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"alibaba/qwen3-coder":{"id":"alibaba/qwen3-coder","name":"Qwen3 Coder 480B A35B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.22,"output":1.8,"cache_read":0.022,"tiers":[{"input":2.7,"output":13.5,"cache_read":0.54,"tier":{"type":"context","size":32001}},{"input":4.5,"output":22.5,"cache_read":0.9,"tier":{"type":"context","size":128001}}]},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"alibaba/qwen3.6-27b":{"id":"alibaba/qwen3.6-27b","name":"Qwen 3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":131072}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.6-27b"},"nvidia/nemotron-3.5-lightning":{"id":"nvidia/nemotron-3.5-lightning","name":"Nemotron 3.5 Lightning 30B","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.05,"output":0.2,"cache_read":0.01},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-nano-9b-v2":{"id":"nvidia/nemotron-nano-9b-v2","name":"Nvidia Nemotron Nano 9B V2","description":"Compact Nemotron model for efficient reasoning and deployable AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2025-08-18","last_updated":"2025-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.06,"output":0.23},"canonical_model_id":"nvidia/nemotron-nano-9b-v2"},"nvidia/nemotron-3-super-120b-a12b":{"id":"nvidia/nemotron-3-super-120b-a12b","name":"NVIDIA Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"cost":{"input":0.15,"output":0.65},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/nemotron-3-ultra-550b-a55b":{"id":"nvidia/nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65000},"cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/nemotron-3.5-lightning-free":{"id":"nvidia/nemotron-3.5-lightning-free","name":"Nemotron 3.5 Lightning Free","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65000},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-3-nano-30b-a3b":{"id":"nvidia/nemotron-3-nano-30b-a3b","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.05,"output":0.2,"cache_read":0.025},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/nemotron-3.5-super-vl":{"id":"nvidia/nemotron-3.5-super-vl","name":"NVIDIA Nemotron 3.5 Super VL","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.2,"output":0.5,"cache_read":0.04},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/nemotron-nano-12b-v2-vl":{"id":"nvidia/nemotron-nano-12b-v2-vl","name":"Nvidia Nemotron Nano 12B V2 VL","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-10-28","last_updated":"2025-10-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.2,"output":0.6},"canonical_model_id":"nvidia/nemotron-nano-12b-v2-vl"},"heygen/heygen-video-1":{"id":"heygen/heygen-video-1","type":"video","name":"HeyGen Video","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-09","last_updated":"2026-10-09","modalities":{"input":["text","image","video","audio"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bytedance/seedance-2.0-fast":{"id":"bytedance/seedance-2.0-fast","type":"video","name":"Seedance 2.0 Fast","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"seed","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-14","last_updated":"2026-04-14","modalities":{"input":["text","image","video","audio"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bytedance/seedance-2.5":{"id":"bytedance/seedance-2.5","type":"video","name":"Seedance 2.5","description":"ByteDance multimodal audio-video model with up to 30-second single-pass generation, multi-round extension, reference inputs and timestamp-level editing.","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-07-31","last_updated":"2026-10-04","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"bytedance-seed/seedance-2.5"},"bytedance/seedance-v1.5-pro":{"id":"bytedance/seedance-v1.5-pro","type":"video","name":"Seedance v1.5 Pro","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"seed","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-16","last_updated":"2025-12-16","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bytedance/seedance-v1.0-pro":{"id":"bytedance/seedance-v1.0-pro","type":"video","name":"Seedance v1.0 Pro","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"seed","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-11","last_updated":"2025-06-11","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bytedance/seedance-v1.0-pro-fast":{"id":"bytedance/seedance-v1.0-pro-fast","type":"video","name":"Seedance v1.0 Pro Fast","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"seed","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-24","last_updated":"2025-10-31","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bytedance/seed-2.1-turbo":{"id":"bytedance/seed-2.1-turbo","name":"Seed 2.1 Turbo","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":2.5,"cache_read":0.1},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"bytedance/seedance-2.0":{"id":"bytedance/seedance-2.0","type":"video","name":"Seedance 2.0","description":"ByteDance unified multimodal audio-video model supporting text, image, video and audio references, editing, extension and synchronized stereo sound.","family":"seed","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-12","last_updated":"2026-10-04","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"bytedance-seed/seedance-2.0"},"bytedance/seed-1.8":{"id":"bytedance/seed-1.8","name":"Seed 1.8","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-28","last_updated":"2025-12-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32768},"cost":{"input":0.25,"output":2,"cache_read":0.05,"tiers":[{"input":0.5,"output":4,"cache_read":0.05,"tier":{"type":"context","size":128001}}]},"canonical_model_id":"bytedance-seed/seed-1-8"},"bytedance/seedream-4.5":{"id":"bytedance/seedream-4.5","type":"image","name":"Seedream 4.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-03","last_updated":"2025-11-28","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bytedance/seedance-2.0-mini":{"id":"bytedance/seedance-2.0-mini","type":"video","name":"Seedance 2.0 Mini","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"seed","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-06-22","last_updated":"2026-06-22","modalities":{"input":["text","image","video","audio"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bytedance/seedream-5.0-pro":{"id":"bytedance/seedream-5.0-pro","type":"image","name":"Seedream 5.0 Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-07-11","last_updated":"2026-07-11","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bytedance/seed-1.6":{"id":"bytedance/seed-1.6","name":"Seed 1.6","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.25,"output":2,"cache_read":0.05,"tiers":[{"input":0.5,"output":4,"cache_read":0.05,"tier":{"type":"context","size":128001}}]},"canonical_model_id":"bytedance-seed/seed-1-6"},"bytedance/seedream-5.0-lite":{"id":"bytedance/seedream-5.0-lite","type":"image","name":"Seedream 5.0 Lite","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-13","last_updated":"2026-01-28","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"bytedance/seedream-4.0":{"id":"bytedance/seedream-4.0","type":"image","name":"Seedream 4.0","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-09-09","last_updated":"2025-08-28","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"zai/glm-4.5-air":{"id":"zai/glm-4.5-air","name":"GLM 4.5 Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":96000},"cost":{"input":0.2,"output":1.1,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-4.5-air"},"zai/glm-4.7-flashx":{"id":"zai/glm-4.7-flashx","name":"GLM 4.7 FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.06,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"zai/glm-4.5v":{"id":"zai/glm-4.5v","name":"GLM 4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":66000,"output":16000},"cost":{"input":0.6,"output":1.8,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5v"},"zai/glm-5.1":{"id":"zai/glm-5.1","name":"GLM 5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":64000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai/glm-4.7":{"id":"zai/glm-4.7","name":"GLM 4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":120000},"cost":{"input":0.6,"output":2.2},"canonical_model_id":"zhipuai/glm-4.7"},"zai/glm-5.3-fast":{"id":"zai/glm-5.3-fast","name":"GLM 5.3 Fast","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":2.1,"output":6.6,"cache_read":0.21},"canonical_model_id":"zhipuai/glm-5.3"},"zai/glm-5-turbo":{"id":"zai/glm-5-turbo","name":"GLM 5 Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202800,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"zai/glm-4.5":{"id":"zai/glm-4.5","name":"GLM 4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":96000},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5"},"zai/glm-5.3-flash":{"id":"zai/glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131000},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai/glm-5.2":{"id":"zai/glm-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.8,"output":2.55,"cache_read":0.16},"canonical_model_id":"zhipuai/glm-5.2"},"zai/glm-4.7-flash":{"id":"zai/glm-4.7-flash","name":"GLM 4.7 Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131000},"cost":{"input":0.07,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"zai/glm-5.3-flashx":{"id":"zai/glm-5.3-flashx","name":"GLM 5.3 FlashX","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.37,"output":1.25,"cache_read":0.075},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai/glm-5":{"id":"zai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":131100},"cost":{"input":1,"output":3.2},"canonical_model_id":"zhipuai/glm-5"},"zai/glm-4.6":{"id":"zai/glm-4.6","name":"GLM 4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":96000},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"zai/glm-5.3":{"id":"zai/glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.3"},"zai/glm-5v-turbo":{"id":"zai/glm-5v-turbo","name":"GLM 5V Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"zai/glm-5.2-fast":{"id":"zai/glm-5.2-fast","name":"GLM 5.2 Fast","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":2.8,"output":8.8,"cache_read":0.56},"canonical_model_id":"zhipuai/glm-5.2"},"meituan/longcat-2.5-preview":{"id":"meituan/longcat-2.5-preview","name":"LongCat 2.5 Preview","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"longcat","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":false,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"meituan/longcat-2.5-preview"},"interfaze/interfaze-beta":{"id":"interfaze/interfaze-beta","name":"Interfaze Beta","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"temperature":true,"release_date":"2025-10-07","last_updated":"2026-04-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":1.5,"output":3.5}},"cohere/rerank-v4-fast":{"id":"cohere/rerank-v4-fast","type":"reranking","name":"Cohere Rerank 4 Fast","description":"Low-latency multilingual reranker for high-throughput retrieval","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"cohere/rerank-v4.0-fast"},"cohere/rerank-v5-pro":{"id":"cohere/rerank-v5-pro","type":"reranking","name":"Cohere Rerank 5 Pro","description":"Reranking model for improving retrieval quality in search and recommendation systems","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-21","last_updated":"2026-10-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"cohere/embed-v5.0-pro":{"id":"cohere/embed-v5.0-pro","type":"embedding","name":"Embed 5 Pro","description":"Cohere's most capable multimodal embedding model for text, images, and mixed documents","family":"cohere-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-30","last_updated":"2026-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":2048},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"cohere/embed-v5.0-pro"},"cohere/embed-v4.0":{"id":"cohere/embed-v4.0","type":"embedding","name":"Embed v4.0","description":"Multimodal embedding model for text, images, and mixed text-and-image documents","family":"cohere-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":1536},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"cohere/embed-v4.0"},"cohere/rerank-v4-pro":{"id":"cohere/rerank-v4-pro","type":"reranking","name":"Cohere Rerank 4 Pro","description":"Multilingual reranker for maximum relevance on documents and semi-structured data","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"cohere/rerank-v4.0-pro"},"cohere/rerank-v5-fast":{"id":"cohere/rerank-v5-fast","type":"reranking","name":"Cohere Rerank 5 Fast","description":"Reranking model for improving retrieval quality in search and recommendation systems","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-21","last_updated":"2026-10-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"cohere/command-a":{"id":"cohere/command-a","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a"},"cohere/rerank-v3.5":{"id":"cohere/rerank-v3.5","type":"reranking","name":"Cohere Rerank 3.5","description":"Multilingual reranker for documents and semi-structured data","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-12-02","last_updated":"2024-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4096,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"cohere/rerank-v3.5"},"cohere/embed-v5.0-fast":{"id":"cohere/embed-v5.0-fast","type":"embedding","name":"Embed 5 Fast","description":"Faster, lighter Embed 5 model sharing Embed 5 Pro's embedding space","family":"cohere-embed","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-30","last_updated":"2026-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":2048},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"cohere/embed-v5.0-fast"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-omni-flash-preview":{"id":"google/gemini-omni-flash-preview","type":"video","name":"Gemini Omni Flash Preview","description":"Omni-modal model for text, vision, audio, and multimodal agent tasks","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":false,"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":57920},"cost":{"input":1.5,"output":9},"canonical_model_id":"google/gemini-omni-flash-preview"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65000},"cost":{"input":0.25,"output":1.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.5-transcribe-live":{"id":"google/gemini-3.5-transcribe-live","type":"transcription","name":"Gemini 3.5 Transcribe Live","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"google/gemini-3.5-transcribe-live"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Gemini multimodal model for text, image, audio, video, and document tasks","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"google/gemini-4-argon"},"google/gemini-3.8-live":{"id":"google/gemini-3.8-live","type":"realtime","name":"Gemini 3.8 Live","description":"Low-latency Live API model for real-time voice agents with interleaved reasoning and async tools","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-09-15","last_updated":"2026-09-15","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":131072,"output":65536},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":0.75,"output":4.5,"input_audio":3,"output_audio":12},"canonical_model_id":"google/gemini-3.8-live"},"google/gemini-3.8-flash-tts":{"id":"google/gemini-3.8-flash-tts","type":"speech","name":"Gemini 3.8 Flash TTS","description":"Flagship text-to-speech model for expressive, studio-grade voices and long-form multi-speaker audio","family":"gemini-flash","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":16384},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":0.5,"output":9,"output_audio":9},"canonical_model_id":"google/gemini-3.8-flash-tts"},"google/gemini-3-flash":{"id":"google/gemini-3-flash","name":"Gemini 3 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65000},"cost":{"input":0.5,"output":3,"cache_read":0.05}},"google/gemini-3.1-flash-lite-image":{"id":"google/gemini-3.1-flash-lite-image","type":"image","name":"Gemini 3.1 Flash Lite Image (Nano Banana 2 Lite)","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":4096},"cost":{"input":0.25,"output":1.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"google/gemini-2.5-flash-image":{"id":"google/gemini-2.5-flash-image","type":"image","name":"Nano Banana (Gemini 2.5 Flash Image)","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":65535},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-2.5-flash-image"},"google/gemini-embedding-001":{"id":"google/gemini-embedding-001","type":"embedding","name":"Gemini Embedding 001","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-05","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2048,"output":1},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"google/gemini-embedding-001"},"google/text-embedding-005":{"id":"google/text-embedding-005","type":"embedding","name":"Text Embedding 005","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-08-01","last_updated":"2024-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65000},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-3.1-flash-image-preview":{"id":"google/gemini-3.1-flash-image-preview","type":"image","name":"Gemini 3.1 Flash Image Preview (Nano Banana 2)","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3.1-flash-image-preview"},"google/gemini-3.5-transcribe":{"id":"google/gemini-3.5-transcribe","type":"transcription","name":"Gemini 3.5 Transcribe","description":"Speech-to-text model with language detection, speaker diarization, word timestamps, and custom vocabulary","family":"gemini","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":98304,"output":32768},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":2,"output":12,"input_audio":2},"canonical_model_id":"google/gemini-3.5-transcribe"},"google/gemini-nano-banana-2.1":{"id":"google/gemini-nano-banana-2.1","type":"image","name":"Gemini Nano Banana 2.1","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"google/gemini-nano-banana-2.1"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200001}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"google/veo-3.1-lite-generate-001":{"id":"google/veo-3.1-lite-generate-001","type":"video","name":"Veo 3.1 Lite Generate","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"google/gemini-3.8-flash-lite-tts":{"id":"google/gemini-3.8-flash-lite-tts","type":"speech","name":"Gemini 3.8 Flash-Lite TTS","description":"Fast, cost-efficient text-to-speech model for high-volume narration and voice agent pipelines","family":"gemini-flash-lite","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":16384},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":0.5,"output":6,"output_audio":6},"canonical_model_id":"google/gemini-3.8-flash-lite-tts"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-3.1-flash-image":{"id":"google/gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3.1-flash-image"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"google/text-multilingual-embedding-002":{"id":"google/text-multilingual-embedding-002","type":"embedding","name":"Text Multilingual Embedding 002","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-03-01","last_updated":"2024-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/veo-3.1-fast-generate-001":{"id":"google/veo-3.1-fast-generate-001","type":"video","name":"Veo 3.1 Fast Generate","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-15","last_updated":"2026-06-08","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"google/gemini-3-pro-image":{"id":"google/gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"google/gemini-3-pro-image"},"google/gemini-3.8-live-extended-thinking":{"id":"google/gemini-3.8-live-extended-thinking","type":"realtime","name":"Gemini 3.8 Live Extended Thinking","description":"Live API voice model with configurable background reasoning for multi-step real-time agents","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-09-15","last_updated":"2026-09-15","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":131072,"output":65536},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"cost":{"input":0.75,"output":4.5,"input_audio":3,"output_audio":12},"canonical_model_id":"google/gemini-3.8-live-extended-thinking"},"google/veo-3.0-fast-generate-001":{"id":"google/veo-3.0-fast-generate-001","type":"video","name":"Veo 3.0 Fast Generate","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-07-31","last_updated":"2026-06-08","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"google/gemini-embedding-2":{"id":"google/gemini-embedding-2","type":"embedding","name":"Gemini Embedding 2","description":"Multimodal embedding model mapping text, images, video, audio, and PDFs into a unified embedding space","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-11","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":3072},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"},"canonical_model_id":"google/gemini-embedding-2"},"google/veo-3.1-generate-001":{"id":"google/veo-3.1-generate-001","type":"video","name":"Veo 3.1","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-15","last_updated":"2026-06-08","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"google/veo-3.0-generate-001":{"id":"google/veo-3.0-generate-001","type":"video","name":"Veo 3.0","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-05-20","last_updated":"2026-06-08","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"provider":{"api":"https://ai-gateway.vercel.sh/v4/ai"}},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"arcee-ai/trinity-large-thinking":{"id":"arcee-ai/trinity-large-thinking","name":"Trinity Large Thinking","description":"Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use","family":"trinity","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262100,"output":80000},"cost":{"input":0.25,"output":0.8,"cache_read":0.06},"canonical_model_id":"arcee-ai/trinity-large-thinking"}}},"runinfra":{"id":"runinfra","env":["RUNINFRA_GATEWAY_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.runinfra.ai/v1","name":"RunInfra","doc":"https://runinfra.ai/docs","models":{"ornith-ai/Ornith-1.5-35B-A3B":{"id":"ornith-ai/Ornith-1.5-35B-A3B","name":"Ornith 1.5 35B A3B","description":"Mixture-of-experts coding-reasoning model for agentic software tasks, tool use, and image understanding","family":"ornith","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-18","last_updated":"2026-08-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"deepreinforce/ornith-1.5-35b-a3b"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":1.3,"output":4.2,"cache_read":0.13},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.11,"output":0.45,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16":{"id":"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16","name":"Nemotron 3.5 Lightning 30B A3B","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.05,"output":0.15,"cache_read":0.01},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"status":"deprecated","cost":{"input":0.13,"output":0.27,"cache_read":0.01},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"status":"deprecated","cost":{"input":0.6,"output":1.9,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"alibaba/qwen3.8-27b"},"Inferact/Qwen3.8-2.4T-A95B-NVFP4":{"id":"Inferact/Qwen3.8-2.4T-A95B-NVFP4","name":"Qwen3.8 2.4T A95B (NVFP4)","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"}}},"hyper":{"id":"hyper","env":["HYPER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://hyper.charm.land/v1","name":"Charm Hyper","doc":"https://hyper.charm.land","models":{"kimi-k2-thinking":{"id":"kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":26214},"cost":{"input":0.6,"output":2.5,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2.5,"output":7.5,"cache_read":0.5},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":1.2,"output":4.8,"cache_read":0.24},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.5,"output":3,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3.8-27b"},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262100,"output":6553},"cost":{"input":0.484,"output":1.852,"cache_read":0.242},"canonical_model_id":"minimax/MiniMax-M2.7"},"prism":{"id":"prism","name":"Prism","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":true,"release_date":"1970-01-01","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":16000},"cost":{"input":3.2664,"output":16.332,"cache_read":0.32664}},"qwen3.8-2.4t-a95b":{"id":"qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":262144},"cost":{"input":0.33,"output":1.31,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":262144},"cost":{"input":1.437216,"output":4.311648,"cache_read":0.047907},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":13107},"cost":{"input":0.178,"output":0.68,"cache_read":0.089},"canonical_model_id":"openai/gpt-oss-120b"},"qwen3.7-flash":{"id":"qwen3.7-flash","name":"Qwen3.7 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":64000},"cost":{"input":0.2,"output":0.8,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max Preview","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-19","last_updated":"2026-07-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16000},"cost":{"input":1.03436,"output":4.3552,"cache_read":0.206872},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":16000},"cost":{"input":3.2664,"output":16.332,"cache_read":0.32664},"canonical_model_id":"moonshotai/kimi-k3"},"gemma-4-26b-a4b-it":{"id":"gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":25600},"cost":{"input":0.098,"output":0.334,"cache_read":0.049},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-flash"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.16332,"output":0.5444,"cache_read":0.031575},"canonical_model_id":"zhipuai/glm-5.3-flash"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":512000},"cost":{"input":0.32664,"output":1.30656,"cache_read":0.064239},"canonical_model_id":"minimax/MiniMax-M3"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":1.52432,"output":4.79072,"cache_read":0.152432},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":262144},"cost":{"input":0.33,"output":1.31,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.44,"output":1.32,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.52432,"output":4.79072,"cache_read":0.283088},"canonical_model_id":"zhipuai/glm-5.3"},"inkling":{"id":"inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":1.0888,"output":4.40964,"cache_read":0.185096},"canonical_model_id":"thinkingmachines/inkling"}}},"stepfun":{"id":"stepfun","env":["STEPFUN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.stepfun.com/v1","name":"StepFun (China)","doc":"https://platform.stepfun.com/docs/zh/overview/concept","models":{"step-tts-2":{"id":"step-tts-2","type":"speech","name":"Step TTS 2","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"step","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-01","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"step-2-16k":{"id":"step-2-16k","name":"Step 2 (16K)","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-01-01","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"input":16384,"output":8192},"status":"deprecated","cost":{"input":5.21,"output":16.44,"cache_read":1.04}},"step-3.7-flash":{"id":"step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.20175,"output":1.21052,"cache_read":0.04035},"canonical_model_id":"stepfun/step-3.7-flash"},"stepaudio-2.5-asr":{"id":"stepaudio-2.5-asr","type":"transcription","name":"StepAudio 2.5 ASR","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"step","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-24","last_updated":"2026-07-02","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0}},"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":1.04613,"output":2.98895,"cache_read":0.10461},"canonical_model_id":"stepfun/step-5-preview"},"step-3.5-flash":{"id":"step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.10461,"output":0.31384,"cache_read":0.02092},"canonical_model_id":"stepfun/step-3.5-flash"},"step-3.5-flash-2603":{"id":"step-3.5-flash-2603","name":"Step 3.5 Flash 2603","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.10461,"output":0.31384,"cache_read":0.02092},"canonical_model_id":"stepfun/step-3.5-flash-2603"},"stepaudio-2.5-tts":{"id":"stepaudio-2.5-tts","type":"speech","name":"StepAudio 2.5 TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"step","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-16","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"step-1-32k":{"id":"step-1-32k","name":"Step 1 (32K)","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-01-01","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":32768},"status":"deprecated","cost":{"input":2.05,"output":9.59,"cache_read":0.41}}}},"alibaba-coding-plan-cn":{"id":"alibaba-coding-plan-cn","env":["ALIBABA_CODING_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://coding.dashscope.aliyuncs.com/v1","name":"Alibaba Coding Plan (China)","doc":"https://help.aliyun.com/zh/model-studio/coding-plan","models":{"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen3-coder-plus":{"id":"qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3-coder-next":{"id":"qwen3-coder-next","name":"Qwen3 Coder Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3.5-plus":{"id":"qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3-max-2026-01-23":{"id":"qwen3-max-2026-01-23","name":"Qwen3 Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-23","last_updated":"2026-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"glm-5":{"id":"glm-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":16384},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":24576},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}}}},"ebcloud":{"id":"ebcloud","env":["EBCLOUD_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://maas-api.ebcloud.com/v1","name":"EBCloud","doc":"https://docs.ebtech.com/ai/model-api.html","models":{"DeepSeek-V4-Pro":{"id":"DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.4286,"output":0.8571},"canonical_model_id":"deepseek/deepseek-v4-pro"},"DeepSeek-V4-Flash":{"id":"DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.143,"output":0.2857},"canonical_model_id":"deepseek/deepseek-v4-flash"},"Kimi-K2.6":{"id":"Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.9286,"output":3.8571},"canonical_model_id":"moonshotai/kimi-k2.6"},"GLM-5.1":{"id":"GLM-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.8571,"output":3.4286},"canonical_model_id":"zhipuai/glm-5.1"}}},"reflection":{"id":"reflection","env":["REFLECTION_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.reflection.ai/openai/v1","name":"Reflection","doc":"https://developers.reflection.ai/models","models":{"Beam-1.0-501B-A23B":{"id":"Beam-1.0-501B-A23B","name":"Beam 1.0 501B A23B","description":"Sparse mixture-of-experts reasoning model with 23B active parameters for coding and agentic workflows","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"reflection/beam-1.0-501b-a23b"}}},"above":{"id":"above","env":["ABOVE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.above.dev/v1","name":"above.dev","doc":"https://above.dev/docs","models":{"mimo-v2.6-pro-ultraspeed":{"id":"mimo-v2.6-pro-ultraspeed","name":"MiMo V2.6 Pro UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":5.0769,"output":10.1538,"cache_read":0.0423},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0.165,"output":0.66,"reasoning":0.66,"cache_read":0.0033},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo V2.6 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.5077,"output":1.0154,"cache_read":0.0042},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo V2.6 Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1692,"output":0.3385,"cache_read":0.0034},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.726,"output":2.178,"reasoning":2.178,"cache_read":0.0242},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen 3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":2.2,"output":6.6,"cache_read":0.275},"canonical_model_id":"alibaba/qwen3.8-max"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.165,"output":0.55,"cache_read":0.0319},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5.2":{"id":"glm-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.54,"output":4.84,"cache_read":0.154},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.165,"output":0.66,"reasoning":0.66,"cache_read":0.0033},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"glm-5.2-fast":{"id":"glm-5.2-fast","name":"GLM 5.2 Fast","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2.31,"output":7.26,"cache_read":0.231},"canonical_model_id":"zhipuai/glm-5.2"}}},"claudinio":{"id":"claudinio","env":["CLAUDINIO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.claudin.io/v1","name":"Claudinio","doc":"https://claudin.io","models":{"claudius":{"id":"claudius","name":"Claudius","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"knowledge":"2026-05","release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":3,"output":8,"cache_read":0.9}},"claudinio":{"id":"claudinio","name":"Claudinio","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"knowledge":"2026-05","release_date":"2026-05-12","last_updated":"2026-06-02","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":0.5,"output":2,"cache_read":0.15}}}},"stackit":{"id":"stackit","env":["STACKIT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.openai-compat.model-serving.eu01.onstackit.cloud/v1","name":"STACKIT","doc":"https://docs.stackit.cloud/products/data-and-ai/ai-model-serving/basics/available-shared-models","models":{"intfloat/e5-mistral-7b-instruct":{"id":"intfloat/e5-mistral-7b-instruct","type":"embedding","name":"E5 Mistral 7B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2023-12-11","last_updated":"2023-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":0.02243,"output":0}},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.18,"output":0.29},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":8192},"cost":{"input":0.53,"output":0.76},"canonical_model_id":"openai/gpt-oss-120b"},"cortecs/Llama-3.3-70B-Instruct-FP8-Dynamic":{"id":"cortecs/Llama-3.3-70B-Instruct-FP8-Dynamic","name":"Llama 3.3 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.53,"output":0.76},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"Qwen/Qwen3.6-27B":{"id":"Qwen/Qwen3.6-27B","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"status":"deprecated","cost":{"input":0.53,"output":0.76},"canonical_model_id":"alibaba/qwen3.6-27b"},"Qwen/Qwen3-VL-Embedding-8B":{"id":"Qwen/Qwen3-VL-Embedding-8B","type":"embedding","name":"Qwen3-VL Embedding 8B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.09,"output":0}},"Qwen/Qwen3-VL-235B-A22B-Instruct-FP8":{"id":"Qwen/Qwen3-VL-235B-A22B-Instruct-FP8","name":"Qwen3-VL 235B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":218000,"output":16384},"cost":{"input":1.76,"output":2.05},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":37000,"output":4096},"status":"deprecated","cost":{"input":0.53,"output":0.76},"canonical_model_id":"google/gemma-3-27b-it"}}},"gitlab":{"id":"gitlab","env":["GITLAB_TOKEN"],"npm":"gitlab-ai-provider","name":"GitLab Duo","doc":"https://docs.gitlab.com/user/duo_agent_platform/","models":{"duo-chat-gpt-5-codex":{"id":"duo-chat-gpt-5-codex","name":"Agentic Chat (GPT-5 Codex)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5-codex"},"duo-chat-gpt-5-6-sol":{"id":"duo-chat-gpt-5-6-sol","name":"Agentic Chat (GPT-5.6 Sol)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-5.6-sol"},"duo-chat-sonnet-5-5":{"id":"duo-chat-sonnet-5-5","name":"Agentic Chat (Claude Sonnet 5.5)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"duo-chat-gpt-5-1":{"id":"duo-chat-gpt-5-1","name":"Agentic Chat (GPT-5.1)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.1"},"duo-chat-opus-4-5":{"id":"duo-chat-opus-4-5","name":"Agentic Chat (Claude Opus 4.5)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"duo-chat-gpt-5-4-nano":{"id":"duo-chat-gpt-5-4-nano","name":"Agentic Chat (GPT-5.4 Nano)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.4-nano"},"duo-chat-opus-4-7":{"id":"duo-chat-opus-4-7","name":"Agentic Chat (Claude Opus 4.7)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-opus-4-7"},"duo-chat-sonnet-4-5":{"id":"duo-chat-sonnet-4-5","name":"Agentic Chat (Claude Sonnet 4.5)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"duo-chat-gpt-6-astra":{"id":"duo-chat-gpt-6-astra","name":"Agentic Chat (GPT-6 Astra)","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-6-astra"},"duo-chat-haiku-4-5":{"id":"duo-chat-haiku-4-5","name":"Agentic Chat (Claude Haiku 4.5)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"duo-chat-fable-5":{"id":"duo-chat-fable-5","name":"Agentic Chat (Claude Fable 5)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-fable-5"},"duo-chat-sonnet-4-6":{"id":"duo-chat-sonnet-4-6","name":"Agentic Chat (Claude Sonnet 4.6)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"duo-chat-gpt-5-5":{"id":"duo-chat-gpt-5-5","name":"Agentic Chat (GPT-5.5)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.5"},"duo-chat-gpt-5-4":{"id":"duo-chat-gpt-5-4","name":"Agentic Chat (GPT-5.4)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.4"},"duo-chat-gpt-5-6-terra":{"id":"duo-chat-gpt-5-6-terra","name":"Agentic Chat (GPT-5.6 Terra)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-5.6-terra"},"duo-chat-gpt-6-sol":{"id":"duo-chat-gpt-6-sol","name":"Agentic Chat (GPT-6 Sol)","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-6-sol"},"duo-chat-sonnet-5":{"id":"duo-chat-sonnet-5","name":"Agentic Chat (Claude Sonnet 5)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-sonnet-5"},"duo-chat-gpt-5-6-luna":{"id":"duo-chat-gpt-5-6-luna","name":"Agentic Chat (GPT-5.6 Luna)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-5.6-luna"},"duo-chat-opus-5":{"id":"duo-chat-opus-5","name":"Agentic Chat (Claude Opus 5)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-opus-5"},"duo-chat-gpt-5-mini":{"id":"duo-chat-gpt-5-mini","name":"Agentic Chat (GPT-5 Mini)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5-mini"},"duo-chat-gpt-5-2-codex":{"id":"duo-chat-gpt-5-2-codex","name":"Agentic Chat (GPT-5.2 Codex)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.2-codex"},"duo-chat-gpt-6-1-sol":{"id":"duo-chat-gpt-6-1-sol","name":"Agentic Chat (GPT-6.1 Sol)","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-6.1-sol"},"duo-chat-opus-5-5":{"id":"duo-chat-opus-5-5","name":"Agentic Chat (Claude Opus 5.5)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-opus-5-5"},"duo-chat-fable-5-1":{"id":"duo-chat-fable-5-1","name":"Agentic Chat (Claude Fable 5.1)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-fable-5-1"},"duo-chat-gpt-6-luna":{"id":"duo-chat-gpt-6-luna","name":"Agentic Chat (GPT-6 Luna)","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"openai/gpt-6-luna"},"duo-chat-opus-4-8":{"id":"duo-chat-opus-4-8","name":"Agentic Chat (Claude Opus 4.8)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-opus-4-8"},"duo-chat-opus-4-6":{"id":"duo-chat-opus-4-6","name":"Agentic Chat (Claude Opus 4.6)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-opus-4-6"},"duo-chat-gpt-5-3-codex":{"id":"duo-chat-gpt-5-3-codex","name":"Agentic Chat (GPT-5.3 Codex)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.3-codex"},"duo-chat-gpt-5-2":{"id":"duo-chat-gpt-5-2","name":"Agentic Chat (GPT-5.2)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.2"},"duo-chat-gpt-5-4-mini":{"id":"duo-chat-gpt-5-4-mini","name":"Agentic Chat (GPT-5.4 Mini)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-5.4-mini"}}},"minimax-coding-plan":{"id":"minimax-coding-plan","env":["MINIMAX_API_KEY"],"npm":"@ai-sdk/anthropic","api":"https://api.minimax.io/anthropic/v1","name":"MiniMax Token Plan (minimax.io)","doc":"https://platform.minimax.io/docs/token-plan/intro","models":{"MiniMax-M2":{"id":"MiniMax-M2","name":"MiniMax-M2","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0}},"MiniMax-M3.1-Flash-Preview":{"id":"MiniMax-M3.1-Flash-Preview","name":"MiniMax-M3.1-Flash-Preview","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-27","last_updated":"2026-09-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":512000},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash-Preview"},"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":524288},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.5-highspeed":{"id":"MiniMax-M2.5-highspeed","name":"MiniMax-M2.5-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M3.1-Flash":{"id":"MiniMax-M3.1-Flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"MiniMax-M2.1":{"id":"MiniMax-M2.1","name":"MiniMax-M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.7-highspeed":{"id":"MiniMax-M2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}}}},"deepinfra":{"id":"deepinfra","env":["DEEPINFRA_API_KEY"],"npm":"@ai-sdk/deepinfra","name":"Deep Infra","doc":"https://deepinfra.com/models","models":{"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":2.85,"output":14.25,"cache_read":0.285},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.68,"output":3.4,"cache_read":0.136},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/Kimi-K2.5":{"id":"moonshotai/Kimi-K2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.45,"output":2.25,"cache_read":0.07},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.75,"output":3.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2.6"},"mistralai/Mistral-Small-24B-Instruct-2501":{"id":"mistralai/Mistral-Small-24B-Instruct-2501","name":"Mistral Small 3","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-01-30","last_updated":"2025-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.05,"output":0.08},"canonical_model_id":"mistral/mistral-small-24b-instruct-2501"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":512000},"cost":{"input":0.28,"output":1.1,"cache_read":0.056},"canonical_model_id":"minimax/MiniMax-M3"},"MiniMaxAI/MiniMax-M2.7":{"id":"MiniMaxAI/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"status":"deprecated","cost":{"input":0.25,"output":1,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M2.7"},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-06","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"status":"deprecated","cost":{"input":0.15,"output":1.15,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"thinkingmachines/Inkling":{"id":"thinkingmachines/Inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":1048576},"cost":{"input":0.95,"output":4.05,"cache_read":0.16},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/Inkling-Small":{"id":"thinkingmachines/Inkling-Small","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":1048576},"cost":{"input":0.45,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.03,"output":0.14},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.037,"output":0.17},"canonical_model_id":"openai/gpt-oss-120b"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50},"canonical_model_id":"anthropic/claude-fable-5"},"XiaomiMiMo/MiMo-V2.5-Pro":{"id":"XiaomiMiMo/MiMo-V2.5-Pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":16384},"status":"deprecated","cost":{"input":1,"output":3,"cache_read":0.2},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"XiaomiMiMo/MiMo-V2.5":{"id":"XiaomiMiMo/MiMo-V2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.5"},"XiaomiMiMo/MiMo-V2.6-Pro":{"id":"XiaomiMiMo/MiMo-V2.6-Pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.43,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"XiaomiMiMo/MiMo-V2.6-Flash":{"id":"XiaomiMiMo/MiMo-V2.6-Flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"tencent/Hy4-preview":{"id":"tencent/Hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"tencent/Hy3":{"id":"tencent/Hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":128000},"status":"deprecated","cost":{"input":0.13,"output":0.53,"cache_read":0.033},"canonical_model_id":"tencent/hy3"},"meta-llama/Llama-3.3-70B-Instruct-Turbo":{"id":"meta-llama/Llama-3.3-70B-Instruct-Turbo","name":"Llama 3.3 70B Turbo","description":"Compact Llama instruction model for fast chat and local deployment","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.1,"output":0.32},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8":{"id":"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8","name":"Llama 4 Maverick 17B FP8","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":16384},"status":"deprecated","cost":{"input":0.2,"output":0.8},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"meta-llama/Llama-Guard-4-12B":{"id":"meta-llama/Llama-Guard-4-12B","name":"Llama Guard 4 12B","description":"Multimodal Llama 4-based safety classifier for moderating prompts and responses","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.18,"output":0.18},"canonical_model_id":"meta/llama-guard-4-12b"},"meta-llama/Llama-4-Scout-17B-16E-Instruct":{"id":"meta-llama/Llama-4-Scout-17B-16E-Instruct","name":"Llama 4 Scout 17B","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":327680,"output":16384},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"zai-org/GLM-5":{"id":"zai-org/GLM-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"status":"deprecated","cost":{"input":0.6,"output":2.08,"cache_read":0.12},"canonical_model_id":"zhipuai/glm-5"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":32768},"cost":{"input":0.75,"output":2.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-4.7-Flash":{"id":"zai-org/GLM-4.7-Flash","name":"GLM-4.7-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"status":"deprecated","cost":{"input":0.06,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flash"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.9,"output":4,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"status":"deprecated","cost":{"input":1.05,"output":3.5,"cache_read":0.205},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-4.7":{"id":"zai-org/GLM-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"cost":{"input":0.4,"output":1.75,"cache_read":0.08},"canonical_model_id":"zhipuai/glm-4.7"},"zai-org/GLM-4.6":{"id":"zai-org/GLM-4.6","name":"GLM-4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.5,"output":2,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-4.6"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"stepfun-ai/Step-3.7-Flash":{"id":"stepfun-ai/Step-3.7-Flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"status":"deprecated","cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"stepfun-ai/Step-5-Preview":{"id":"stepfun-ai/Step-5-Preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1000000},"cost":{"input":1,"output":2.7,"cache_read":0.1},"canonical_model_id":"stepfun/step-5-preview"},"nvidia/Llama-3.3-Nemotron-Super-49B-v1.5":{"id":"nvidia/Llama-3.3-Nemotron-Super-49B-v1.5","name":"Llama 3.3 Nemotron Super 49B v1.5","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":0.4,"output":0.4},"canonical_model_id":"nvidia/llama-3.3-nemotron-super-49b-v1.5"},"nvidia/Nemotron-3-Nano-30B-A3B":{"id":"nvidia/Nemotron-3-Nano-30B-A3B","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.05,"output":0.2,"cache_read":0.025},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning":{"id":"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning","name":"Nemotron 3 Nano Omni 30B A3B Reasoning","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.2,"output":0.8},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"deepseek-ai/DeepSeek-V4.1-Flash-1017":{"id":"deepseek-ai/DeepSeek-V4.1-Flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.2,"output":0.6,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-ai/DeepSeek-V4-Flash-Vision-Exp":{"id":"deepseek-ai/DeepSeek-V4-Flash-Vision-Exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.06,"output":0.18,"cache_read":0.015},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-R1-0528":{"id":"deepseek-ai/DeepSeek-R1-0528","name":"DeepSeek-R1-0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":64000},"cost":{"input":0.5,"output":2.15,"cache_read":0.35},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek-ai/DeepSeek-V3.2":{"id":"deepseek-ai/DeepSeek-V3.2","name":"DeepSeek-V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":64000},"cost":{"input":0.26,"output":0.38,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":16384},"cost":{"input":1.3,"output":2.6,"cache_read":0.1},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/DeepSeek-V3":{"id":"deepseek-ai/DeepSeek-V3","name":"DeepSeek-V3","description":"Open DeepSeek MoE chat model for coding, math, and general reasoning","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":8192},"cost":{"input":0.32,"output":0.89},"canonical_model_id":"deepseek/deepseek-v3"},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek-V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":8192},"cost":{"input":0.25,"output":0.95,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek-ai/DeepSeek-V4-Flash":{"id":"deepseek-ai/DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":16384},"cost":{"input":0.09,"output":0.18,"cache_read":0.018},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.3,"output":2.6,"cache_read":0.1},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek-ai/DeepSeek-V4.1-Flash":{"id":"deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.2,"output":0.6,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-ai/DeepSeek-V3-0324":{"id":"deepseek-ai/DeepSeek-V3-0324","name":"DeepSeek V3 0324","description":"March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"status":"deprecated","cost":{"input":0.24,"output":0.9,"cache_read":0.135},"canonical_model_id":"deepseek/deepseek-v3-0324"},"Qwen/Qwen3.6-27B":{"id":"Qwen/Qwen3.6-27B","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.32,"output":3.2},"canonical_model_id":"alibaba/qwen3.6-27b"},"Qwen/Qwen3.7-Max":{"id":"Qwen/Qwen3.7-Max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.5,"tiers":[{"input":5,"output":15,"cache_read":1,"tier":{"type":"context","size":32000}},{"input":6.25,"output":18.5,"cache_read":1.25,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.7-max"},"Qwen/Qwen3.6-35B-A3B":{"id":"Qwen/Qwen3.6-35B-A3B","name":"Qwen3.6 35B A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":81920},"cost":{"input":0.1,"output":0.95,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"Qwen/Qwen3.5-27B":{"id":"Qwen/Qwen3.5-27B","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.26,"output":2.6},"canonical_model_id":"alibaba/qwen3.5-27b"},"Qwen/Qwen3-Next-80B-A3B-Instruct":{"id":"Qwen/Qwen3-Next-80B-A3B-Instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.09,"output":1.1},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B-A22B Instruct 2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.09,"output":0.55},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"Qwen/Qwen3-VL-30B-A3B-Instruct":{"id":"Qwen/Qwen3-VL-30B-A3B-Instruct","name":"Qwen3 VL 30B A3B Instruct","description":"Open vision-language MoE model for image and video understanding, documents, and visual agents","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"Qwen/Qwen3.5-35B-A3B":{"id":"Qwen/Qwen3.5-35B-A3B","name":"Qwen 3.5 35B A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":81920},"cost":{"input":0.14,"output":1,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":2.5,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3-VL-235B-A22B-Instruct":{"id":"Qwen/Qwen3-VL-235B-A22B-Instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":0.88,"cache_read":0.11},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"Qwen/Qwen3.5-397B-A17B":{"id":"Qwen/Qwen3.5-397B-A17B","name":"Qwen 3.5 397B A17B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":81920},"cost":{"input":0.45,"output":3,"cache_read":0.22},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"Qwen/Qwen3-30B-A3B":{"id":"Qwen/Qwen3-30B-A3B","name":"Qwen3 30B A3B","description":"Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.5},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo":{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo","name":"Qwen3 Coder 480B A35B Instruct Turbo","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":66536},"cost":{"input":0.3,"output":1,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"Qwen/Qwen3-Max":{"id":"Qwen/Qwen3-Max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"status":"deprecated","cost":{"input":1.2,"output":6,"cache_read":0.24,"tiers":[{"input":2.4,"output":12,"cache_read":0.48,"tier":{"type":"context","size":32000}},{"input":3,"output":15,"cache_read":0.6,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3-max"},"Qwen/Qwen3.8-Flash":{"id":"Qwen/Qwen3.8-Flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.113,"output":0.382,"cache_read":0.0141},"canonical_model_id":"alibaba/qwen3.8-flash"},"Qwen/Qwen3.5-122B-A10B":{"id":"Qwen/Qwen3.5-122B-A10B","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.29,"output":2.4},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"Qwen/Qwen3.5-9B":{"id":"Qwen/Qwen3.5-9B","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.1,"output":0.15},"canonical_model_id":"alibaba/qwen3.5-9b"},"Qwen/Qwen3.8-Max":{"id":"Qwen/Qwen3.8-Max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":131072},"cost":{"input":1.65,"output":4.951,"cache_read":0.206},"canonical_model_id":"alibaba/qwen3.8-max"},"Qwen/Qwen3.8-2.4T-A95B":{"id":"Qwen/Qwen3.8-2.4T-A95B","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"status":"deprecated","cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"Qwen/Qwen3-32B":{"id":"Qwen/Qwen3-32B","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.08,"output":0.28},"canonical_model_id":"alibaba/qwen3-32b"},"ByteDance/Seed-2.0-mini":{"id":"ByteDance/Seed-2.0-mini","name":"Seed 2.0 Mini","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.1,"output":0.4,"cache_read":0.02,"tiers":[{"input":0.2,"output":0.8,"cache_read":0.2,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"ByteDance/Seed-2.0-code":{"id":"ByteDance/Seed-2.0-code","name":"Seed 2.0 Code","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":131072},"cost":{"input":0.5,"output":3,"cache_read":0.1,"tiers":[{"input":1,"output":6,"cache_read":0.2,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"ByteDance/Seed-2.0-pro":{"id":"ByteDance/Seed-2.0-pro","name":"Seed 2.0 Pro","description":"Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.5,"output":3,"cache_read":0.1,"tiers":[{"input":1,"output":6,"cache_read":0.2,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-pro"},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemma-4-26B-A4B-it":{"id":"google/gemma-4-26B-A4B-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.07,"output":0.34},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.08,"output":0.16},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemma-3-4b-it":{"id":"google/gemma-3-4b-it","name":"Gemma 3 4B IT","description":"Open multimodal Gemma instruction model for efficient text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.1},"canonical_model_id":"google/gemma-3-4b-it"},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.15},"canonical_model_id":"google/gemma-3-12b-it"},"google/gemma-4-E4B-it":{"id":"google/gemma-4-E4B-it","name":"Gemma 4 E4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.02,"output":0.1},"canonical_model_id":"google/gemma-4-E4B-it"}}},"lowrouter":{"id":"lowrouter","env":["LOWROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.lowrouter.ai/v1","name":"LowRouter","doc":"https://lowrouter.ai/docs/models/available","models":{"auto/mistralai/ministral-3b-2512":{"id":"auto/mistralai/ministral-3b-2512","name":"Ministral 3 3B","description":"Compact open vision-language model for edge deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.1046,"output":0.1046,"cache_read":0.0105},"canonical_model_id":"mistral/ministral-3-3b-instruct-2512"},"auto/mistralai/mistral-large-4":{"id":"auto/mistralai/mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":262144},"status":"beta","cost":{"input":1.4367,"output":4.4156,"cache_read":0.1479},"canonical_model_id":"mistral/mistral-large-4"},"auto/mistralai/ministral-8b-2512":{"id":"auto/mistralai/ministral-8b-2512","name":"Ministral 3 8B","description":"Compact open vision-language model for edge deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.1569,"output":0.1569,"cache_read":0.0157},"canonical_model_id":"mistral/ministral-3-8b-instruct-2512"},"auto/mistralai/mistral-small-2603":{"id":"auto/mistralai/mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.1569,"output":0.6275,"cache_read":0.0157},"canonical_model_id":"mistral/mistral-small-2603"},"auto/mistralai/codestral-2508":{"id":"auto/mistralai/codestral-2508","name":"Codestral 25.08","description":"Mistral code model for low-latency completion, fill-in-the-middle, and code generation","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.3138,"output":0.9413,"cache_read":0.0314},"canonical_model_id":"mistral/codestral-2508"},"auto/mistralai/ministral-14b-2512":{"id":"auto/mistralai/ministral-14b-2512","name":"Ministral 3 14B","description":"Open vision-language model for efficient local deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.2092,"output":0.2092,"cache_read":0.0209},"canonical_model_id":"mistral/ministral-3-14b-instruct-2512"},"auto/mistralai/mistral-medium-3.5":{"id":"auto/mistralai/mistral-medium-3.5","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.5688,"output":7.8442,"cache_read":0.1569},"canonical_model_id":"mistral/mistral-medium-2604"},"auto/mistralai/mistral-large-2512":{"id":"auto/mistralai/mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5229,"output":1.5688,"cache_read":0.0523},"canonical_model_id":"mistral/mistral-large-2512"},"auto/openai/gpt-6-sol":{"id":"auto/openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"auto/openai/gpt-4.1-mini":{"id":"auto/openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"auto/openai/gpt-6-astra":{"id":"auto/openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"auto/openai/gpt-oss-safeguard-120b":{"id":"auto/openai/gpt-oss-safeguard-120b","name":"GPT OSS Safeguard 120B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.18,"output":0.73},"canonical_model_id":"openai/gpt-oss-safeguard-120b"},"auto/openai/gpt-oss-20b":{"id":"auto/openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.113,"output":0.4745},"canonical_model_id":"openai/gpt-oss-20b"},"auto/openai/gpt-5.6-terra":{"id":"auto/openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"auto/openai/gpt-5-mini":{"id":"auto/openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"auto/openai/gpt-5-nano":{"id":"auto/openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"auto/openai/gpt-5.2":{"id":"auto/openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"auto/openai/gpt-oss-safeguard-20b":{"id":"auto/openai/gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.08,"output":0.24},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"auto/openai/gpt-5.5":{"id":"auto/openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"auto/openai/o3":{"id":"auto/openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"auto/openai/gpt-5.4":{"id":"auto/openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"auto/openai/gpt-6-luna":{"id":"auto/openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"auto/openai/gpt-oss-120b":{"id":"auto/openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.1695,"output":0.6779},"canonical_model_id":"openai/gpt-oss-120b"},"auto/openai/gpt-5.6-sol":{"id":"auto/openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"auto/openai/gpt-6.1-sol":{"id":"auto/openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"auto/openai/gpt-5.6-luna":{"id":"auto/openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"auto/openai/gpt-4o":{"id":"auto/openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"auto/openai/gpt-4o-mini":{"id":"auto/openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"auto/openai/gpt-4.1":{"id":"auto/openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"auto/openai/gpt-5":{"id":"auto/openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"auto/openai/gpt-5.4-mini":{"id":"auto/openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"}}},"infomaniak":{"id":"infomaniak","env":["INFOMANIAK_API_KEY","INFOMANIAK_PRODUCT_ID"],"npm":"@ai-sdk/openai-compatible","api":"https://api.infomaniak.com/2/ai/${INFOMANIAK_PRODUCT_ID}/openai/v1","name":"Infomaniak","doc":"https://www.infomaniak.com/en/hosting/ai-services/open-source-models","models":{"bge_multilingual_gemma2":{"id":"bge_multilingual_gemma2","type":"embedding","name":"BGE Multilingual Gemma2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-07-25","last_updated":"2026-08-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"input":8000,"output":3584},"cost":{"input":0.08,"output":0}},"mini_lm_l12_v2":{"id":"mini_lm_l12_v2","type":"embedding","name":"All-MiniLM-L12-v2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2021-08-30","last_updated":"2021-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128,"input":128,"output":384},"cost":{"input":0,"output":0},"canonical_model_id":"sentence-transformers/all-minilm-l12-v2"},"moonshotai/Kimi-K2.6":{"id":"moonshotai/Kimi-K2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"status":"beta","cost":{"input":0.74,"output":3.72},"canonical_model_id":"moonshotai/kimi-k2.6"},"mistralai/Ministral-3-14B-Instruct-2512":{"id":"mistralai/Ministral-3-14B-Instruct-2512","name":"Ministral 3 14B Instruct","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":100000,"input":100000,"output":25600},"status":"beta","cost":{"input":0.37,"output":0.5},"canonical_model_id":"mistral/ministral-3-14b-instruct-2512"},"mistralai/Mistral-Small-4-119B-2603":{"id":"mistralai/Mistral-Small-4-119B-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.25,"output":0.93},"canonical_model_id":"mistral/mistral-small-2603"},"swiss-ai/Apertus-v1.5-70B":{"id":"swiss-ai/Apertus-v1.5-70B","name":"Apertus v1.5 70B","description":"Open, ethically-sourced Swiss AI model for multilingual, multimodal chat and instruction following","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-07-24","last_updated":"2026-08-01","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":100000,"input":100000,"output":8192},"status":"beta","cost":{"input":0.87,"output":3.1}},"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8":{"id":"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8","name":"Nemotron 3 Nano 30B A3B FP8","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":262144},"status":"deprecated","cost":{"input":0.06,"output":0.25},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"Qwen/Qwen3.5-122B-A10B-FP8":{"id":"Qwen/Qwen3.5-122B-A10B-FP8","name":"Qwen3.5 122B-A10B FP8","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65536},"status":"deprecated","cost":{"input":0.5,"output":3.97},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"Qwen/Qwen3.5-397B-A17B-FP8":{"id":"Qwen/Qwen3.5-397B-A17B-FP8","name":"Qwen3.5 397B-A17B FP8","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65536},"status":"beta","cost":{"input":0.99,"output":4.46},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":100000,"input":100000,"output":32768},"cost":{"input":0.25,"output":0.5},"canonical_model_id":"google/gemma-4-31b-it"}}},"blueclaw":{"id":"blueclaw","env":["BLUECLAW_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://openai.blueclaw.network/v1","name":"Blue Claw","doc":"https://blueclaw.network","models":{"Qwen3.8-27B":{"id":"Qwen3.8-27B","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"beta","canonical_model_id":"alibaba/qwen3.8-27b"}}},"cortecs":{"id":"cortecs","env":["CORTECS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.cortecs.ai/v1","name":"Cortecs","doc":"https://api.cortecs.ai/v1/models","models":{"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.825,"output":4.125,"cache_read":0.082,"cache_write":0.084},"canonical_model_id":"google/gemini-3.8-flash"},"pixtral-large-2502":{"id":"pixtral-large-2502","name":"Pixtral Large (25.02)","description":"Pixtral Large (25.02) is a 124B open-weight multimodal model built on Mistral Large 2, offering advanced image understanding and strong performance across text and code tasks.","family":"pixtral","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-04-08","last_updated":"2025-04-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":1.993,"output":5.978},"canonical_model_id":"mistral/pixtral-large-2502"},"devstral-2512":{"id":"devstral-2512","name":"Devstral 2","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.478,"output":2.392,"cache_read":0.045},"canonical_model_id":"mistral/devstral-2512"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.272,"output":1.631,"cache_read":0.025,"cache_write":0.082},"canonical_model_id":"google/gemini-3.1-flash-lite"},"claude-opus4-5":{"id":"claude-opus4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5.313,"output":26.568,"cache_read":0.531,"cache_write":6.645},"canonical_model_id":"anthropic/claude-opus-4-5"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.4,"output":11.999,"cache_read":0.24,"cache_write":3},"canonical_model_id":"openai/gpt-6-sol"},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.299,"output":2.491,"cache_read":0.029,"cache_write":0.097},"canonical_model_id":"google/gemini-2.5-flash"},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.375,"output":10.96,"cache_read":0.156},"canonical_model_id":"openai/gpt-5.1"},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.434,"output":1.704,"cache_read":0.134},"canonical_model_id":"openai/gpt-4.1-mini"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":12.272,"output":61.359,"cache_read":1.228,"cache_write":15.34},"canonical_model_id":"openai/gpt-6-astra"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":2.25,"output":11.249,"cache_read":0.113},"canonical_model_id":"google/gemini-4-argon"},"gpt-oss-safeguard-120b":{"id":"gpt-oss-safeguard-120b","name":"GPT OSS Safeguard 120B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.179,"output":0.697},"canonical_model_id":"openai/gpt-oss-safeguard-120b"},"qwen3guard-gen-0.6b":{"id":"qwen3guard-gen-0.6b","name":"qwen3guard-gen-0.6b","description":"Qwen3Guard-Gen-0.6B is a lightweight multilingual safety moderation model that classifies prompts and responses into safe, controversial, or unsafe categories.","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-02-04","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"cost":{"input":0,"output":0}},"apertus-70b":{"id":"apertus-70b","name":"Apertus 70B","description":"Apertus 70B is an open, multilingual language model designed for research, long-context reasoning, and sovereignty-focused AI systems.","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-09","release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":1.393,"output":2.228},"canonical_model_id":"swiss-ai/apertus-70b"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.045,"output":0.167},"canonical_model_id":"openai/gpt-oss-20b"},"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":500000},"cost":{"input":0.1,"output":0.4,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.8-27b"},"claude-sonnet-5.5":{"id":"claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.25,"output":11.249,"cache_read":0.225,"cache_write":2.813},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.2,"output":13.199,"cache_read":0.219,"cache_write":2.749},"canonical_model_id":"openai/gpt-5.6-terra"},"ministral-3b-2512":{"id":"ministral-3b-2512","name":"Ministral 3 3B 2512","description":"Ministral 3 3B is a compact, efficient multimodal model with strong language, vision capabilities, and ideal for custom fine-tuning.","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.123,"output":0.123,"cache_read":0.012},"canonical_model_id":"mistral/ministral-3b-2512"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.279,"output":2.192,"cache_read":0.056},"canonical_model_id":"openai/gpt-5-mini"},"qwen3-vl-235b-a22b":{"id":"qwen3-vl-235b-a22b","name":"qwen3-vl-235b-a22b","description":"Qwen3 VL 235B A22B is a 235B-parameter MoE vision-language flagship model (≈22B active) designed for frontier-level multimodal understanding across text, images, documents, and long videos.","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.617,"output":3.119,"cache_read":0.052},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"claude-4-5-sonnet":{"id":"claude-4-5-sonnet","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":2.989,"output":14.945,"cache_read":0.326,"cache_write":4.078},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":0.996,"output":4.982,"cache_read":0.099,"cache_write":1.186},"canonical_model_id":"anthropic/claude-haiku-4-5"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":32768},"cost":{"input":0.167,"output":0.557},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen3.8-2.4t-a95b":{"id":"qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":2.5,"output":6,"cache_read":0.625},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"qwen3-coder-next":{"id":"qwen3-coder-next","name":"Qwen3 Coder Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.167,"output":0.891},"canonical_model_id":"alibaba/qwen3-coder-next"},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.06,"output":0.439,"cache_read":0.019},"canonical_model_id":"openai/gpt-5-nano"},"mistral-large-2402":{"id":"mistral-large-2402","name":"mistral-large-2402","description":"Mistral Large (24.02) is Mistral AI’s most advanced language model, built for complex multilingual reasoning, code generation, and deep text understanding.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-05-26","last_updated":"2025-05-26","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":8192},"cost":{"input":4.284,"output":12.952}},"qwen3guard-gen-8b":{"id":"qwen3guard-gen-8b","name":"qwen3guard-gen-8b","description":"Qwen3Guard-Gen-8B is a large-scale multilingual safety moderation model designed for high-accuracy prompt and response classification.","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-02-04","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"cost":{"input":0,"output":0}},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.475,"output":2.97,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.6"},"qwen3-30b-a3b-instruct-2507":{"id":"qwen3-30b-a3b-instruct-2507","name":"qwen3-30b-a3b-instruct-2507","description":"Qwen3-30B-A3B-Instruct-2507 is an advanced Mixture-of-Experts model optimized for reasoning, coding, and multilingual instruction following.","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.099,"output":0.299},"canonical_model_id":"alibaba/qwen3-30b-a3b-instruct-2507"},"gemma-3-27b-it":{"id":"gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Gemma 3 is a family of lightweight, multimodal models from Google, supporting text and image inputs, multilingual capabilities, and a 131K context window.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":110000},"cost":{"input":0.099,"output":0.299},"canonical_model_id":"google/gemma-3-27b-it"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":65536},"cost":{"input":0.5,"output":1.499,"cache_read":0.15},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.511,"output":1.534,"cache_read":0.128},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":2,"output":3.999,"cache_read":0.5},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2.898,"output":15.453,"cache_read":0.242},"canonical_model_id":"openai/gpt-5.4"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.12,"output":0.6,"cache_read":0.012,"cache_write":0.15},"canonical_model_id":"openai/gpt-6-luna"},"qwen3-235b-a22b-instruct-2507":{"id":"qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B-A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.199,"output":0.598,"cache_read":0.018},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.7,"output":3.4,"cache_read":0.15},"canonical_model_id":"deepseek/deepseek-v4-pro"},"qwen2.5-vl-72b-instruct":{"id":"qwen2.5-vl-72b-instruct","name":"qwen2.5-vl-72b-instruct","description":"Qwen2.5-VL is a powerful vision-language model with advanced capabilities in visual understanding, long video reasoning, and structured output generation.","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2024-04","release_date":"2025-01-26","last_updated":"2025-01-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000},"cost":{"input":1.014,"output":1.014},"canonical_model_id":"alibaba/qwen2-5-vl-72b-instruct"},"minimax-m2.1":{"id":"minimax-m2.1","name":"MiniMax-M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196000,"output":196000},"cost":{"input":0.359,"output":1.435},"canonical_model_id":"minimax/MiniMax-M2.1"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":202752},"cost":{"input":1.384,"output":4.348,"cache_read":0.346},"canonical_model_id":"zhipuai/glm-5.1"},"nova-pro-v1":{"id":"nova-pro-v1","name":"Nova Pro 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.918,"output":3.671},"canonical_model_id":"amazon/nova-pro-v1"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.089,"output":0.446,"cache_read":0.01},"canonical_model_id":"openai/gpt-oss-120b"},"nova-lite-v1":{"id":"nova-lite-v1","name":"Nova Lite 1.0","description":"Nova Lite is a fast, low-cost multimodal foundation model capable of reasoning over text, images, and video in 200+ languages.","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.069,"output":0.275},"canonical_model_id":"amazon/nova-lite-v1"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4.399,"output":21.998,"cache_read":0.44,"cache_write":5.5},"canonical_model_id":"openai/gpt-5.6-sol"},"ministral-8b-2512":{"id":"ministral-8b-2512","name":"Ministral 3 8B 2512","description":"Ministral 3 8B is a balanced, efficient multimodal model offering strong text and vision capabilities, optimized for edge and local deployment.","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.179,"output":0.179,"cache_read":0.017},"canonical_model_id":"mistral/ministral-8b-2512"},"mistral-small-2603":{"id":"mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.156,"output":0.625,"cache_read":0.016},"canonical_model_id":"mistral/mistral-small-2603"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.656,"output":3.3,"cache_read":0.18},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"mistral-large-4-0":{"id":"mistral-large-4-0","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":262144},"cost":{"input":0.765,"output":2.351,"cache_read":0.079},"canonical_model_id":"mistral/mistral-large-4"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.33,"output":2.749,"cache_read":0.033,"cache_write":0.082},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.4,"output":11.999,"cache_read":0.12,"cache_write":3},"canonical_model_id":"openai/gpt-6.1-sol"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":14.999},"canonical_model_id":"moonshotai/kimi-k3"},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.668,"output":4.01},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.219,"output":1.32,"cache_read":0.022,"cache_write":0.275},"canonical_model_id":"openai/gpt-5.6-luna"},"qwen3-coder-30b-a3b-instruct":{"id":"qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.179,"output":0.697,"cache_read":0.014},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"codestral-2508":{"id":"codestral-2508","name":"Codestral 25.08","description":"Mistral coding model for code completion, generation, and developer workflows","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.368,"output":1.103,"cache_read":0.037},"canonical_model_id":"mistral/codestral-2508"},"mistral-small-3.2-24b-instruct-2506":{"id":"mistral-small-3.2-24b-instruct-2506","name":"mistral-small-3.2-24b-instruct-2506","description":"Mistral-Small-3.2-24B-Instruct-2506 is a 24B parameter instruction-tuned model with enhanced long-context support (128k) and state-of-the-art vision understanding.","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-10-31","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.1,"output":0.312},"canonical_model_id":"mistral/mistral-small-3.2-24b-instruct"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.498,"cache_read":0.55,"cache_write":6.874},"canonical_model_id":"anthropic/claude-opus-5"},"claude-opus-5.5":{"id":"claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4.399,"output":21.998,"cache_read":0.219,"cache_write":5.5},"canonical_model_id":"anthropic/claude-opus-5-5"},"qwen3.8-flash-next":{"id":"qwen3.8-flash-next","name":"Qwen3.8 Flash Next","description":"Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":64000},"cost":{"input":0.201,"output":0.5,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.8-flash-next"},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.724,"output":0.724},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"nemotron-nano-v2-12b":{"id":"nemotron-nano-v2-12b","name":"nemotron-nano-v2-12b","description":"NVIDIA Nemotron Nano v2 12B is a 12-billion-parameter multimodal reasoning model designed for advanced video understanding, document intelligence, and visual reasoning, built with a hybrid Transformer-Mamba architecture for high efficiency and low latency.","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-10-28","last_updated":"2025-10-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.24,"output":0.707},"canonical_model_id":"nvidia/nemotron-nano-12b-v2-vl"},"gemma-4-26b-a4b-it":{"id":"gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":81920},"cost":{"input":0.111,"output":0.557},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000},"cost":{"input":2.659,"output":10.635,"cache_read":1.33},"canonical_model_id":"openai/gpt-4o"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.038},"canonical_model_id":"google/gemini-3.6-flash"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.038},"canonical_model_id":"google/gemini-3.7-flash"},"mistral-nemo-instruct-2407":{"id":"mistral-nemo-instruct-2407","name":"mistral-nemo-instruct-2407","description":"A 12B parameter, instruct-tuned language model by Mistral AI and NVIDIA, designed for advanced instruction following, multi-turn conversations, and generating text and code across multiple languages.","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.145,"output":0.145,"cache_read":0.014},"canonical_model_id":"mistral/mistral-nemo"},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.111,"output":0.434,"cache_read":0.056},"canonical_model_id":"openai/gpt-4.1-nano"},"gpt-4o-mini":{"id":"gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000},"cost":{"input":0.159,"output":0.638,"cache_read":0.081},"canonical_model_id":"openai/gpt-4o-mini"},"nvidia-nemotron-3-nano-30b-a3b":{"id":"nvidia-nemotron-3-nano-30b-a3b","name":"nvidia-nemotron-3-nano-30b-a3b","description":"Nemotron-Nano-3-30B-A3B is a compact Mixture-of-Experts model optimized for efficient reasoning, chat, and coding, with strong multilingual support and long-context RAG and agent workflows.","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.06,"output":0.24},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"mixtral-8x7B-instruct-v0.1":{"id":"mixtral-8x7B-instruct-v0.1","name":"Mixtral 8x7B Instruct v0.1","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mistral","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2023-12-11","last_updated":"2023-12-11","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.488,"output":0.758},"canonical_model_id":"mistral/mixtral-8x7b-instruct-v0.1"},"claude-opus4-8":{"id":"claude-opus4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.437,"output":27.186,"cache_read":0.544,"cache_write":6.797},"canonical_model_id":"anthropic/claude-opus-4-8"},"minimax-m2":{"id":"minimax-m2","name":"MiniMax-M2","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":400000,"output":196000},"cost":{"input":0.349,"output":1.405},"canonical_model_id":"minimax/MiniMax-M2"},"minicpm-v-4.5":{"id":"minicpm-v-4.5","name":"minicpm-v-4.5","description":"MiniCPM-V 4.5 is a compact, high-performance vision-language model excelling in video understanding, OCR, and multimodal reasoning with efficient deployment.","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":32000},"cost":{"input":0.651,"output":1.097}},"claude-opus4-7":{"id":"claude-opus4-7","name":"Claude Opus 4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.437,"output":27.186,"cache_read":0.544,"cache_write":6.797},"canonical_model_id":"anthropic/claude-opus-4-7"},"ministral-14b-2512":{"id":"ministral-14b-2512","name":"Ministral 3 14B 2512","description":"Ministral 3 14B is a frontier-level 14B multimodal model optimized for local deployment, delivering state-of-the-art text and vision reasoning with a 256K context window and strong agentic capabilities.","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.24,"output":0.24,"cache_read":0.022},"canonical_model_id":"mistral/ministral-14b-2512"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.1,"output":0.35,"cache_read":0.018},"canonical_model_id":"zhipuai/glm-5.3-flash"},"qwen3-32b":{"id":"qwen3-32b","name":"Qwen3 32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":0.179,"output":0.697},"canonical_model_id":"alibaba/qwen3-32b"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1000000},"cost":{"input":0.9,"output":3.24,"cache_read":0.189},"canonical_model_id":"zhipuai/glm-5.2"},"nova-2-lite":{"id":"nova-2-lite","name":"Nova 2 Lite","description":"Nova 2 Lite is an advanced multimodal reasoning model that combines efficiency and performance, delivering reliable AI for agentic workflows and enterprise applications.","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.373,"output":3.144},"canonical_model_id":"amazon/nova-2-lite"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":1.649,"output":9.899,"cache_read":0.165,"cache_write":1},"canonical_model_id":"google/gemini-3.5-flash"},"glm-4.7-flash":{"id":"glm-4.7-flash","name":"GLM-4.7-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":203000},"cost":{"input":0.08,"output":0.478},"canonical_model_id":"zhipuai/glm-4.7-flash"},"claude-haiku-5.5":{"id":"claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.113,"output":0.563,"cache_read":0.011,"cache_write":0.14},"canonical_model_id":"anthropic/claude-haiku-5-5"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.201,"output":0.6,"cache_read":0.004},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.2,"output":11,"cache_read":0.219,"cache_write":2.749},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2.192,"output":8.769,"cache_read":0.546},"canonical_model_id":"openai/gpt-4.1"},"mistral-medium-3.5":{"id":"mistral-medium-3.5","name":"mistral-medium-3.5","description":"Mistral Medium 3.5 is a frontier multimodal 128B model combining reasoning, coding, and instruction-following with strong agentic performance and efficient deployment.","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1.532,"output":7.843,"cache_read":0.154},"canonical_model_id":"mistral/mistral-medium-3-5"},"qwen3.5-9b":{"id":"qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.111,"output":0.167},"canonical_model_id":"alibaba/qwen3.5-9b"},"mistral-small-2503":{"id":"mistral-small-2503","name":"mistral-small-2503","description":"Combines advanced text and vision capabilities with 24 billion parameters, supporting multilingual tasks and long contexts up to 131k tokens, making it versatile for various applications without sacrificing performance.","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-06","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.111,"output":0.334},"canonical_model_id":"mistral/mistral-small-3-1-24b-instruct-2503"},"minimax-m2.5":{"id":"minimax-m2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196000,"output":196000},"cost":{"input":0.348,"output":1.39,"cache_read":0.075},"canonical_model_id":"minimax/MiniMax-M2.5"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.375,"output":10.96,"cache_read":0.156},"canonical_model_id":"openai/gpt-5"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.082,"output":0.413,"cache_read":0.06},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"nova-micro-v1":{"id":"nova-micro-v1","name":"Nova Micro 1.0","description":"Nova Micro is a multilingual text-to-text foundation model with strong reasoning capabilities and broad language coverage across 200+ languages.","family":"nova-micro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.04,"output":0.159},"canonical_model_id":"amazon/nova-micro-v1"},"mistral-7b-instruct-v0.2":{"id":"mistral-7b-instruct-v0.2","name":"mistral-7b-instruct-v0.2","description":"Mistral 7B Instruct is a compact, 7B parameter model optimized for fast and efficient text and code generation with a 32K token context window.","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2025-05-26","last_updated":"2025-05-26","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":8192},"cost":{"input":0.159,"output":0.219}},"gemma-4-31b-it":{"id":"gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.223,"output":0.39},"canonical_model_id":"google/gemma-4-31b-it"},"claude-4-6-sonnet":{"id":"claude-4-6-sonnet","name":"Claude Sonnet 4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3.196,"output":15.94,"cache_read":0.32,"cache_write":3.999},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.511,"output":2.045,"cache_read":0.154},"canonical_model_id":"zhipuai/glm-5.3"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":1.495,"output":9.964,"cache_read":0.242,"cache_write":0.434},"canonical_model_id":"google/gemini-2.5-pro"},"hermes-4-405b":{"id":"hermes-4-405b","name":"Hermes 4 405B","description":"Hermes 4 405B is a frontier hybrid-mode reasoning model built on Llama 3.1, optimized for advanced logic, math, coding, and structured output generation.","family":"hermes","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-08-31","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.996,"output":2.989},"canonical_model_id":"nousresearch/hermes-4-405b"},"voxtral-small-2507":{"id":"voxtral-small-2507","name":"voxtral-small-2507","description":"Voxtral Small is a multimodal model with audio input, combining advanced speech capabilities with strong text performance for transcription, translation, and audio understanding.","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000},"cost":{"input":0.123,"output":0.368,"cache_read":0.012},"canonical_model_id":"mistral/voxtral-small-24b-2507"},"claude-opus4-6":{"id":"claude-opus4-6","name":"Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.313,"output":26.561,"cache_read":0.531,"cache_write":6.645},"canonical_model_id":"anthropic/claude-opus-4-6"},"qwen3.6-27b":{"id":"qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.446,"output":3.008},"canonical_model_id":"alibaba/qwen3.6-27b"},"mistral-large-2512":{"id":"mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.613,"output":1.838,"cache_read":0.061},"canonical_model_id":"mistral/mistral-large-2512"}}},"venice":{"id":"venice","env":["VENICE_API_KEY"],"npm":"venice-ai-sdk-provider","name":"Venice AI","doc":"https://docs.venice.ai","models":{"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"zai-org-glm-5":{"id":"zai-org-glm-5","name":"GLM 5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":32000},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"minimax-m27":{"id":"minimax-m27","name":"MiniMax M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":32768},"cost":{"input":0.375,"output":1.5,"cache_read":0.06875},"canonical_model_id":"minimax/MiniMax-M2.7"},"grok-4-20":{"id":"grok-4-20","name":"Grok 4.20","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-03-12","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":128000},"cost":{"input":1.42,"output":2.83,"cache_read":0.23,"tiers":[{"input":2.83,"output":5.67,"cache_read":0.45,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.83,"output":5.67,"cache_read":0.45}}},"openai-gpt-56-luna-pro":{"id":"openai-gpt-56-luna-pro","name":"GPT-5.6 Luna Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"z-ai-glm-5v-turbo":{"id":"z-ai-glm-5v-turbo","name":"GLM 5V Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32768},"cost":{"input":1.5,"output":5,"cache_read":0.3},"canonical_model_id":"zhipuai/glm-5v-turbo"},"seed-2-1-turbo":{"id":"seed-2-1-turbo","name":"Seed 2.1 Turbo","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"cost":{"input":0.625,"output":3.125,"cache_read":0.125},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"mercury-2-5":{"id":"mercury-2-5","name":"Mercury 2.5","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":260000,"output":65536},"cost":{"input":0.25,"output":0.9375,"cache_read":0.025},"canonical_model_id":"inception/mercury-2.5"},"qwen-3-8-max":{"id":"qwen-3-8-max","name":"Qwen 3.8 Max","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-07-19","last_updated":"2026-07-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.3125,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"deepseek-v4-flash-0731-fast":{"id":"deepseek-v4-flash-0731-fast","name":"DeepSeek V4 Flash 0731 Fast","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":0.35,"output":0.7,"cache_read":0.0875},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"qwen3-6-27b":{"id":"qwen3-6-27b","name":"Qwen 3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.325,"output":3.25},"canonical_model_id":"alibaba/qwen3.6-27b"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Gemini multimodal model for text, image, audio, video, and document tasks","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":2.5,"output":12.5,"cache_read":0.125},"canonical_model_id":"google/gemini-4-argon"},"xiaomi-mimo-v2-5":{"id":"xiaomi-mimo-v2-5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0.4,"output":2,"cache_read":0.08},"canonical_model_id":"xiaomi/mimo-v2.5"},"grok-4-6":{"id":"grok-4-6","name":"Grok 4.6","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":200000},"cost":{"input":2.27,"output":6.8,"cache_read":0.57,"tiers":[{"input":4.53,"output":13.6,"cache_read":1.13,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4.53,"output":13.6,"cache_read":1.13}},"canonical_model_id":"xai/grok-4.6"},"grok-4-7":{"id":"grok-4-7","name":"Grok 4.7","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":200000},"cost":{"input":2.27,"output":6.8,"cache_read":0.57,"tiers":[{"input":4.53,"output":13.6,"cache_read":1.13,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4.53,"output":13.6,"cache_read":1.13}},"canonical_model_id":"xai/grok-4.7"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"qwen-3-6-plus":{"id":"qwen-3-6-plus","name":"Qwen 3.6 Plus Uncensored","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.625,"output":3.75,"cache_read":0.0625,"cache_write":0.78,"tiers":[{"input":2.5,"output":7.5,"cache_read":0.0625,"cache_write":0.78,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2.5,"output":7.5,"cache_read":0.0625,"cache_write":0.78}},"canonical_model_id":"alibaba/qwen3.6-plus"},"aion-labs-aion-3-0":{"id":"aion-labs-aion-3-0","name":"Aion 3.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-05","last_updated":"2026-05-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":3.75,"output":7.5,"cache_read":0.9375},"canonical_model_id":"aion-labs/aion-3.0"},"nvidia-nemotron-3-ultra-550b-a55b":{"id":"nvidia-nemotron-3-ultra-550b-a55b","name":"NVIDIA Nemotron 3 Ultra","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"cost":{"input":0.625,"output":3.125,"cache_read":0.1875},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"qwen3-vl-235b-a22b":{"id":"qwen3-vl-235b-a22b","name":"Qwen3 VL 235B","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-01-16","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.21,"output":1.9,"cache_read":0.1}},"openai-gpt-55":{"id":"openai-gpt-55","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":131072},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"qwen3-next-80b":{"id":"qwen3-next-80b","name":"Qwen 3 Next 80b","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.35,"output":1.9},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"zai-org-glm-4.6":{"id":"zai-org-glm-4.6","name":"GLM 4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":16384},"cost":{"input":0.43,"output":1.75,"cache_read":0.08},"canonical_model_id":"zhipuai/glm-4.6"},"venice-uncensored-1-2":{"id":"venice-uncensored-1-2","name":"Venice Uncensored 1.2","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"venice","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-04-01","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.2,"output":0.9}},"qwen3-6-35b-a3b":{"id":"qwen3-6-35b-a3b","name":"Qwen 3.6 35B A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.1,"output":1},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"z-ai-glm-5-3-flash":{"id":"z-ai-glm-5-3-flash","name":"GLM 5.3 Flash","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":198000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash 0423","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":0.138,"output":0.275,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"mercury-3":{"id":"mercury-3","name":"Mercury 3","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-16","last_updated":"2026-10-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1,"output":0.35,"cache_read":0.01},"canonical_model_id":"inception/mercury-3"},"deepseek-v4-1-flash-1017":{"id":"deepseek-v4-1-flash-1017","name":"DeepSeek v4.1 Flash (1017)","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.0075},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"openai-gpt-56-sol":{"id":"openai-gpt-56-sol","name":"GPT-5.6 Sol","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai-gpt-61-sol-ultrafast":{"id":"openai-gpt-61-sol-ultrafast","name":"GPT-6.1 Sol Ultrafast","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":12,"output":60,"cache_read":0.6,"cache_write":15,"tiers":[{"input":24,"output":90,"cache_read":1.2,"cache_write":30,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":24,"output":90,"cache_read":1.2,"cache_write":30}},"canonical_model_id":"openai/gpt-6.1-sol"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":160000,"output":32768},"status":"deprecated","cost":{"input":0.33,"output":0.48,"cache_read":0.16},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":1.65,"output":4.95,"cache_read":0.165},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"gemini-3-5-flash-lite":{"id":"gemini-3-5-flash-lite","name":"Gemini 3.5 Flash-Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.375,"output":3.125,"cache_read":0.0375},"canonical_model_id":"google/gemini-3.5-flash-lite"},"aion-labs-aion-3-5-mini":{"id":"aion-labs-aion-3-5-mini","name":"Aion 3.5 Mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.875,"output":1.75,"cache_read":0.225}},"qwen3-235b-a22b-instruct-2507":{"id":"qwen3-235b-a22b-instruct-2507","name":"Qwen 3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.75},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"abliteration-abliterated-model-large-v2":{"id":"abliteration-abliterated-model-large-v2","name":"Abliterated Large V2","description":"Flagship model for demanding analysis, coding, and production agent workflows","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":3,"output":5,"cache_read":0.3}},"qwen3-coder-480b-a35b-instruct-turbo":{"id":"qwen3-coder-480b-a35b-instruct-turbo","name":"Qwen 3 Coder 480B Turbo","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-01-27","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.35,"output":1.5,"cache_read":0.04}},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":1.65,"output":3.301,"cache_read":0.33},"canonical_model_id":"deepseek/deepseek-v4-pro"},"kimi-k2-6":{"id":"kimi-k2-6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.75,"output":3.5,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"venice-uncensored-role-play":{"id":"venice-uncensored-role-play","name":"Venice Role Play Uncensored","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"venice","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-02-20","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.5,"output":2}},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"mercury-2":{"id":"mercury-2","name":"Mercury 2","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":50000},"cost":{"input":0.3125,"output":0.9375,"cache_read":0.03125},"canonical_model_id":"inception/mercury-2"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"aion-labs-aion-3-0-mini":{"id":"aion-labs-aion-3-0-mini","name":"Aion 3.0 Mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-14","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":0.875,"output":1.75,"cache_read":0.225},"canonical_model_id":"aion-labs/aion-3.0-mini"},"kimi-k2-7-code":{"id":"kimi-k2-7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.75,"output":3.5,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"nvidia-nemotron-3-5-super-vl-120b-a12b":{"id":"nvidia-nemotron-3-5-super-vl-120b-a12b","name":"NVIDIA Nemotron 3.5 Super VL 120B","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.25,"output":1.25,"cache_read":0.125},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"openai-gpt-56-terra-pro":{"id":"openai-gpt-56-terra-pro","name":"GPT-5.6 Terra Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai-gpt-oss-120b":{"id":"openai-gpt-oss-120b","name":"OpenAI GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.07,"output":0.3},"canonical_model_id":"openai/gpt-oss-120b"},"qwen-3-8-27b":{"id":"qwen-3-8-27b","name":"Qwen 3.8 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.45,"output":3.2},"canonical_model_id":"alibaba/qwen3.8-27b"},"openai-gpt-54-pro":{"id":"openai-gpt-54-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"openai-gpt-54-mini":{"id":"openai-gpt-54-mini","name":"GPT-5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":3.75,"output":18.75,"cache_read":0.375},"canonical_model_id":"moonshotai/kimi-k3"},"kimi-k3-fast-api":{"id":"kimi-k3-fast-api","name":"Kimi K3 Fast","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":4.5,"output":22.5,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"kimi-k2-5":{"id":"kimi-k2-5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.56,"output":3.5,"cache_read":0.22},"canonical_model_id":"moonshotai/kimi-k2.5"},"qwen3-5-35b-a3b":{"id":"qwen3-5-35b-a3b","name":"Qwen 3.5 35B A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.15,"output":1,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"hermes-3-llama-3.1-405b":{"id":"hermes-3-llama-3.1-405b","name":"Hermes 3 Llama 3.1 405b","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-16","last_updated":"2024-08-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":1.1,"output":3},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-405b"},"claude-opus-5-fast":{"id":"claude-opus-5-fast","name":"Claude Opus 5 Fast","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-opus-5"},"openai-gpt-6-luna":{"id":"openai-gpt-6-luna","name":"GPT-6 Luna","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai-gpt-4o-mini-2024-07-18":{"id":"openai-gpt-4o-mini-2024-07-18","name":"GPT-4o Mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"llama-3.2-3b":{"id":"llama-3.2-3b","name":"Llama 3.2 3B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"meta/llama-3.2-3b"},"grok-4-20-multi-agent":{"id":"grok-4-20-multi-agent","name":"Grok 4.20 Multi-Agent","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"release_date":"2026-03-12","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":128000},"cost":{"input":1.42,"output":2.83,"cache_read":0.23,"tiers":[{"input":2.83,"output":5.67,"cache_read":0.45,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.83,"output":5.67,"cache_read":0.45}}},"openai-gpt-4o-2024-11-20":{"id":"openai-gpt-4o-2024-11-20","name":"GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"qwen-3-7-max":{"id":"qwen-3-7-max","name":"Qwen 3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":2.7,"output":8.05,"cache_read":0.27,"cache_write":3.35},"canonical_model_id":"alibaba/qwen3.7-max"},"openai-gpt-56-luna":{"id":"openai-gpt-56-luna","name":"GPT-5.6 Luna","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"gemini-3-6-flash":{"id":"gemini-3-6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.9375,"output":4.6875,"cache_read":0.09375},"canonical_model_id":"google/gemini-3.6-flash"},"nvidia-nemotron-3-nano-30b-a3b":{"id":"nvidia-nemotron-3-nano-30b-a3b","name":"NVIDIA Nemotron 3 Nano 30B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.075,"output":0.3},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"google-gemma-4-31b-it":{"id":"google-gemma-4-31b-it","name":"Google Gemma 4 31B Instruct","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.12,"output":0.36,"cache_read":0.09},"canonical_model_id":"google/gemma-4-31b-it"},"grok-build-0-1":{"id":"grok-build-0-1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}},"canonical_model_id":"xai/grok-build-0.1"},"deepseek-v4-1-flash":{"id":"deepseek-v4-1-flash","name":"DeepSeek V4.1 Flash","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.0075},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"openai-gpt-54":{"id":"openai-gpt-54","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":131072},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"gemini-3-8-flash":{"id":"gemini-3-8-flash","name":"Gemini 3.8 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.9375,"output":4.6875,"cache_read":0.09375},"canonical_model_id":"google/gemini-3.8-flash"},"aion-labs-aion-3-5":{"id":"aion-labs-aion-3-5","name":"Aion 3.5","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":3.75,"output":7.5,"cache_read":0.9375}},"claude-opus-4-8-fast":{"id":"claude-opus-4-8-fast","name":"Claude Opus 4.8 Fast","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-opus-4-8"},"google-gemma-3-27b-it":{"id":"google-gemma-3-27b-it","name":"Google Gemma 3 27B Instruct","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":16384},"cost":{"input":0.12,"output":0.2},"canonical_model_id":"google/gemma-3-27b-it"},"openai-gpt-56-sol-pro":{"id":"openai-gpt-56-sol-pro","name":"GPT-5.6 Sol Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"muse-wonder":{"id":"muse-wonder","name":"Muse Wonder","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-10-12","last_updated":"2026-10-15","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":5,"output":25,"cache_read":0.25}},"gemini-3-5-flash":{"id":"gemini-3-5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.55,"output":9.45,"cache_read":0.155,"cache_write":0.086},"canonical_model_id":"google/gemini-3.5-flash"},"minimax-m3-preview":{"id":"minimax-m3-preview","name":"MiniMax M3 Preview","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax-m3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"release_date":"2026-06-12","last_updated":"2026-06-13","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"cost":{"input":0.3,"output":1.2,"cache_read":0.06}},"google-gemma-4-26b-a4b-it":{"id":"google-gemma-4-26b-a4b-it","name":"Google Gemma 4 26B A4B Instruct","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.13,"output":0.4,"cache_read":0.05},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"z-ai-glm-5-turbo":{"id":"z-ai-glm-5-turbo","name":"GLM 5 Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32768},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"qwen-3-7-plus":{"id":"qwen-3-7-plus","name":"Qwen 3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":2,"cache_read":0.05,"cache_write":0.625,"tiers":[{"input":1.5,"output":6,"cache_read":0.15,"cache_write":1.875,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.5,"output":6,"cache_read":0.15,"cache_write":1.875}},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen-3-8-flash":{"id":"qwen-3-8-flash","name":"Qwen 3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.14,"output":0.49,"cache_read":0.014},"canonical_model_id":"alibaba/qwen3.8-flash"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":198000,"output":32768},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"openai-gpt-55-pro":{"id":"openai-gpt-55-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai-gpt-6-sol":{"id":"openai-gpt-6-sol","name":"GPT-6 Sol","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"qwen3-235b-a22b-thinking-2507":{"id":"qwen3-235b-a22b-thinking-2507","name":"Qwen 3 235B A22B Thinking 2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.45,"output":3.5},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"xiaomi-mimo-v2-6-flash":{"id":"xiaomi-mimo-v2-6-flash","name":"MiMo-V2.6-Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.175,"output":0.35,"cache_read":0.00375},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"z-ai-glm-5-3":{"id":"z-ai-glm-5-3","name":"GLM 5.3","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.75,"output":5.5,"cache_read":0.325},"canonical_model_id":"zhipuai/glm-5.3"},"claude-opus-5-5-fast":{"id":"claude-opus-5-5-fast","name":"Claude Opus 5.5 Fast","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":8,"output":40,"cache_read":0.4,"cache_write":10},"canonical_model_id":"anthropic/claude-opus-5-5"},"zai-org-glm-5-2":{"id":"zai-org-glm-5-2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"qwen-3-8-2-4t-a95b":{"id":"qwen-3-8-2-4t-a95b","name":"Qwen 3.8 2.4T","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":2.5,"output":7.5,"cache_read":0.3125},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"mistral-small-3-2-24b-instruct":{"id":"mistral-small-3-2-24b-instruct","name":"Mistral Small 3.2 24B Instruct","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.09375,"output":0.25},"canonical_model_id":"mistral/mistral-small-3.2-24b-instruct"},"grok-4-5":{"id":"grok-4-5","name":"Grok 4.5","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":32000},"cost":{"input":2.27,"output":6.8,"cache_read":0.34,"tiers":[{"input":4.53,"output":13.6,"cache_read":0.68,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4.53,"output":13.6,"cache_read":0.68}},"canonical_model_id":"xai/grok-4.5"},"minimax-m3-1-flash":{"id":"minimax-m3-1-flash","name":"MiniMax M3.1 Flash","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1875,"output":0.75,"cache_read":0.01875},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":0.13,"output":0.26,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"cost":{"input":0.7,"output":3.75,"cache_read":0.07},"canonical_model_id":"google/gemini-3-flash-preview"},"grok-4-3":{"id":"grok-4-3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":1.42,"output":2.83,"cache_read":0.23,"tiers":[{"input":2.83,"output":5.67,"cache_read":0.45,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.83,"output":5.67,"cache_read":0.45}},"canonical_model_id":"xai/grok-4.3"},"gemma-4-uncensored":{"id":"gemma-4-uncensored","name":"Gemma 4 Uncensored","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-04-13","last_updated":"2026-06-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.1625,"output":0.5}},"openai-gpt-6-astra-pro":{"id":"openai-gpt-6-astra-pro","name":"GPT-6 Astra Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai-gpt-53-codex":{"id":"openai-gpt-53-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"qwen3-5-397b-a17b":{"id":"qwen3-5-397b-a17b","name":"Qwen 3.5 397B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.75,"output":4.5},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"zai-org-glm-5-1":{"id":"zai-org-glm-5-1","name":"GLM 5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":80000},"cost":{"input":1.54,"output":4.84,"cache_read":0.286},"canonical_model_id":"zhipuai/glm-5.1"},"hy4-preview":{"id":"hy4-preview","name":"Hy4 Preview","description":"Preview model for early access evaluation, prototyping, and compatibility testing","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":1.0425,"output":3.12625,"cache_read":0.0525},"canonical_model_id":"tencent/hy4-preview"},"gemini-3-7-flash":{"id":"gemini-3-7-flash","name":"Gemini 3.7 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.9375,"output":4.6875,"cache_read":0.09375},"canonical_model_id":"google/gemini-3.7-flash"},"gemini-3-1-pro-preview":{"id":"gemini-3-1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":2.5,"output":15,"cache_read":0.5,"cache_write":0.5,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"cache_write":0.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5,"cache_write":0.5}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"zai-org-glm-4.7":{"id":"zai-org-glm-4.7","name":"GLM 4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":16384},"cost":{"input":0.55,"output":2.65,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"openai-gpt-61-sol":{"id":"openai-gpt-61-sol","name":"GPT-6.1 Sol","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"olafangensan-glm-4.7-flash-heretic":{"id":"olafangensan-glm-4.7-flash-heretic","name":"GLM 4.7 Flash Heretic","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-02-04","last_updated":"2026-06-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":24000},"cost":{"input":0.07,"output":0.4,"cache_read":0.035}},"openai-gpt-52":{"id":"openai-gpt-52","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":272000,"output":65536},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"inkling":{"id":"inkling","name":"Inkling","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":65536},"status":"deprecated","cost":{"input":0.5625,"output":1.5,"cache_read":0.125},"canonical_model_id":"thinkingmachines/inkling"},"llama-3.3-70b":{"id":"llama-3.3-70b","name":"Llama 3.3 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.7,"output":2.8},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"openai-gpt-6-astra":{"id":"openai-gpt-6-astra","name":"GPT-6 Astra","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai-gpt-56-terra":{"id":"openai-gpt-56-terra","name":"GPT-5.6 Terra","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"muse-spark-open":{"id":"muse-spark-open","name":"Muse Spark Open","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.5625,"output":5.3125,"cache_read":0.1875},"canonical_model_id":"meta/muse-spark-open"},"minimax-m25":{"id":"minimax-m25","name":"MiniMax M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":198000,"output":32768},"cost":{"input":0.27,"output":0.95,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"}}},"impossibl":{"id":"impossibl","env":["IMPOSSIBL_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.impossibl.com/v1","name":"Impossibl","doc":"https://impossibl.com/docs/models","models":{"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"groq/gpt-oss-20b":{"id":"groq/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-oss-20b"},"groq/gpt-oss-120b":{"id":"groq/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-oss-120b"},"cerebras/gpt-oss-120b":{"id":"cerebras/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.35,"output":0.75},"canonical_model_id":"openai/gpt-oss-120b"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":1.87,"output":4.68,"cache_read":0.374},"canonical_model_id":"thinkingmachines/inkling"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"cache_read":3,"tiers":[{"input":60,"output":270,"cache_read":3,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270,"cache_read":3}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"cache_read":3,"tiers":[{"input":60,"output":270,"cache_read":3,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270,"cache_read":3}},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-5-codex":{"id":"openai/gpt-5-codex","name":"GPT-5-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5-codex"},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5-turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.19,"output":0.51,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.74,"output":3.48,"cache_read":0.145},"canonical_model_id":"deepseek/deepseek-v4-pro"},"anthropic/claude-opus-4-8":{"id":"anthropic/claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-haiku-4-5":{"id":"anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-sonnet-4-5":{"id":"anthropic/claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-4-5":{"id":"anthropic/claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.003},"canonical_model_id":"xiaomi/mimo-v2.5"},"fireworks/gpt-oss-20b":{"id":"fireworks/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.07,"output":0.3,"cache_read":0.035},"canonical_model_id":"openai/gpt-oss-20b"},"fireworks/gpt-oss-120b":{"id":"fireworks/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"openai/gpt-oss-120b"},"fireworks/glm-5.2":{"id":"fireworks/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"xai/grok-4.5":{"id":"xai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"xai/grok-4.20-0309-non-reasoning":{"id":"xai/grok-4.20-0309-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"xai/grok-4.3":{"id":"xai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"xai/grok-4.20-0309-reasoning":{"id":"xai/grok-4.20-0309-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"xai/grok-build-0.1":{"id":"xai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}},"canonical_model_id":"xai/grok-build-0.1"},"zai/glm-4.5-air":{"id":"zai/glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.2,"output":1.1,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-4.5-air"},"zai/glm-5.1":{"id":"zai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai/glm-4.7":{"id":"zai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"zai/glm-5-turbo":{"id":"zai/glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"zai/glm-4.5":{"id":"zai/glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5"},"zai/glm-5.2":{"id":"zai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai/glm-5":{"id":"zai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"zai/glm-4.6":{"id":"zai/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.5},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.24,"tier":{"type":"context","size":262144}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.24}},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3.8-max-preview":{"id":"qwen/qwen3.8-max-preview","name":"Qwen3.8 Max Preview","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-19","last_updated":"2026-07-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":2.5,"output":7.5},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.05,"tiers":[{"input":1,"output":4,"cache_read":0.2,"tier":{"type":"context","size":262144}}],"context_over_200k":{"input":1,"output":4,"cache_read":0.2}},"canonical_model_id":"alibaba/qwen3.6-flash"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"}}},"kenari":{"id":"kenari","env":["KENARI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://kenari.id/v1","name":"Kenari","doc":"https://kenari.id/docs","models":{"gpt-5-6-luna":{"id":"gpt-5-6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0.2,"output":1.14285714,"cache_read":0.02,"cache_write":0.25714286},"canonical_model_id":"openai/gpt-5.6-luna"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.71428571,"output":29.71428571,"cache_read":0.57142857,"cache_write":7.42857143},"canonical_model_id":"anthropic/claude-opus-4-8"},"gpt-5-6-terra":{"id":"gpt-5-6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"status":"deprecated","cost":{"input":2.4,"output":14.28571429},"canonical_model_id":"openai/gpt-5.6-terra"},"nano-banana-2":{"id":"nano-banana-2","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","video","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"canonical_model_id":"google/gemini-3.1-flash-image"},"whisper-large-v3-turbo":{"id":"whisper-large-v3-turbo","type":"transcription","name":"Whisper Large v3 Turbo","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-10-01","last_updated":"2024-10-01","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":448},"canonical_model_id":"openai/whisper-large-v3-turbo"},"minimax-m2-7":{"id":"minimax-m2-7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.36,"output":1.42857143,"cache_read":0.06857143},"canonical_model_id":"minimax/MiniMax-M2.7"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.036,"output":0.15428571,"cache_read":0.00371429},"canonical_model_id":"openai/gpt-oss-20b"},"grok-4-6":{"id":"grok-4-6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2.28571429,"output":6.85714286,"cache_read":0.34285714},"canonical_model_id":"xai/grok-4.6"},"grok-imagine-image-2-0":{"id":"grok-imagine-image-2-0","type":"image","name":"Grok Imagine Image 2.0","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-08-07","last_updated":"2026-08-07","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":64000,"output":0},"canonical_model_id":"xai/grok-imagine-image-2.0"},"minimax-m2-7-highspeed":{"id":"minimax-m2-7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.59428571,"output":2.4,"cache_read":0.11428571},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.71428571,"output":29.71428571,"cache_read":0.57142857,"cache_write":7.42857143},"canonical_model_id":"anthropic/claude-opus-4-7"},"nemotron-3-super-120b-a12b":{"id":"nemotron-3-super-120b-a12b","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.09714286,"output":0.48,"cache_read":0.00971429},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"grok-imagine-image":{"id":"grok-imagine-image","type":"image","name":"Grok Imagine Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-28","last_updated":"2026-01-28","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":16000,"output":0},"canonical_model_id":"xai/grok-imagine-image"},"step-3-7-flash":{"id":"step-3-7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.24,"output":1.37142857,"cache_read":0.048},"canonical_model_id":"stepfun/step-3.7-flash"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15714286,"output":0.31428571,"cache_read":0.00371429},"canonical_model_id":"deepseek/deepseek-v4-flash"},"nemotron-3-ultra-550b-a55b":{"id":"nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.68571429,"output":4.28571429,"cache_read":0.06857143},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"gpt-5-6-sol":{"id":"gpt-5-6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2.97142857,"output":17.71428571,"cache_read":1.18857143,"cache_write":3.71428571},"canonical_model_id":"openai/gpt-5.6-sol"},"mimo-v2-5:free":{"id":"mimo-v2-5:free","name":"MiMo-V2.5 (Free)","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"qwen3-7-plus":{"id":"qwen3-7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.38285714,"output":1.48571429,"cache_read":0.07428571,"cache_write":0.48},"canonical_model_id":"alibaba/qwen3.7-plus"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3.6,"output":17.71428571,"cache_read":0.36,"cache_write":4.45714286},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"hy3:free":{"id":"hy3:free","name":"Hy3 (Free)","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"tencent/hy3"},"deepseek-v4-flash:free":{"id":"deepseek-v4-flash:free","name":"DeepSeek V4 Flash (Free)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"mimo-v2-5":{"id":"mimo-v2-5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.16571429,"output":0.33142857},"canonical_model_id":"xiaomi/mimo-v2.5"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.57142857,"output":1.14285714,"cache_read":0.00571429},"canonical_model_id":"deepseek/deepseek-v4-pro"},"glm-4-7-flash:free":{"id":"glm-4-7-flash:free","name":"GLM-4.7-Flash (Free)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"gemini-3-1-pro":{"id":"gemini-3-1-pro","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.2,"output":7.14285714,"cache_read":0.12,"cache_write":0.22285714},"canonical_model_id":"google/gemini-3.1-pro-preview"},"kimi-k2-6":{"id":"kimi-k2-6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.57142857,"output":2.85714286,"cache_read":0.28571429},"canonical_model_id":"moonshotai/kimi-k2.6"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":12,"output":57.14285714,"cache_read":1.2,"cache_write":14.85714286},"canonical_model_id":"anthropic/claude-fable-5"},"gpt-image-2":{"id":"gpt-image-2","type":"image","name":"GPT-Image-2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":272000,"output":16384},"canonical_model_id":"openai/gpt-image-2"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.036,"output":0.2,"cache_read":0.0036},"canonical_model_id":"openai/gpt-oss-120b"},"kimi-k2-7-code":{"id":"kimi-k2-7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.57142857,"output":2.85714286,"cache_read":0.11428571},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"gemini-2-5-flash":{"id":"gemini-2-5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","canonical_model_id":"google/gemini-2.5-flash"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.85714286,"output":8.57142857,"cache_read":0.11428571},"canonical_model_id":"moonshotai/kimi-k3"},"nemotron-3-nano-30b-a3b":{"id":"nemotron-3-nano-30b-a3b","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.71428571,"output":28.57142857,"cache_read":2.28571429,"cache_write":5.71428571},"canonical_model_id":"anthropic/claude-opus-5"},"glm-5-3-flash":{"id":"glm-5-3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.11428571,"output":0.34285714,"cache_read":0.02285714},"canonical_model_id":"zhipuai/glm-5.3-flash"},"gemini-2-5-flash-lite":{"id":"gemini-2-5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","canonical_model_id":"google/gemini-2.5-flash-lite"},"nano-banana-2-lite":{"id":"nano-banana-2-lite","type":"image","name":"Nano Banana 2 Lite","description":"Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":4096},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"kimi-k2-6:free":{"id":"kimi-k2-6:free","name":"Kimi K2.6 (Free)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"mimo-v2-5-pro":{"id":"mimo-v2-5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.52,"output":1.02857143},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"gemini-3-6-flash":{"id":"gemini-3-6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.34285714,"output":1.71428571,"cache_read":0.03428571,"cache_write":0.01714286},"canonical_model_id":"google/gemini-3.6-flash"},"grok-build-0-1":{"id":"grok-build-0-1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"status":"deprecated","canonical_model_id":"xai/grok-build-0.1"},"deepseek-v4-1-flash":{"id":"deepseek-v4-1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15714286,"output":0.31428571,"cache_read":0.00371429},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.36,"output":1.42857143,"cache_read":0.06857143},"canonical_model_id":"minimax/MiniMax-M3"},"gemini-3-5-flash":{"id":"gemini-3-5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","canonical_model_id":"google/gemini-3.5-flash"},"glm-5-3":{"id":"glm-5-3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.28571429,"output":3.42857143,"cache_read":0.05314286},"canonical_model_id":"zhipuai/glm-5.3"},"qwen3-8-max":{"id":"qwen3-8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.4,"output":6.85714286,"cache_read":0.29714286,"cache_write":2.97142857},"canonical_model_id":"alibaba/qwen3.8-max"},"gemini-3-1-flash-tts":{"id":"gemini-3-1-flash-tts","type":"speech","name":"Gemini 3.1 Flash TTS Preview","description":"Low-latency speech generation with steerable prompts and expressive audio tags","family":"gemini-flash","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-15","last_updated":"2026-04-15","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":16384},"canonical_model_id":"google/gemini-3.1-flash-tts-preview"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":1.42857143,"output":7.14285714,"cache_read":0.57142857,"cache_write":1.71428571},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-5-4-mini":{"id":"gpt-5-4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":0.85714286,"output":5.37142857},"canonical_model_id":"openai/gpt-5.4-mini"},"glm-5-2":{"id":"glm-5-2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.28571429,"output":3.42857143,"cache_read":0.05314286},"canonical_model_id":"zhipuai/glm-5.2"},"nemotron-3-super-120b-a12b:free":{"id":"nemotron-3-super-120b-a12b:free","name":"Nemotron 3 Super 120B A12B (Free)","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"grok-4-5":{"id":"grok-4-5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2.28571429,"output":6.85714286,"cache_read":0.34285714},"canonical_model_id":"xai/grok-4.5"},"nano-banana-pro":{"id":"nano-banana-pro","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"canonical_model_id":"google/gemini-3-pro-image"},"mistral-large:free":{"id":"mistral-large:free","name":"Mistral Large (Free)","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"mistral/mistral-large-latest"},"gemma-4-31b-it":{"id":"gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.10285714,"output":0.40571429,"cache_read":0.04114286},"canonical_model_id":"google/gemma-4-31b-it"},"mistral-medium-3-5:free":{"id":"mistral-medium-3-5:free","name":"Mistral Medium 3.5 (Free)","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"mistral/mistral-medium-2604"},"step-3-7-flash:free":{"id":"step-3-7-flash:free","name":"Step 3.7 Flash (Free)","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"stepfun/step-3.7-flash"},"gemini-3-7-flash":{"id":"gemini-3-7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.34285714,"output":1.71428571,"cache_read":0.03428571,"cache_write":0.01714286},"canonical_model_id":"google/gemini-3.7-flash"},"hy3":{"id":"hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0.15428571,"output":0.62857143,"cache_read":0.01542857},"canonical_model_id":"tencent/hy3"},"kimi-k2-7-code:free":{"id":"kimi-k2-7-code:free","name":"Kimi K2.7 Code (Free)","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"glm-5-1":{"id":"glm-5-1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"status":"deprecated","canonical_model_id":"zhipuai/glm-5.1"},"gpt-5-5":{"id":"gpt-5-5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5.71428571,"output":36,"cache_read":0.57142857},"canonical_model_id":"openai/gpt-5.5"},"gemini-3-1-flash-lite":{"id":"gemini-3-1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.28571429,"output":1.71428571,"cache_read":0.02857143,"cache_write":0.08571429},"canonical_model_id":"google/gemini-3.1-flash-lite"}}},"perplexity":{"id":"perplexity","env":["PERPLEXITY_API_KEY"],"npm":"@ai-sdk/perplexity","name":"Perplexity","doc":"https://docs.perplexity.ai","models":{"sonar-reasoning-pro":{"id":"sonar-reasoning-pro","name":"Sonar Reasoning Pro","description":"Web-grounded Sonar for multi-step research questions that need cited reasoning","family":"sonar-reasoning","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":2,"output":8},"canonical_model_id":"perplexity/sonar-reasoning-pro"},"sonar-deep-research":{"id":"sonar-deep-research","name":"Perplexity Sonar Deep Research","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-02-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"status":"deprecated","cost":{"input":2,"output":8,"reasoning":3},"canonical_model_id":"perplexity/sonar-deep-research"},"sonar-pro":{"id":"sonar-pro","name":"Sonar Pro","description":"Deeper Sonar search model with broader retrieval and stronger synthesis","family":"sonar-pro","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"status":"deprecated","cost":{"input":3,"output":15},"canonical_model_id":"perplexity/sonar-pro"},"sonar":{"id":"sonar","name":"Sonar","description":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","family":"sonar","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":1,"output":1},"canonical_model_id":"perplexity/sonar"}}},"regolo-ai":{"id":"regolo-ai","env":["REGOLO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.regolo.ai/v1","name":"Regolo AI","doc":"https://docs.regolo.ai/","models":{"apertus-70b":{"id":"apertus-70b","name":"Apertus 70B","description":"Fully open 70B multilingual LLM supporting 1800+ languages with 65K context. Trained on 15T tokens of compliant open data. Apache 2.0, EU AI Act compliant.","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":30000,"output":30000},"cost":{"input":0.46,"output":2.42},"canonical_model_id":"swiss-ai/apertus-70b"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT-OSS-20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.11215,"output":0.47103},"canonical_model_id":"openai/gpt-oss-20b"},"faster-whisper-large-v3":{"id":"faster-whisper-large-v3","type":"transcription","name":"Faster Whisper Large v3","description":"Open Whisper checkpoint for robust multilingual transcription and captioning","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"openai/whisper-large-v3"},"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":120000,"output":120000},"cost":{"input":0.58,"output":2.42},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen3-coder-next":{"id":"qwen3-coder-next","name":"Qwen3-Coder-Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"status":"deprecated","cost":{"input":0.3,"output":1.2},"canonical_model_id":"alibaba/qwen3-coder-next"},"qwen3-embedding-8b":{"id":"qwen3-embedding-8b","type":"embedding","name":"Qwen3-Embedding-8B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0.1,"output":0.1},"canonical_model_id":"alibaba/qwen3-embedding-8b"},"brick-v1-beta":{"id":"brick-v1-beta","name":"Brick v1 Beta","description":"Semantic router by Regolo.ai that directs each request to the most suitable model, optimizing costs and performance","family":"model-router","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-02-06","last_updated":"2026-02-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":100000,"output":15000},"status":"beta","cost":{"input":0,"output":0}},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT-OSS-120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":1.1215,"output":4.7103},"canonical_model_id":"openai/gpt-oss-120b"},"mistral-small-4-119b":{"id":"mistral-small-4-119b","name":"Mistral Small 4 119B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.56075,"output":2.35515},"canonical_model_id":"mistral/mistral-small-2603"},"qwen-image":{"id":"qwen-image","type":"image","name":"Qwen-Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-03-01","last_updated":"2026-03-01","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":4096},"cost":{"input":0.5,"output":2}},"qwen3.5-122b":{"id":"qwen3.5-122b","name":"Qwen3.5-122B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":1.1215,"output":4.7103},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen3-reranker-4b":{"id":"qwen3-reranker-4b","type":"reranking","name":"Qwen3-Reranker-4B","description":"Reranking model for improving retrieval quality in search and recommendation systems","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-01","last_updated":"2026-02-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0.12,"output":0.12}},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama 3.3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":0.6,"output":2.7},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"glm5.2":{"id":"glm5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":96000,"output":96000},"cost":{"input":2.31,"output":6},"canonical_model_id":"zhipuai/glm-5.2"},"gemma4-31b":{"id":"gemma4-31b","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":100000,"output":100000},"cost":{"input":0.46,"output":2.42},"canonical_model_id":"google/gemma-4-31b-it"},"brick-complexity-pro":{"id":"brick-complexity-pro","name":"Brick Complexity Pro","description":"Complexity classifier that powers the Brick semantic router by extracting query difficulty","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-02-06","last_updated":"2026-02-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":100000,"output":15000},"cost":{"input":0.12,"output":0.46}},"qwen3.5-9b":{"id":"qwen3.5-9b","name":"Qwen3.5-9B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.078505,"output":0.392525},"canonical_model_id":"alibaba/qwen3.5-9b"},"deepseek-ocr-2":{"id":"deepseek-ocr-2","name":"DeepSeek OCR 2","description":"High-accuracy OCR model for extracting text from documents, screenshots, receipts, and natural scenes","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":4000,"output":4000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-ocr-2"}}},"neuralwatt":{"id":"neuralwatt","env":["NEURALWATT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.neuralwatt.com/v1","name":"Neuralwatt","doc":"https://portal.neuralwatt.com/docs","models":{"deepseek-v4-flash-flex":{"id":"deepseek-v4-flash-flex","name":"DeepSeek V4 Flash Flex","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"cost":{"input":0.091,"output":0.182,"cache_read":0.0182},"canonical_model_id":"deepseek/deepseek-v4-flash"},"kimi-k2.7-code-fast":{"id":"kimi-k2.7-code-fast","name":"Kimi K2.7 Code Fast","description":"Kimi K2.7 Code with reasoning capped to a short budget for lower latency; reasoning cannot be disabled on this model","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262128,"output":262128},"status":"deprecated","cost":{"input":0.95,"output":4,"cache_read":0.095},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"glm-5.2-short-fast":{"id":"glm-5.2-short-fast","name":"GLM 5.2 Short Fast","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":199984,"output":32000},"status":"deprecated","cost":{"input":1.45,"output":4.5,"cache_read":0.145},"canonical_model_id":"zhipuai/glm-5.2"},"qwen3.6-35b-fast":{"id":"qwen3.6-35b-fast","name":"Qwen3.6 35B Fast","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131056,"output":131056},"cost":{"input":0.29,"output":1.15,"cache_read":0.029},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"kimi-k3-flex":{"id":"kimi-k3-flex","name":"Kimi K3 Flex","description":"Kimi K3 on the flex tier: discounted, best-effort latency, requests may be held under load","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"cost":{"input":1.95,"output":9.75,"cache_read":0.195},"canonical_model_id":"moonshotai/kimi-k3"},"mimo-v2.6-pro-flex":{"id":"mimo-v2.6-pro-flex","name":"MiMo-V2.6-Pro Flex","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"cost":{"input":0.5655,"output":1.131,"cache_read":0.0234},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"glm-5.3-flex":{"id":"glm-5.3-flex","name":"GLM 5.3 Flex","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"cost":{"input":0.9425,"output":2.925,"cache_read":0.09425},"canonical_model_id":"zhipuai/glm-5.3"},"glm-5.2-flex":{"id":"glm-5.2-flex","name":"GLM 5.2 Flex","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"status":"deprecated","cost":{"input":0.9425,"output":2.925,"cache_read":0.09425},"canonical_model_id":"zhipuai/glm-5.2"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"cost":{"input":0.87,"output":1.74,"cache_read":0.036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"status":"deprecated","cost":{"input":1,"output":3,"cache_read":0.1},"canonical_model_id":"deepseek/deepseek-v4-pro"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262128,"output":262128},"status":"deprecated","cost":{"input":0.95,"output":4,"cache_read":0.095},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"kimi-k2.7-code-flex":{"id":"kimi-k2.7-code-flex","name":"Kimi K2.7 Code Flex","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262128,"output":262128},"status":"deprecated","cost":{"input":0.6175,"output":2.6,"cache_read":0.06175},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"deepseek-v4.1-flash-flex":{"id":"deepseek-v4.1-flash-flex","name":"DeepSeek V4.1 Flash Flex","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"cost":{"input":0.0975,"output":0.39,"cache_read":0.00975},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"glm-5.3-flash-flex":{"id":"glm-5.3-flash-flex","name":"GLM-5.3 Flash Flex","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"cost":{"input":0.0975,"output":0.325,"cache_read":0.0195},"canonical_model_id":"zhipuai/glm-5.3-flash"},"gemma-4-31b":{"id":"gemma-4-31b","name":"Gemma 4 31B","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262128,"output":16384},"cost":{"input":0.144,"output":0.42,"cache_read":0.0144},"canonical_model_id":"google/gemma-4-31b-it"},"glm-5.2-short":{"id":"glm-5.2-short","name":"GLM 5.2 Short","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":199984,"output":32000},"status":"deprecated","cost":{"input":1.45,"output":4.5,"cache_read":0.145},"canonical_model_id":"zhipuai/glm-5.2"},"qwen3.6-35b":{"id":"qwen3.6-35b","name":"Qwen3.6 35B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131056,"output":131056},"cost":{"input":0.29,"output":1.15,"cache_read":0.029},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen3.6-35b-flex":{"id":"qwen3.6-35b-flex","name":"Qwen3.6 35B Flex","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131056,"output":131056},"cost":{"input":0.1885,"output":0.7475,"cache_read":0.01885},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen-3.8-27b":{"id":"qwen-3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262128,"output":131072},"cost":{"input":0.45,"output":3.2,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-27b"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5.2-short-flex":{"id":"glm-5.2-short-flex","name":"GLM 5.2 Short Flex","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":199984,"output":32000},"status":"deprecated","cost":{"input":0.9425,"output":2.925,"cache_read":0.09425},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5.2":{"id":"glm-5.2","name":"GLM 5.2","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"status":"deprecated","cost":{"input":1.45,"output":4.5,"cache_read":0.145},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek-v4-flash-speed":{"id":"deepseek-v4-flash-speed","name":"DeepSeek V4 Flash (Speed)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":393216},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"kimi-k3-fast":{"id":"kimi-k3-fast","name":"Kimi K3 Fast","description":"Kimi K3 with thinking disabled for low-latency tool calling, vision, and JSON work","family":"kimi-k3","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"glm-5.2-short-fast-flex":{"id":"glm-5.2-short-fast-flex","name":"GLM 5.2 Short Fast Flex","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":199984,"output":32000},"status":"deprecated","cost":{"input":0.9425,"output":2.925,"cache_read":0.09425},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5.3":{"id":"glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"cost":{"input":1.45,"output":4.5,"cache_read":0.145},"canonical_model_id":"zhipuai/glm-5.3"},"qwen-3.8-27b-flex":{"id":"qwen-3.8-27b-flex","name":"Qwen3.8 27B Flex","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262128,"output":131072},"cost":{"input":0.2925,"output":2.08,"cache_read":0.1625},"canonical_model_id":"alibaba/qwen3.8-27b"},"glm-5.2-fast":{"id":"glm-5.2-fast","name":"GLM 5.2 Fast","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048560,"output":1048560},"status":"deprecated","cost":{"input":1.45,"output":4.5,"cache_read":0.145},"canonical_model_id":"zhipuai/glm-5.2"}}},"ofox":{"id":"ofox","env":["OFOX_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.ofox.ai/v1","name":"Ofox","doc":"https://ofox.ai/docs","models":{"volcengine/doubao-seed-2.0-pro":{"id":"volcengine/doubao-seed-2.0-pro","name":"Seed 2.0 Pro","description":"Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.67,"output":3.36,"cache_read":0.14,"cache_write":0.0024},"canonical_model_id":"bytedance-seed/seed-2.0-pro"},"volcengine/doubao-seed-2.0-code":{"id":"volcengine/doubao-seed-2.0-code","name":"Seed 2.0 Code","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.67,"output":3.36,"cache_read":0.14,"cache_write":0.0024},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"volcengine/doubao-seed-1-6-flash":{"id":"volcengine/doubao-seed-1-6-flash","name":"Seed 1.6 Flash","description":"Low-latency ByteDance Seed model for high-throughput chat, extraction, and lightweight tool use","family":"seed","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"status":"deprecated","cost":{"input":0.03,"output":0.22,"cache_read":0.0043},"canonical_model_id":"bytedance-seed/seed-1-6-flash"},"volcengine/doubao-seed-character":{"id":"volcengine/doubao-seed-character","name":"Seed Character","description":"ByteDance Seed model optimized for character-driven dialogue and consistent conversational behavior","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.177,"output":0.884,"cache_read":0.024,"cache_write":0.0025},"canonical_model_id":"bytedance-seed/seed-character"},"volcengine/doubao-seed-2.0-lite":{"id":"volcengine/doubao-seed-2.0-lite","name":"Seed 2.0 Lite","description":"Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.13,"output":0.76,"cache_read":0.03,"cache_write":0.0024},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"volcengine/doubao-seed-evolving":{"id":"volcengine/doubao-seed-evolving","name":"Seed Evolving","description":"Rolling ByteDance Seed model for rapidly updated reasoning, coding, and agent capabilities","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.884,"output":4.42,"cache_read":0.177,"cache_write":0.0025},"canonical_model_id":"bytedance-seed/seed-evolving"},"volcengine/doubao-seed-1-6-vision":{"id":"volcengine/doubao-seed-1-6-vision","name":"Seed 1.6 Vision","description":"ByteDance Seed multimodal model for image understanding, visual reasoning, and tool-assisted tasks","family":"seed","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-15","last_updated":"2025-08-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"status":"deprecated","cost":{"input":0.12,"output":1.15,"cache_read":0.023},"canonical_model_id":"bytedance-seed/seed-1-6-vision"},"volcengine/doubao-seed-1-6":{"id":"volcengine/doubao-seed-1-6","name":"Seed 1.6","description":"ByteDance Seed model for long-context reasoning, instruction following, and tool-assisted tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0.12,"output":0.29,"cache_read":0.023},"canonical_model_id":"bytedance-seed/seed-1-6"},"volcengine/doubao-seed-2.1-turbo":{"id":"volcengine/doubao-seed-2.1-turbo","name":"Seed 2.1 Turbo","description":"Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.3536,"output":1.7696,"cache_read":0.068,"cache_write":0.0019},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"volcengine/doubao-seed-1-8":{"id":"volcengine/doubao-seed-1-8","name":"Seed 1.8","description":"ByteDance Seed model for multimodal reasoning, long-context analysis, and agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-28","last_updated":"2025-12-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0.12,"output":0.29,"cache_read":0.023},"canonical_model_id":"bytedance-seed/seed-1-8"},"volcengine/doubao-seed-2.1-pro":{"id":"volcengine/doubao-seed-2.1-pro","name":"Seed 2.1 Pro","description":"Flagship ByteDance Seed 2.1 model for complex multimodal reasoning, coding, and agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.7072,"output":3.536,"cache_read":0.1416,"cache_write":0.002},"canonical_model_id":"bytedance-seed/seed-2.1-pro"},"volcengine/doubao-seed-2.0-mini":{"id":"volcengine/doubao-seed-2.0-mini","name":"Seed 2.0 Mini","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.06,"output":0.56,"cache_read":0.02,"cache_write":0.0024},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"z-ai/glm-4.7-flashx":{"id":"z-ai/glm-4.7-flashx","name":"GLM-4.7-FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.072,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.4,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-5.3-flashx":{"id":"z-ai/glm-5.3-flashx","name":"GLM-5.3-FlashX","description":"High-speed GLM-5.3-Flash serving option for coding and agent workflows","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-4.6":{"id":"z-ai/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-5v-turbo":{"id":"z-ai/glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k2.7-code-highspeed":{"id":"moonshotai/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32768},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.25,"output":2,"cache_read":0.03},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.18},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.18},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.05,"output":0.4,"cache_read":0.01},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.18},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":3.125,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"cache_write":6.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5,"cache_write":6.25}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5.1-codex-max":{"id":"openai/gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5.1-codex-max"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.25,"output":2,"cache_read":0.03},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.ofox.ai/v1"},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.29,"output":0.43,"cache_read":0.06},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4-pro-0423":{"id":"deepseek/deepseek-v4-pro-0423","name":"DeepSeek V4 Pro 0423","description":"DeepSeek V4 Pro initial snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.15},"canonical_model_id":"deepseek/deepseek-v4-pro-0423"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-v4-flash-0423":{"id":"deepseek/deepseek-v4-flash-0423","name":"DeepSeek V4 Flash 0423","description":"Initial DeepSeek V4 Flash snapshot for economical reasoning, coding, and million-token agent workloads","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.19,"output":0.51,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0423"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.ofox.ai/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"minimax/minimax-m2.1-lightning":{"id":"minimax/minimax-m2.1-lightning","name":"MiniMax-M2.1 Lightning","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":2.4,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/m2-her":{"id":"minimax/m2-her","name":"MiniMax-M2 Her","description":"MiniMax M2 variant tuned for conversational and character-driven agent interactions","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-01-23","last_updated":"2026-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":2048},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2-Her"},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax-M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-m2.5-lightning":{"id":"minimax/minimax-m2.5-lightning","name":"MiniMax-M2.5 Lightning","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":2.4,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5-highspeed"},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.29,"output":1.83,"cache_read":0.29},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1064000,"output":64000},"cost":{"input":1.71,"output":5.14,"cache_read":0.17,"cache_write":2.14},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1064000,"output":64000},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3-max":{"id":"qwen/qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":0.36,"output":1.43,"cache_read":0.072},"canonical_model_id":"alibaba/qwen3-max"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1131072,"output":131072},"cost":{"input":0.5,"output":1.71,"cache_read":0.043,"cache_write":0.63},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen-max":{"id":"qwen/qwen-max","name":"Qwen Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":8000},"cost":{"input":0.35,"output":1.38,"cache_read":0.069},"canonical_model_id":"alibaba/qwen-max"},"qwen/qwen-flash":{"id":"qwen/qwen-flash","name":"Qwen Flash","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":0.022,"output":0.22,"cache_read":0.0043,"cache_write":0.027},"canonical_model_id":"alibaba/qwen-flash"},"qwen/qwen3-coder-plus":{"id":"qwen/qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Hosted Qwen coder for software agents, repo edits, and long-context code","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":1.8,"output":9,"cache_read":0.2,"cache_write":1},"canonical_model_id":"alibaba/qwen3-coder-plus"},"qwen/qwen-vl-max":{"id":"qwen/qwen-vl-max","name":"Qwen-VL Max","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-08","last_updated":"2025-08-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8000},"cost":{"input":0.23,"output":0.58,"cache_read":0.023},"canonical_model_id":"alibaba/qwen-vl-max"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.29,"output":2.29,"cache_read":0.29},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3-coder-next":{"id":"qwen/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.2,"output":1.5},"canonical_model_id":"alibaba/qwen3-coder-next"},"qwen/qwen3.5-flash":{"id":"qwen/qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"alibaba/qwen3.5-flash"},"qwen/qwen3.8-max-0902":{"id":"qwen/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.71,"output":5.14,"cache_read":0.17,"cache_write":2.14},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"qwen/qwen3.6-max-preview":{"id":"qwen/qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"cost":{"input":2.15,"output":12.86,"cache_read":0.2,"cache_write":1.17},"canonical_model_id":"alibaba/qwen3.6-max-preview"},"qwen/qwen3.8-max":{"id":"qwen/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.71,"output":5.14,"cache_read":0.17,"cache_write":2.14},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.29,"output":2.05,"cache_read":0.29},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.31},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.55,"output":3.5,"cache_read":0.55},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen-turbo":{"id":"qwen/qwen-turbo","name":"Qwen Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-11-01","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000},"cost":{"input":0.043,"output":0.09,"cache_read":0.0086},"canonical_model_id":"alibaba/qwen-turbo"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.11,"output":0.39,"cache_read":0.011,"cache_write":0.14},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3.5-plus":{"id":"qwen/qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":2.4,"cache_read":0.04,"cache_write":0.4},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen/qwen-plus":{"id":"qwen/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":0.12,"output":0.29,"cache_read":0.023},"canonical_model_id":"alibaba/qwen-plus"},"qwen/qwen3-coder-flash":{"id":"qwen/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.5,"output":2.5,"cache_read":0.06,"cache_write":0.27},"canonical_model_id":"alibaba/qwen3-coder-flash"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.43,"output":2.57},"canonical_model_id":"alibaba/qwen3.6-27b"},"bailian/qwen3.5-35b-a3b":{"id":"bailian/qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.29,"output":1.83,"cache_read":0.29},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"bailian/qwen3.7-max":{"id":"bailian/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.5,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.7-max"},"bailian/qwen3.7-plus":{"id":"bailian/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen3.7-plus"},"bailian/qwen3-max":{"id":"bailian/qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.36,"output":1.43,"cache_read":0.072},"canonical_model_id":"alibaba/qwen3-max"},"bailian/qwen3.8-27b":{"id":"bailian/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1131072,"output":131072},"cost":{"input":0.45,"output":3.2,"cache_read":0.05,"cache_write":0.5625},"canonical_model_id":"alibaba/qwen3.8-27b"},"bailian/qwen-max":{"id":"bailian/qwen-max","name":"Qwen Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":0.35,"output":1.38,"cache_read":0.069},"canonical_model_id":"alibaba/qwen-max"},"bailian/qwen-flash":{"id":"bailian/qwen-flash","name":"Qwen Flash","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.022,"output":0.22,"cache_read":0.0043,"cache_write":0.027},"canonical_model_id":"alibaba/qwen-flash"},"bailian/qwen3-coder-plus":{"id":"bailian/qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Hosted Qwen coder for software agents, repo edits, and long-context code","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.8,"output":9,"cache_read":0.2,"cache_write":1},"canonical_model_id":"alibaba/qwen3-coder-plus"},"bailian/qwen-vl-max":{"id":"bailian/qwen-vl-max","name":"Qwen-VL Max","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-08","last_updated":"2025-08-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.23,"output":0.58,"cache_read":0.046},"canonical_model_id":"alibaba/qwen-vl-max"},"bailian/qwen3.5-122b-a10b":{"id":"bailian/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.29,"output":2.29,"cache_read":0.29},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"bailian/qwen3-coder-next":{"id":"bailian/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.2,"output":1.5},"canonical_model_id":"alibaba/qwen3-coder-next"},"bailian/qwen3.5-flash":{"id":"bailian/qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"alibaba/qwen3.5-flash"},"bailian/qwen3.8-max-0902":{"id":"bailian/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"bailian/qwen3.6-max-preview":{"id":"bailian/qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":2.15,"output":12.86,"cache_read":0.2,"cache_write":1.17},"canonical_model_id":"alibaba/qwen3.6-max-preview"},"bailian/qwen3.8-max":{"id":"bailian/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"bailian/qwen3.5-27b":{"id":"bailian/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.29,"output":2.05,"cache_read":0.29},"canonical_model_id":"alibaba/qwen3.5-27b"},"bailian/qwen3.6-flash":{"id":"bailian/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.31},"canonical_model_id":"alibaba/qwen3.6-flash"},"bailian/qwen3.5-397b-a17b":{"id":"bailian/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.55,"output":3.5,"cache_read":0.55},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"bailian/qwen-turbo":{"id":"bailian/qwen-turbo","name":"Qwen Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-11-01","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.05,"output":0.09,"cache_read":0.0086},"canonical_model_id":"alibaba/qwen-turbo"},"bailian/qwen3.8-flash":{"id":"bailian/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"bailian/qwen3.5-plus":{"id":"bailian/qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.4,"output":2.4,"cache_read":0.04,"cache_write":0.4},"canonical_model_id":"alibaba/qwen3.5-plus"},"bailian/qwen3.6-plus":{"id":"bailian/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625},"canonical_model_id":"alibaba/qwen3.6-plus"},"bailian/qwen-plus":{"id":"bailian/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.12,"output":0.29,"cache_read":0.023},"canonical_model_id":"alibaba/qwen-plus"},"bailian/qwen3-coder-flash":{"id":"bailian/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":2.5,"cache_read":0.06,"cache_write":0.27},"canonical_model_id":"alibaba/qwen3-coder-flash"},"bailian/qwen3.6-27b":{"id":"bailian/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.6-27b"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://api.ofox.ai/gemini/v1beta"},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.0415,"input_audio":0.75},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":1,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":1,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://api.ofox.ai/gemini/v1beta"},"cost":{"input":2,"output":10,"cache_read":0.1,"input_audio":2},"canonical_model_id":"google/gemini-4-argon"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://api.ofox.ai/gemini/v1beta"},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.083,"input_audio":0.3},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://api.ofox.ai/gemini/v1beta"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":4.5,"input_audio":2},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://api.ofox.ai/gemini/v1beta"},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.0415,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://api.ofox.ai/gemini/v1beta"},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.0415,"input_audio":0.75},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://api.ofox.ai/gemini/v1beta"},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":0.083,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":1,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":1,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"cache_write":4.5,"input_audio":1.25},"canonical_model_id":"google/gemini-2.5-pro"},"x-ai/grok-4.5":{"id":"x-ai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":65536},"cost":{"input":2,"output":6,"cache_read":0.3},"canonical_model_id":"xai/grok-4.5"},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":65536},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"x-ai/grok-4.7":{"id":"x-ai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":65536},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"x-ai/grok-4.1-fast":{"id":"x-ai/grok-4.1-fast","name":"Grok 4.1 Fast","description":"xAI's fast agentic tool-calling model with a 2M context window; non-reasoning variant for low-latency responses","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast"},"x-ai/grok-4.20":{"id":"x-ai/grok-4.20","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":128000},"cost":{"input":4,"output":12,"cache_read":0.4},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"}}},"iflowcn":{"id":"iflowcn","env":["IFLOW_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://apis.iflow.cn/v1","name":"iFlow","doc":"https://platform.iflow.cn/en/docs","models":{"qwen3-235b-a22b-instruct":{"id":"qwen3-235b-a22b-instruct","name":"Qwen3-235B-A22B-Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"qwen3-max":{"id":"qwen3-max","name":"Qwen3-Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-max"},"qwen3-coder-plus":{"id":"qwen3-coder-plus","name":"Qwen3-Coder-Plus","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-coder-plus"},"kimi-k2-0905":{"id":"kimi-k2-0905","name":"Kimi-K2-0905","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"moonshotai/kimi-k2-0905"},"qwen3-vl-plus":{"id":"qwen3-vl-plus","name":"Qwen3-VL-Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-vl-plus"},"deepseek-v3":{"id":"deepseek-v3","name":"DeepSeek-V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v3"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek-V3.2-Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":64000},"status":"deprecated","cost":{"input":0,"output":0}},"deepseek-r1":{"id":"deepseek-r1","name":"DeepSeek-R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-r1"},"qwen3-max-preview":{"id":"qwen3-max-preview","name":"Qwen3-Max-Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2025-09-05","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-max-preview"},"qwen3-32b":{"id":"qwen3-32b","name":"Qwen3-32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-32b"},"qwen3-235b-a22b-thinking-2507":{"id":"qwen3-235b-a22b-thinking-2507","name":"Qwen3-235B-A22B-Thinking","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"qwen3-235b":{"id":"qwen3-235b","name":"Qwen3-235B-A22B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-235b-a22b"},"glm-4.6":{"id":"glm-4.6","name":"GLM-4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.6"},"kimi-k2":{"id":"kimi-k2","name":"Kimi-K2","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":64000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"moonshotai/kimi-k2"}}},"aki-io":{"id":"aki-io","env":["AKI_IO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://aki.io/v1","name":"AKI.IO","doc":"https://aki.io/docs/","models":{"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.33645,"output":2.4673,"cache_read":0.11215},"canonical_model_id":"alibaba/qwen3.8-27b"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.168225,"output":0.616825},"canonical_model_id":"openai/gpt-oss-120b"},"mistral4-119b":{"id":"mistral4-119b","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":81920},"cost":{"input":0.2243,"output":0.6729},"canonical_model_id":"mistral/mistral-small-2603"},"qwen3.6-35b":{"id":"qwen3.6-35b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"status":"deprecated","cost":{"input":0.168225,"output":0.56075},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"gemma4-26b":{"id":"gemma4-26b","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"cost":{"input":0.11215,"output":0.56075},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"deepseek-v4.1-flash-552b":{"id":"deepseek-v4.1-flash-552b","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":81920},"cost":{"input":0.280375,"output":1.1215,"cache_read":0.11215},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"glm5.3-754b":{"id":"glm5.3-754b","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":81920},"cost":{"input":1.1215,"output":3.92525,"cache_read":0.280375},"canonical_model_id":"zhipuai/glm-5.3"}}},"privatemode-ai":{"id":"privatemode-ai","env":["PRIVATEMODE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"http://localhost:8080/v1","name":"Privatemode AI","doc":"https://docs.privatemode.ai/api/overview","models":{"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":262144},"status":"deprecated","cost":{"input":1.791,"output":8.9436,"cache_read":0.1733},"canonical_model_id":"moonshotai/kimi-k2.6"},"glm-latest":{"id":"glm-latest","name":"GLM (latest)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.791,"output":6.6326,"cache_read":0.1733},"canonical_model_id":"zhipuai/glm-5.3"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"gpt-oss-120b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.2311,"output":0.7511,"cache_read":0.0462},"canonical_model_id":"openai/gpt-oss-120b"},"qwen3-embedding-4b":{"id":"qwen3-embedding-4b","type":"embedding","name":"Qwen3-Embedding 4B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-06","release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":2560},"cost":{"input":0.1502,"output":0},"canonical_model_id":"alibaba/qwen3-embedding-4b"},"kimi-latest":{"id":"kimi-latest","name":"Kimi (latest)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":262144},"status":"deprecated","cost":{"input":1.791,"output":8.9436,"cache_read":0.1733},"canonical_model_id":"moonshotai/kimi-k2.6"},"whisper-large-v3":{"id":"whisper-large-v3","type":"transcription","name":"Whisper large-v3","description":"Open Whisper checkpoint for robust multilingual transcription and captioning","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":4096},"cost":{"input":0.01618,"output":0},"canonical_model_id":"openai/whisper-large-v3"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.2311,"output":0.7511,"cache_read":0.0578},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.791,"output":6.6326,"cache_read":0.1733},"canonical_model_id":"zhipuai/glm-5.3"},"deepseek-ocr-2":{"id":"deepseek-ocr-2","name":"DeepSeek OCR 2","description":"High-accuracy OCR model for extracting text from documents, screenshots, receipts, and natural scenes","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"status":"beta","cost":{"input":0.8897,"output":1.4675,"cache_read":0.0924},"canonical_model_id":"deepseek/deepseek-ocr-2"},"glm-flash-latest":{"id":"glm-flash-latest","name":"GLM Flash (latest)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.2311,"output":0.7511,"cache_read":0.0578},"canonical_model_id":"zhipuai/glm-5.3-flash"},"voxtral-mini-3b":{"id":"voxtral-mini-3b","type":"transcription","name":"Voxtral Mini 3B","description":"Speech-to-text model for audio transcription, translation, and audio understanding","family":"voxtral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000},"cost":{"input":0.00462,"output":0},"canonical_model_id":"mistral/voxtral-mini-3b-2507"}}},"stepfun-ai":{"id":"stepfun-ai","env":["STEPFUN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.stepfun.ai/v1","name":"StepFun (Global)","doc":"https://platform.stepfun.ai/docs/en/overview/concept","models":{"step-tts-2":{"id":"step-tts-2","type":"speech","name":"Step TTS 2","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"step","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-01","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"step-2-16k":{"id":"step-2-16k","name":"Step 2 (16K)","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-01-01","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"input":16384,"output":8192},"status":"deprecated","cost":{"input":5.21,"output":16.44,"cache_read":1.04}},"step-3.7-flash":{"id":"step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"stepaudio-2.5-asr":{"id":"stepaudio-2.5-asr","type":"transcription","name":"StepAudio 2.5 ASR","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"step","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-24","last_updated":"2026-07-02","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0}},"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"input":1024000,"output":65536},"cost":{"input":1,"output":2.7,"cache_read":0.1},"canonical_model_id":"stepfun/step-5-preview"},"step-3.5-flash":{"id":"step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.1,"output":0.3,"cache_read":0.02},"canonical_model_id":"stepfun/step-3.5-flash"},"step-3.5-flash-2603":{"id":"step-3.5-flash-2603","name":"Step 3.5 Flash 2603","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.1,"output":0.3,"cache_read":0.02},"canonical_model_id":"stepfun/step-3.5-flash-2603"},"stepaudio-2.5-tts":{"id":"stepaudio-2.5-tts","type":"speech","name":"StepAudio 2.5 TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"step","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-16","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":0,"output":0}},"step-1-32k":{"id":"step-1-32k","name":"Step 1 (32K)","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-01-01","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":32768},"status":"deprecated","cost":{"input":2.05,"output":9.59,"cache_read":0.41}}}},"sap-ai-core":{"id":"sap-ai-core","env":["AICORE_SERVICE_KEY"],"npm":"@jerome-benoit/sap-ai-provider-v2","name":"SAP AI Core","doc":"https://help.sap.com/docs/sap-ai-core","models":{"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-3.8-flash"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"gemini-3.1-flash-lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"anthropic--claude-3-sonnet":{"id":"anthropic--claude-3-sonnet","name":"anthropic--claude-3-sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-08-31","release_date":"2024-03-04","last_updated":"2024-03-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":4096},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75}},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"gemini-2.5-flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"gpt-4.1-mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"canonical_model_id":"openai/gpt-6-astra"},"anthropic--claude-3-haiku":{"id":"anthropic--claude-3-haiku","name":"anthropic--claude-3-haiku","description":"Legacy model retained for compatibility with older integrations","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-08-31","release_date":"2024-03-13","last_updated":"2024-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":4096},"cost":{"input":0.25,"output":1.25,"cache_read":0.03,"cache_write":0.3},"canonical_model_id":"anthropic/claude-3-haiku-20240307"},"anthropic--claude-3-opus":{"id":"anthropic--claude-3-opus","name":"anthropic--claude-3-opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-08-31","release_date":"2024-02-29","last_updated":"2024-02-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":4096},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75}},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"gpt-5.6-terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"openai/gpt-5.6-terra"},"anthropic--claude-4.5-haiku":{"id":"anthropic--claude-4.5-haiku","name":"anthropic--claude-4.5-haiku","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"gpt-5-mini":{"id":"gpt-5-mini","name":"gpt-5-mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"amazon--nova-micro":{"id":"amazon--nova-micro","name":"amazon--nova-micro","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.035,"output":0.14},"canonical_model_id":"amazon/nova-micro"},"text-embedding-3-large":{"id":"text-embedding-3-large","type":"embedding","name":"text-embedding-3-large","description":"Most capable OpenAI embedding model for English and non-English text","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8191,"output":3072},"cost":{"input":0.13,"output":0},"canonical_model_id":"openai/text-embedding-3-large"},"nvidia--llama-3.2-nv-embedqa-1b":{"id":"nvidia--llama-3.2-nv-embedqa-1b","name":"nvidia--llama-3.2-nv-embedqa-1b","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":4096},"status":"deprecated"},"sap-abap-1":{"id":"sap-abap-1","name":"sap-abap-1","description":"SAP-hosted model for ABAP code generation and enterprise development tasks","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-11-26","last_updated":"2025-11-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":4096},"status":"deprecated"},"gpt-5-nano":{"id":"gpt-5-nano","name":"gpt-5-nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"gpt-5.2":{"id":"gpt-5.2","name":"gpt-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"anthropic--claude-4-opus":{"id":"anthropic--claude-4-opus","name":"anthropic--claude-4-opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-20250514"},"gpt-5.5":{"id":"gpt-5.5","name":"gpt-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"gemini-embedding":{"id":"gemini-embedding","type":"embedding","name":"Gemini Embedding 001","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-05","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2048,"output":1},"canonical_model_id":"google/gemini-embedding-001"},"mistralai--mistral-medium-instruct":{"id":"mistralai--mistral-medium-instruct","name":"mistralai--mistral-medium-instruct","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/mistral-medium-2505"},"gpt-5.4":{"id":"gpt-5.4","name":"gpt-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"canonical_model_id":"openai/gpt-6-luna"},"anthropic--claude-4.5-opus":{"id":"anthropic--claude-4.5-opus","name":"anthropic--claude-4.5-opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic--claude-4.6-sonnet":{"id":"anthropic--claude-4.6-sonnet","name":"anthropic--claude-4.6-sonnet","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"gpt-5.6-sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"tiers":[{"input":8,"output":30,"cache_read":0.8,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8}},"canonical_model_id":"openai/gpt-5.6-sol"},"text-embedding-3-small":{"id":"text-embedding-3-small","type":"embedding","name":"text-embedding-3-small","description":"Small, efficient OpenAI text embedding model","family":"text-embedding","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8191,"output":1536},"cost":{"input":0.02,"output":0},"canonical_model_id":"openai/text-embedding-3-small"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-3.5-flash-lite"},"anthropic--claude-3.7-sonnet":{"id":"anthropic--claude-3.7-sonnet","name":"anthropic--claude-3.7-sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2024-10-31","release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-3-7-sonnet-20250219"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"canonical_model_id":"openai/gpt-6.1-sol"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"gpt-5.6-luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04}},"canonical_model_id":"openai/gpt-5.6-luna"},"amazon--nova-pro":{"id":"amazon--nova-pro","name":"amazon--nova-pro","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":8192},"cost":{"input":0.8,"output":3.2},"canonical_model_id":"amazon/nova-pro"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"canonical_model_id":"google/gemini-3.6-flash"},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"gpt-4.1-nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":0.1,"output":0.4},"canonical_model_id":"openai/gpt-4.1-nano"},"nvidia--llama-nemotron-embed-1b":{"id":"nvidia--llama-nemotron-embed-1b","type":"embedding","name":"Llama Nemotron Embed 1B v2","description":"Multilingual text embedding model for retrieval and question answering pipelines","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-16","last_updated":"2025-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":2048},"canonical_model_id":"nvidia/llama-nemotron-embed-1b-v2"},"sonar-deep-research":{"id":"sonar-deep-research","name":"sonar-deep-research","description":"Sonar search model for autonomous research and citation-backed long-form reports","family":"sonar","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-02-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":2,"output":8,"reasoning":3},"canonical_model_id":"perplexity/sonar-deep-research"},"anthropic--claude-4.8-opus":{"id":"anthropic--claude-4.8-opus","name":"anthropic--claude-4.8-opus","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"mistralai--mistral-medium":{"id":"mistralai--mistral-medium","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"canonical_model_id":"mistral/mistral-medium-2604"},"anthropic--claude-3.5-sonnet":{"id":"anthropic--claude-3.5-sonnet","name":"anthropic--claude-3.5-sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04-30","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-3-5-sonnet-20241022"},"cohere--command-a-reasoning":{"id":"cohere--command-a-reasoning","name":"cohere--command-a-reasoning","description":"Cohere reasoning model for multilingual enterprise agents, tools, and complex workflows","family":"command-a","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]},{"type":"budget_tokens","min":1}],"tool_call":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"status":"deprecated","cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a-reasoning-08-2025"},"anthropic--claude-4.5-sonnet":{"id":"anthropic--claude-4.5-sonnet","name":"anthropic--claude-4.5-sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic--claude-4.7-opus":{"id":"anthropic--claude-4.7-opus","name":"anthropic--claude-4.7-opus","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"gemini-3.5-flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"amazon--titan-embed-text":{"id":"amazon--titan-embed-text","name":"amazon--titan-embed-text","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-04-30","last_updated":"2024-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1536},"cost":{"input":0.02,"output":0}},"gpt-4.1":{"id":"gpt-4.1","name":"gpt-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"status":"deprecated","cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"gpt-5":{"id":"gpt-5","name":"gpt-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"anthropic--claude-4-sonnet":{"id":"anthropic--claude-4-sonnet","name":"anthropic--claude-4-sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"anthropic--claude-4.6-opus":{"id":"anthropic--claude-4.6-opus","name":"anthropic--claude-4.6-opus","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"gemini-2.5-flash-lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"mistralai--mistral-small":{"id":"mistralai--mistral-small","name":"mistralai--mistral-small","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.15,"output":0.6},"canonical_model_id":"mistral/mistral-small-2603"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"gemini-2.5-pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"google/gemini-2.5-pro"},"sonar-pro":{"id":"sonar-pro","name":"sonar-pro","description":"Deeper Sonar search model with broader retrieval and stronger synthesis","family":"sonar-pro","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":3,"output":15},"canonical_model_id":"perplexity/sonar-pro"},"amazon--nova-lite":{"id":"amazon--nova-lite","name":"amazon--nova-lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.3,"output":2.5},"canonical_model_id":"amazon/nova-2-lite"},"sonar":{"id":"sonar","name":"sonar","description":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","family":"sonar","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":1,"output":1},"canonical_model_id":"perplexity/sonar"},"gemini-embedding-2":{"id":"gemini-embedding-2","type":"embedding","name":"Gemini Embedding 2","description":"Multimodal embedding model mapping text, images, video, audio, and PDFs into a unified embedding space","family":"gemini","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-11","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":3072},"canonical_model_id":"google/gemini-embedding-2"}}},"meganova":{"id":"meganova","env":["MEGANOVA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.meganova.ai/v1","name":"Meganova","doc":"https://docs.meganova.ai","models":{"moonshotai/Kimi-K2.5":{"id":"moonshotai/Kimi-K2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2026-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.45,"output":2.25,"cache_read":0.07},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/Kimi-K2-Thinking":{"id":"moonshotai/Kimi-K2-Thinking","name":"Kimi K2 Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":2.6},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"mistralai/Mistral-Nemo-Instruct-2407":{"id":"mistralai/Mistral-Nemo-Instruct-2407","name":"Mistral Nemo Instruct 2407","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.02,"output":0.04},"canonical_model_id":"mistral/mistral-nemo"},"mistralai/Mistral-Small-3.2-24B-Instruct-2506":{"id":"mistralai/Mistral-Small-3.2-24B-Instruct-2506","name":"Mistral Small 3.2 24B Instruct","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"mistral/mistral-small-2506"},"MiniMaxAI/MiniMax-M2.1":{"id":"MiniMaxAI/MiniMax-M2.1","name":"MiniMax M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.28,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.1"},"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.15},"canonical_model_id":"minimax/MiniMax-M2.5"},"XiaomiMiMo/MiMo-V2-Flash":{"id":"XiaomiMiMo/MiMo-V2-Flash","name":"MiMo V2 Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12-01","release_date":"2025-12-16","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32000},"status":"deprecated","cost":{"input":0.1,"output":0.3},"canonical_model_id":"xiaomi/mimo-v2-flash"},"meta-llama/Llama-3.3-70B-Instruct":{"id":"meta-llama/Llama-3.3-70B-Instruct","name":"Llama 3.3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"zai-org/GLM-5":{"id":"zai-org/GLM-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.8,"output":2.56,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"zai-org/GLM-4.7":{"id":"zai-org/GLM-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.4,"output":1.75,"cache_read":0.08},"canonical_model_id":"zhipuai/glm-4.7"},"zai-org/GLM-4.6":{"id":"zai-org/GLM-4.6","name":"GLM-4.6","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.45,"output":1.9},"canonical_model_id":"zhipuai/glm-4.6"},"deepseek-ai/DeepSeek-R1-0528":{"id":"deepseek-ai/DeepSeek-R1-0528","name":"DeepSeek R1 0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-07","release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":64000},"cost":{"input":0.5,"output":2.15,"cache_read":0.4},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek-ai/DeepSeek-V3.2":{"id":"deepseek-ai/DeepSeek-V3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.26,"output":0.38,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.19,"output":0.79},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek-ai/DeepSeek-V3.2-Exp":{"id":"deepseek-ai/DeepSeek-V3.2-Exp","name":"DeepSeek V3.2 Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-10","last_updated":"2025-10-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":164000},"cost":{"input":0.26,"output":0.38,"cache_read":0.13}},"deepseek-ai/DeepSeek-V3-0324":{"id":"deepseek-ai/DeepSeek-V3-0324","name":"DeepSeek V3 0324","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.275,"output":1.1},"canonical_model_id":"deepseek/deepseek-v3-0324"},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"cost":{"input":0.09,"output":0.1},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"Qwen/Qwen2.5-VL-32B-Instruct":{"id":"Qwen/Qwen2.5-VL-32B-Instruct","name":"Qwen2.5 VL 32B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":0.2,"output":0.6}},"Qwen/Qwen3.5-Plus":{"id":"Qwen/Qwen3.5-Plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.4,"output":2.4,"reasoning":2.4},"canonical_model_id":"alibaba/qwen3.5-plus"}}},"hpc-ai":{"id":"hpc-ai","env":["HPC_AI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.hpc-ai.com/inference/v1","name":"HPC-AI","doc":"https://www.hpc-ai.com/doc/docs/quickstart/","models":{"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1002000,"output":128000},"cost":{"input":1.74,"output":3.48,"cache_read":0.145},"canonical_model_id":"deepseek/deepseek-v4-pro"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-7"},"zai-org/glm-5.1":{"id":"zai-org/glm-5.1","name":"GLM 5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202000,"output":202000},"cost":{"input":0.615,"output":2.46,"cache_read":0.133},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/glm-5.2":{"id":"zai-org/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196000,"output":195000},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"}}},"databricks":{"id":"databricks","env":["DATABRICKS_HOST","DATABRICKS_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"https://${DATABRICKS_HOST}/ai-gateway/mlflow/v1","name":"Databricks","doc":"https://docs.databricks.com/aws/en/machine-learning/foundation-models/","models":{"databricks-claude-opus-4-1":{"id":"databricks-claude-opus-4-1","name":"Claude Opus 4.1 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"databricks-claude-haiku-4-5":{"id":"databricks-claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"databricks-gpt-5-4":{"id":"databricks-gpt-5-4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"databricks-gpt-5-6-terra":{"id":"databricks-gpt-5-6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":4,"output":24,"cache_read":0.4,"cache_write":5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"databricks-gemini-3-pro":{"id":"databricks-gemini-3-pro","name":"Gemini 3 Pro Preview","description":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3-pro-preview"},"databricks-claude-opus-4-7":{"id":"databricks-claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"databricks-gemini-3-1-pro":{"id":"databricks-gemini-3-1-pro","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"experimental":{"modes":{"fast":{"cost":{"input":3.6,"output":21.6,"cache_read":0.36,"input_audio":3.6},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2,"output":12,"cache_read":0.2,"input_audio":2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"databricks-gpt-5-mini":{"id":"databricks-gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":0.45,"output":3.6,"cache_read":0.045},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"databricks-kimi-k2-7-code":{"id":"databricks-kimi-k2-7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"databricks-glm-5-2":{"id":"databricks-glm-5-2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"experimental":{"modes":{"fast":{"cost":{"input":2.45,"output":7.7,"cache_read":0.455},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"databricks-gemini-3-flash":{"id":"databricks-gemini-3-flash","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"experimental":{"modes":{"fast":{"cost":{"input":0.9,"output":5.4,"cache_read":0.09,"input_audio":1.8},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"databricks-gpt-oss-20b":{"id":"databricks-gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":25000},"cost":{"input":0.07,"output":0.3},"canonical_model_id":"openai/gpt-oss-20b"},"databricks-gemini-3-5-flash":{"id":"databricks-gemini-3-5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"experimental":{"modes":{"fast":{"cost":{"input":2.7,"output":16.2,"cache_read":0.27,"input_audio":2.7},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"databricks-gemini-2-5-pro":{"id":"databricks-gemini-2-5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"databricks-claude-sonnet-4-5":{"id":"databricks-claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"databricks-gpt-5":{"id":"databricks-gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":2.5,"output":20,"cache_read":0.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"databricks-gpt-5-6-sol":{"id":"databricks-gpt-5-6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":8,"output":40,"cache_read":0.8,"cache_write":10},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"databricks-gpt-5-4-nano":{"id":"databricks-gpt-5-4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"databricks-gpt-5-5":{"id":"databricks-gpt-5-5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"databricks-gpt-5-2":{"id":"databricks-gpt-5-2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":3.5,"output":28,"cache_read":0.35},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"databricks-gpt-5-6-luna":{"id":"databricks-gpt-5-6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":0.4,"output":2.4,"cache_read":0.04,"cache_write":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"databricks-claude-sonnet-4":{"id":"databricks-claude-sonnet-4","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"databricks-claude-opus-4-5":{"id":"databricks-claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"databricks-gpt-5-nano":{"id":"databricks-gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"databricks-gpt-5-1":{"id":"databricks-gpt-5-1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":2.5,"output":20,"cache_read":0.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"databricks-claude-opus-4-6":{"id":"databricks-claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"databricks-claude-sonnet-4-6":{"id":"databricks-claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"databricks-gpt-5-4-mini":{"id":"databricks-gpt-5-4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":1.5,"output":9,"cache_read":0.15},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"databricks-gemini-3-1-flash-lite":{"id":"databricks-gemini-3-1-flash-lite","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"experimental":{"modes":{"fast":{"cost":{"input":0.45,"output":2.7,"cache_read":0.045,"input_audio":0.9},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"databricks-gpt-oss-120b":{"id":"databricks-gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":25000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"databricks-gemini-2-5-flash":{"id":"databricks-gemini-2-5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"}}},"moonshotai-cn":{"id":"moonshotai-cn","env":["MOONSHOT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.moonshot.cn/v1","name":"Moonshot AI (China)","doc":"https://platform.moonshot.cn/docs/api/chat","models":{"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.9711,"output":4.0337,"cache_read":0.1643},"canonical_model_id":"moonshotai/kimi-k2.6"},"kimi-k2.7-code-highspeed":{"id":"kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code HighSpeed","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9421,"output":8.0673,"cache_read":0.3884},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.9711,"output":4.0337,"cache_read":0.1942},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":2.9879,"output":14.9395,"cache_read":0.2988,"cache_write":2.9879},"canonical_model_id":"moonshotai/kimi-k3"}}},"drun":{"id":"drun","env":["DRUN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://chat.d.run/v1","name":"D.Run (China)","doc":"https://www.d.run","models":{"public/deepseek-v3":{"id":"public/deepseek-v3","name":"DeepSeek V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.28,"output":1.1},"canonical_model_id":"deepseek/deepseek-v3"},"public/deepseek-r1":{"id":"public/deepseek-r1","name":"DeepSeek R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32000},"cost":{"input":0.55,"output":2.2},"canonical_model_id":"deepseek/deepseek-r1"},"public/minimax-m25":{"id":"public/minimax-m25","name":"MiniMax M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.29,"output":1.16},"canonical_model_id":"minimax/MiniMax-M2.5"}}},"aiand":{"id":"aiand","env":["AIAND_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.aiand.com/v1","name":"ai&","doc":"https://docs.aiand.com/","models":{"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.75,"output":3.5,"cache_read":0.2},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":12.5,"cache_read":0.5},"canonical_model_id":"moonshotai/kimi-k3"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.08},"canonical_model_id":"openai/gpt-oss-120b"},"zai-org/glm-5.3-flash":{"id":"zai-org/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048550,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai-org/glm-5.2":{"id":"zai-org/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1,"output":4,"cache_read":0.3},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/glm-5.3":{"id":"zai-org/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1,"output":4,"cache_read":0.3},"canonical_model_id":"zhipuai/glm-5.3"},"deepseek-ai/deepseek-v4-flash":{"id":"deepseek-ai/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.15,"output":0.25,"cache_read":0.08},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-ai/deepseek-v4-pro":{"id":"deepseek-ai/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1,"output":2.5,"cache_read":0.25},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/deepseek-v4.1-flash":{"id":"deepseek-ai/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.3,"output":0.6,"cache_read":0.02},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.4,"output":3,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.32,"output":3.2,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3.6-27b"},"motif-technologies/motif-3":{"id":"motif-technologies/motif-3","name":"Motif 3","description":"Motif 3 is a large-scale, decoder-only Mixture-of-Experts (MoE) language model with 314 billion total parameters and 13.2 billion parameters activated per token.","family":"motif","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-08-07","last_updated":"2026-08-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":2,"cache_read":0.2},"canonical_model_id":"motif-technologies/motif-3"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"google/gemma-4-31b-it"}}},"nano-gpt":{"id":"nano-gpt","env":["NANO_GPT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://nano-gpt.com/api/v1","name":"NanoGPT","doc":"https://docs.nano-gpt.com","models":{"command-a-reasoning-08-2025":{"id":"command-a-reasoning-08-2025","name":"Cohere Command A (08/2025)","description":"Cohere reasoning model for multilingual enterprise agents, tools, and complex workflows","family":"command-a","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":8192},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"cohere/command-a-reasoning-08-2025"},"auto-model-premium":{"id":"auto-model-premium","name":"Auto model (Premium)","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2025-04-16","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":1000000},"cost":{"input":9.996,"output":19.992,"cache_read":4.998}},"glm-4-plus-0111":{"id":"glm-4-plus-0111","name":"GLM 4 Plus 0111","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":4096},"cost":{"input":9.996,"output":9.996,"cache_read":4.998}},"gemini-2.5-flash-preview-04-17:thinking":{"id":"gemini-2.5-flash-preview-04-17:thinking","name":"Gemini 2.5 Flash Preview Thinking","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"release_date":"2025-04-17","last_updated":"2025-04-17","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"status":"deprecated","cost":{"input":0.15,"output":3.5,"cache_read":0.015}},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-2.5-flash"},"gemma-4-12b-it-semancer":{"id":"gemma-4-12b-it-semancer","name":"Gemma 4 12B Semancer","description":"Gemma 4 12B Semancer is an open-weight roleplay finetune with image understanding, tool calling, optional reasoning, and a 131,072-token context window.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-09","last_updated":"2026-09-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.05,"output":0.25,"cache_read":0.025}},"gemini-2.0-pro-exp-02-05":{"id":"gemini-2.0-pro-exp-02-05","name":"Gemini 2.0 Pro 0205","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-02-05","last_updated":"2025-02-05","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":2097152,"input":2097152,"output":8192},"cost":{"input":1.989,"output":7.956,"cache_read":0.49725}},"gemini-2.5-flash-lite-preview-06-17":{"id":"gemini-2.5-flash-lite-preview-06-17","name":"Gemini 2.5 Flash Lite Preview","description":"First preview of Gemini 2.5 Flash-Lite, the low-cost, low-latency 2.5 model","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"google/gemini-2.5-flash-lite-preview-06-17"},"nano-gpt-help":{"id":"nano-gpt-help","name":"NanoGPT Help","description":"Text-only NanoGPT support assistant. Questions are processed by the Help inference provider; do not paste secrets or account credentials. Covers the website, models, API, pricing, memory, media generation, and support.","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-06-06","last_updated":"2026-06-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":6000,"input":6000,"output":512},"cost":{"input":0,"output":0}},"Gemma-4-31B-MeroMero-v2:thinking":{"id":"Gemma-4-31B-MeroMero-v2:thinking","name":"Gemma 4 31B MeroMero v2 Thinking","description":"Gemma 4 31B MeroMero v2 with thinking enabled for more deliberate emotive dialogue, relationship scenes, creative writing, and multimodal roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-08-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"doubao-seed-2-0-code-preview-260215":{"id":"doubao-seed-2-0-code-preview-260215","name":"Doubao Seed 2.0 Code Preview","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":128000},"cost":{"input":0.782,"output":3.893,"cache_read":0.391},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"doubao-1.5-pro-256k":{"id":"doubao-1.5-pro-256k","name":"Doubao 1.5 Pro 256k","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":16384},"cost":{"input":0.799,"output":1.445,"cache_read":0.3995}},"auto-model":{"id":"auto-model","name":"Auto model","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2025-04-16","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":1000000},"cost":{"input":0,"output":0}},"gemma-4-26b-a4b-it-moonlight":{"id":"gemma-4-26b-a4b-it-moonlight","name":"Moonlight Dusk","description":"Moonlight Dusk is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"auto-model-basic":{"id":"auto-model-basic","name":"Auto model (Basic)","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2025-04-16","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":1000000},"cost":{"input":9.996,"output":19.992,"cache_read":4.998}},"claw-medium":{"id":"claw-medium","name":"Claw Medium","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-11","last_updated":"2026-05-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2}},"glm-4.1v-thinking-flash":{"id":"glm-4.1v-thinking-flash","name":"GLM 4.1V Thinking Flash","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"glm-flash","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":64000,"input":64000,"output":8192},"cost":{"input":0.3,"output":0.3,"cache_read":0.15}},"holo3-35b-a3b:thinking":{"id":"holo3-35b-a3b:thinking","name":"Holo3-35B-A3B Thinking","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.25,"output":1.8,"cache_read":0.125}},"gemma-4-12b-it-station-keeper":{"id":"gemma-4-12b-it-station-keeper","name":"Gemma 4 12B StationKeeper","description":"Gemma 4 12B StationKeeper is an open-weight roleplay finetune with image understanding, tool calling, optional reasoning, and a 131,072-token context window.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-09","last_updated":"2026-09-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.05,"output":0.25,"cache_read":0.025}},"gemma-4-12b-it":{"id":"gemma-4-12b-it","name":"Gemma 4 12B Instruct","description":"Compact Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.05,"output":0.25,"cache_read":0.025},"canonical_model_id":"google/gemma-4-12b-it"},"ernie-5.0-thinking-preview":{"id":"ernie-5.0-thinking-preview","name":"Ernie 5.0 Thinking Preview","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"ernie","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":1,"output":3.5,"cache_read":0.5}},"mistral-code-latest":{"id":"mistral-code-latest","name":"Mistral Code Latest","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"mistral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.15}},"gemma-4-31b-it-fabled":{"id":"gemma-4-31b-it-fabled","name":"Fabled","description":"Fabled is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"glm-4-air-0111":{"id":"glm-4-air-0111","name":"GLM 4 Air 0111","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-01-11","last_updated":"2025-01-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":4096},"cost":{"input":0.1394,"output":0.1394,"cache_read":0.0697}},"doubao-1.5-vision-pro-32k":{"id":"doubao-1.5-vision-pro-32k","name":"Doubao 1.5 Vision Pro 32k","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-11-20","last_updated":"2025-01-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"input":32000,"output":8192},"cost":{"input":0.459,"output":1.377,"cache_read":0.2295}},"glm-z1-airx":{"id":"glm-z1-airx","name":"GLM Z1 AirX","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"input":32000,"output":16384},"cost":{"input":0.7,"output":0.7,"cache_read":0.35}},"gemma-4-31b-it-novelist":{"id":"gemma-4-31b-it-novelist","name":"Novelist","description":"Novelist is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"doubao-seed-2-0-lite-260215":{"id":"doubao-seed-2-0-lite-260215","name":"Doubao Seed 2.0 Lite","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":32000},"cost":{"input":0.1462,"output":0.8738,"cache_read":0.0731},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"mercury-coder-small":{"id":"mercury-coder-small","name":"Mercury Coder Small","description":"Model by Inception AI. A diffusion large language model that runs incredibly quickly (500+ tokens/second) while matching Claude 3.5 Haiku and GPT-4o-mini. 1st in speed on Copilot arena, and matching 2nd in quality.","family":"mercury","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.25,"output":1,"cache_read":0.125}},"agnes-3.0-flash":{"id":"agnes-3.0-flash","name":"Agnes 3.0 Flash","description":"Agnes 3.0 Flash is a low-cost model for coding, tool use, and multi-turn agent tasks. It supports text and image input, optional thinking, and a 512K-token context window.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-09","last_updated":"2026-09-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"input":524288,"output":65536},"cost":{"input":0.05,"output":0.15,"cache_read":0.005}},"gemini-3-pro-image-preview":{"id":"gemini-3-pro-image-preview","type":"image","name":"Gemini 3 Pro Image","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"input":65536,"output":32768},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"google/gemini-3-pro-image-preview"},"gemma-4-26b-a4b-it-luminous":{"id":"gemma-4-26b-a4b-it-luminous","name":"Luminous Mirror","description":"Luminous Mirror is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"gemini-2.5-flash-lite-preview-09-2025-thinking":{"id":"gemini-2.5-flash-lite-preview-09-2025-thinking","name":"Gemini 2.5 Flash Lite Preview (09/2025) – Thinking","description":"September 2025 preview of Gemini 2.5 Flash-Lite with better instruction following and multimodal quality","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"status":"deprecated","cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"google/gemini-2.5-flash-lite-preview-09-2025"},"pokee-isaac":{"id":"pokee-isaac","name":"Pokee-Isaac 28B","description":"Pokee-Isaac is a 28B agentic model with a roughly 10-million-token context window, function calling, and OpenAI-compatible structured output. Pokee bills in $0.01 increments, rounding each non-zero request up to the next cent.","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-08-04","last_updated":"2026-08-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":10000000,"input":10000000,"output":60000},"cost":{"input":0.15,"output":1,"cache_read":0.075}},"hermes-medium":{"id":"hermes-medium","name":"Hermes Medium","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"hermes","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-11","last_updated":"2026-05-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2}},"doubao-1.5-pro-32k":{"id":"doubao-1.5-pro-32k","name":"Doubao 1.5 Pro 32k","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-11-20","last_updated":"2025-01-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"input":32000,"output":8192},"cost":{"input":0.1343,"output":0.3349,"cache_read":0.06715}},"mercury-2":{"id":"mercury-2","name":"Mercury 2","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":50000},"cost":{"input":0.25,"output":0.75,"cache_read":0.025},"canonical_model_id":"inception/mercury-2"},"gemma-4-26b-a4b-it-shadowsiren":{"id":"gemma-4-26b-a4b-it-shadowsiren","name":"Shadow Siren","description":"Shadow Siren is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"claw-high":{"id":"claw-high","name":"Claw High","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-11","last_updated":"2026-05-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2}},"Gemma-4-26B-A4B-MeroMero":{"id":"Gemma-4-26B-A4B-MeroMero","name":"Gemma 4 26B A4B MeroMero","description":"Gemma 4 26B A4B MeroMero is an NVFP4 multimodal mixture-of-experts fine-tune for emotive dialogue, relationship scenes, creative writing, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"asi1-mini":{"id":"asi1-mini","name":"ASI1 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-03-25","last_updated":"2025-03-25","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":1,"output":1,"cache_read":0.5}},"longcat-2.0:thinking":{"id":"longcat-2.0:thinking","name":"LongCat 2.0 Thinking","description":"Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":131072},"cost":{"input":0.75,"output":3,"cache_read":0.015},"canonical_model_id":"meituan/longcat-2.0"},"deepseek-chat":{"id":"deepseek-chat","name":"DeepSeek V3/Deepseek Chat","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.1,"output":0.425,"cache_read":0.05},"canonical_model_id":"deepseek/deepseek-chat"},"gemini-2.5-flash-preview-09-2025":{"id":"gemini-2.5-flash-preview-09-2025","name":"Gemini 2.5 Flash Preview (09/2025)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03}},"gemma-4-26b-a4b-it-opusdistill":{"id":"gemma-4-26b-a4b-it-opusdistill","name":"Opus Distill","description":"Opus Distill is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"gemini-2.5-flash-preview-09-2025-thinking":{"id":"gemini-2.5-flash-preview-09-2025-thinking","name":"Gemini 2.5 Flash Preview (09/2025) – Thinking","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03}},"ernie-5.1:thinking":{"id":"ernie-5.1:thinking","name":"ERNIE 5.1 Thinking","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"ernie","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2026-05-10","last_updated":"2026-05-10","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":119000,"input":119000,"output":64000},"cost":{"input":0.75,"output":3,"cache_read":0.75}},"gemma-4-26b-a4b-it-musica":{"id":"gemma-4-26b-a4b-it-musica","name":"Musica","description":"Musica is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"Gemma-4-31B-Cognitive-Unshackled":{"id":"Gemma-4-31B-Cognitive-Unshackled","name":"Gemma 4 31B Cognitive Unshackled","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.306,"output":0.306,"cache_read":0.153}},"Gemma-4-31B-GarnetV2":{"id":"Gemma-4-31B-GarnetV2","name":"Gemma 4 31B Garnet V2","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.306,"output":0.306,"cache_read":0.153}},"doubao-seed-2-0-pro-260215":{"id":"doubao-seed-2-0-pro-260215","name":"Doubao Seed 2.0 Pro","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":128000},"cost":{"input":0.782,"output":3.876,"cache_read":0.391},"canonical_model_id":"bytedance-seed/seed-2.0-pro"},"gemma-4-31b-it-scarlet":{"id":"gemma-4-31b-it-scarlet","name":"Gemma 4 31B Scarlet","description":"Scarlet is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-08","last_updated":"2026-10-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"doubao-seed-2-0-mini-260215":{"id":"doubao-seed-2-0-mini-260215","name":"Doubao Seed 2.0 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":32000},"cost":{"input":0.0493,"output":0.4845,"cache_read":0.02465},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"gemini-2.5-flash-nothinking":{"id":"gemini-2.5-flash-nothinking","name":"Gemini 2.5 Flash (No Thinking)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03}},"longcat-2.0":{"id":"longcat-2.0","name":"LongCat 2.0","description":"Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window","family":"longcat","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":131072},"cost":{"input":0.75,"output":3,"cache_read":0.015},"canonical_model_id":"meituan/longcat-2.0"},"Meta-Llama-3-1-8B-Instruct-FP8":{"id":"Meta-Llama-3-1-8B-Instruct-FP8","name":"Llama 3.1 8B (decentralized)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.02,"output":0.03,"cache_read":0.01}},"Gemma-4-31B-Queen":{"id":"Gemma-4-31B-Queen","name":"Gemma 4 31B Queen","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.306,"output":0.306,"cache_read":0.153}},"Qwen3.5-27B-BlueStar-v3-Derestricted":{"id":"Qwen3.5-27B-BlueStar-v3-Derestricted","name":"Qwen3.5 27B BlueStar v3 Derestricted","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.306,"output":0.306,"cache_read":0.153}},"sarvam-105b":{"id":"sarvam-105b","name":"Sarvam 105B","description":"Flagship Indian-language reasoning model for enterprise multilingual applications","family":"sarvam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-18","last_updated":"2026-02-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":4096},"cost":{"input":0.054,"output":0.2124,"cache_read":0.0336},"canonical_model_id":"sarvam/sarvam-105b"},"gemini-2.5-flash-lite-preview-09-2025":{"id":"gemini-2.5-flash-lite-preview-09-2025","name":"Gemini 2.5 Flash Lite Preview (09/2025)","description":"September 2025 preview of Gemini 2.5 Flash-Lite with better instruction following and multimodal quality","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"status":"deprecated","cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"google/gemini-2.5-flash-lite-preview-09-2025"},"hermes-low":{"id":"hermes-low","name":"Hermes Low","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"hermes","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-11","last_updated":"2026-05-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2,"cache_write":0.08333}},"Gemma-4-31B-DarkIdol":{"id":"Gemma-4-31B-DarkIdol","name":"Gemma 4 31B DarkIdol","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.306,"output":0.306,"cache_read":0.153}},"Gemma-4-31B-MeroMero-v2":{"id":"Gemma-4-31B-MeroMero-v2","name":"Gemma 4 31B MeroMero v2","description":"Gemma 4 31B MeroMero v2 is a LoRA finetune for emotive dialogue, relationship scenes, creative writing, and multimodal roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-08-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"hermes-high":{"id":"hermes-high","name":"Hermes High","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"hermes","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-11","last_updated":"2026-05-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2}},"gemma-4-31b-it-garnet":{"id":"gemma-4-31b-it-garnet","name":"Garnet","description":"Garnet is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"gemma-4-31b-it-isometry":{"id":"gemma-4-31b-it-isometry","name":"Isometry","description":"Isometry is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"gemini-2.5-flash-preview-05-20":{"id":"gemini-2.5-flash-preview-05-20","name":"Gemini 2.5 Flash 0520","description":"Preview snapshot of Gemini 2.5 Flash with controllable thinking budgets","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048000,"input":1048000,"output":65536},"status":"deprecated","cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"google/gemini-2.5-flash-preview-05-20"},"GLM-4.6-Derestricted-v5":{"id":"GLM-4.6-Derestricted-v5","name":"GLM 4.6 Derestricted v5","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":8192},"cost":{"input":0.4,"output":1.5,"cache_read":0.2}},"claw-low":{"id":"claw-low","name":"Claw Low","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-11","last_updated":"2026-05-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2,"cache_write":0.08333}},"gemma-4-26b-a4b-it-chimeraxheretic":{"id":"gemma-4-26b-a4b-it-chimeraxheretic","name":"Gemma 4 26B A4B Chimera X Heretic","description":"Chimera X Heretic is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-08","last_updated":"2026-10-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"deepseek-chat-cheaper":{"id":"deepseek-chat-cheaper","name":"DeepSeek V3/Chat Cheaper","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"deepseek","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.1,"output":0.425,"cache_read":0.05}},"qvq-max":{"id":"qvq-max","name":"Qwen: QvQ Max","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-03-28","last_updated":"2025-03-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":1.2,"output":4.8,"cache_read":0.6}},"gemini-2.5-flash-preview-04-17":{"id":"gemini-2.5-flash-preview-04-17","name":"Gemini 2.5 Flash Preview","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"release_date":"2025-04-17","last_updated":"2025-04-17","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"status":"deprecated","cost":{"input":0.15,"output":0.6,"cache_read":0.015}},"universal-summarizer":{"id":"universal-summarizer","name":"Universal Summarizer","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-12-23","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":30,"output":30}},"kimi-k2-instruct-fast":{"id":"kimi-k2-instruct-fast","name":"Kimi K2 0711 Fast","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"kimi-k2","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":16384},"status":"deprecated","cost":{"input":0.4,"output":1.8,"cache_read":0.2},"canonical_model_id":"moonshotai/kimi-k2-instruct"},"fastgpt":{"id":"fastgpt","name":"Web Answer","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-12-23","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":7.5,"output":7.5}},"venice-uncensored":{"id":"venice-uncensored","name":"Venice Uncensored","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"venice","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-10-01","last_updated":"2025-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.4,"output":1.8,"cache_read":0.4}},"gemini-2.5-pro-exp-03-25":{"id":"gemini-2.5-pro-exp-03-25","name":"Gemini 2.5 Pro Experimental 0325","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"release_date":"2025-03-25","last_updated":"2025-03-25","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2.5,"output":10,"cache_read":0.25}},"gemma-4-31b-it-gembrain":{"id":"gemma-4-31b-it-gembrain","name":"Gembrain","description":"Gembrain is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"gemini-exp-1206":{"id":"gemini-exp-1206","name":"Gemini 2.0 Pro 1206","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":2097152,"input":2097152,"output":8192},"cost":{"input":1.258,"output":4.998,"cache_read":0.629}},"holo3-35b-a3b":{"id":"holo3-35b-a3b","name":"Holo3-35B-A3B","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.25,"output":1.8,"cache_read":0.125}},"glm-4.1v-thinking-flashx":{"id":"glm-4.1v-thinking-flashx","name":"GLM 4.1V Thinking FlashX","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"glm-flash","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":64000,"input":64000,"output":8192},"cost":{"input":0.3,"output":0.3,"cache_read":0.15}},"phi-4-multimodal-instruct":{"id":"phi-4-multimodal-instruct","name":"Phi 4 Multimodal","description":"Compact Microsoft multimodal model for text, image, and speech understanding","family":"phi","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-02-26","last_updated":"2025-02-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.07,"output":0.11,"cache_read":0.035},"canonical_model_id":"microsoft/phi-4-multimodal-instruct"},"mistral-code-agent-latest":{"id":"mistral-code-agent-latest","name":"Mistral Code Agent Latest","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"mistral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.4,"output":2,"cache_read":0.2}},"gemma-4-26b-a4b-it-chimerax":{"id":"gemma-4-26b-a4b-it-chimerax","name":"Chimera X","description":"Chimera X is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"gemini-2.0-pro-reasoner":{"id":"gemini-2.0-pro-reasoner","name":"Gemini 2.0 Pro Reasoner","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-02-05","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":1.292,"output":4.998,"cache_read":0.323}},"Gemma-4-26B-A4B-MeroMero:thinking":{"id":"Gemma-4-26B-A4B-MeroMero:thinking","name":"Gemma 4 26B A4B MeroMero Thinking","description":"Gemma 4 26B A4B MeroMero with thinking enabled for more deliberate emotive dialogue, relationship scenes, creative writing, and multimodal roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"phi-4-mini-instruct":{"id":"phi-4-mini-instruct","name":"Phi 4 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"phi","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-02-26","last_updated":"2025-02-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.17,"output":0.68,"cache_read":0.085},"canonical_model_id":"microsoft/phi-4-mini"},"ernie-5.1":{"id":"ernie-5.1","name":"ERNIE 5.1","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"ernie","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-05-10","last_updated":"2026-05-10","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":119000,"input":119000,"output":64000},"cost":{"input":0.75,"output":3,"cache_read":0.75}},"gemma-4-31b-it-darkidol":{"id":"gemma-4-31b-it-darkidol","name":"DarkIdol","description":"DarkIdol is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"gemini-2.5-pro-preview-06-05":{"id":"gemini-2.5-pro-preview-06-05","name":"Gemini 2.5 Pro Preview 0605","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01-31","release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2.5,"output":10,"cache_read":0.25},"canonical_model_id":"google/gemini-2.5-pro-preview"},"gemma-4-26b-a4b-it-darksoul":{"id":"gemma-4-26b-a4b-it-darksoul","name":"Dark Soul","description":"Dark Soul is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.12,"output":0.38,"cache_read":0.06}},"gemma-4-31b-it-gemsicle":{"id":"gemma-4-31b-it-gemsicle","name":"Gemsicle","description":"Gemsicle is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-07-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"Gemma-4-31B-Claude-4.6-Opus-Reasoning-Distilled":{"id":"Gemma-4-31B-Claude-4.6-Opus-Reasoning-Distilled","name":"Gemma 4 31B Claude 4.6 Opus Reasoning Distilled","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"claude","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-05-01","last_updated":"2026-05-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.306,"output":0.306,"cache_read":0.0306}},"Qwen3.5-27B-Queen-Derestricted":{"id":"Qwen3.5-27B-Queen-Derestricted","name":"Qwen3.5 27B Queen Derestricted","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.306,"output":0.306,"cache_read":0.153}},"gemini-2.5-flash-preview-05-20:thinking":{"id":"gemini-2.5-flash-preview-05-20:thinking","name":"Gemini 2.5 Flash 0520 Thinking","description":"Preview snapshot of Gemini 2.5 Flash with controllable thinking budgets","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048000,"input":1048000,"output":65536},"status":"deprecated","cost":{"input":0.15,"output":3.5,"cache_read":0.015},"canonical_model_id":"google/gemini-2.5-flash-preview-05-20"},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"Gemini 2.5 Flash Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"google/gemini-2.5-flash-lite"},"qwen3-vl-235b-a22b-instruct-original":{"id":"qwen3-vl-235b-a22b-instruct-original","name":"Qwen3 VL 235B A22B Instruct Original","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":0.3,"output":1.9,"cache_read":0.15},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"auto-model-standard":{"id":"auto-model-standard","name":"Auto model (Standard)","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2025-04-16","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":1000000},"cost":{"input":9.996,"output":19.992,"cache_read":4.998}},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"cache_write":0.375},"canonical_model_id":"google/gemini-2.5-pro"},"glm-4-long":{"id":"glm-4-long","name":"GLM-4 Long","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-08-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":4095},"cost":{"input":0.2006,"output":0.2006,"cache_read":0.1003}},"gemini-2.5-pro-preview-03-25":{"id":"gemini-2.5-pro-preview-03-25","name":"Gemini 2.5 Pro Preview 0325","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"release_date":"2025-03-25","last_updated":"2025-03-25","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2.5,"output":10,"cache_read":0.25}},"celeris-1":{"id":"celeris-1","name":"Celeris 1","description":"Celeris 1 is a diffusion language model built for ultra-low-latency classification, extraction, judging, query rewriting, and other short structured responses.","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-07-25","last_updated":"2026-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"input":8192,"output":8192},"cost":{"input":2,"output":6,"cache_read":1}},"MiniMax-M1":{"id":"MiniMax-M1","name":"MiniMax M1","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"minimax","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-01-08","last_updated":"2025-06-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":131072},"cost":{"input":0.1394,"output":1.3328,"cache_read":0.0697}},"longcat-2.5-preview":{"id":"longcat-2.5-preview","name":"LongCat 2.5 Preview","description":"Meituan's multimodal reasoning model for coding and agent tasks, with image understanding and a 1M-token context window","family":"longcat","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":131072},"cost":{"input":0.75,"output":2.95,"cache_read":0.015},"canonical_model_id":"meituan/longcat-2.5-preview"},"gemini-2.5-pro-preview-05-06":{"id":"gemini-2.5-pro-preview-05-06","name":"Gemini 2.5 Pro Preview 0506","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"release_date":"2025-05-06","last_updated":"2025-05-06","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2.5,"output":10,"cache_read":0.25}},"deepseek-v3-0324":{"id":"deepseek-v3-0324","name":"DeepSeek Chat 0324","description":"March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.2,"output":0.77,"cache_read":0.135},"canonical_model_id":"deepseek/deepseek-v3-0324"},"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.3-contributor":{"id":"meta/muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Meta's Muse Spark 1.3 Contributor is a frontier multimodal reasoning model for long-horizon coding and agentic workflows, with strong gains in computer use, browsing, professional tool use, codebase understanding, and million-token retrieval. It accepts text, images, audio, video, and files, supports tool calling and structured output, and always reasons before answering. Prompts and outputs may be used by Meta for training and to improve its products.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":943718},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/muse-glimmer-30b":{"id":"meta/muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":117964},"cost":{"input":0.35,"output":1.5,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"meta/muse-spark-1.2-contributor":{"id":"meta/muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor (Data Used for Training)","description":"A much cheaper opt-in version of Muse Spark 1.2 with the same multimodal coding and agentic capabilities. Prompts and outputs sent to this Contributor model may be used by Meta for training and to improve its products; use the standard Muse Spark 1.2 model if you do not want your data used for training.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/muse-spark-open":{"id":"meta/muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"sakana/fugu-ultra-v1.1":{"id":"sakana/fugu-ultra-v1.1","name":"Fugu Ultra v1.1","description":"Sakana AI's upgraded Fugu Ultra release with stronger coding, agentic task execution, and advanced reasoning through dynamic orchestration of frontier models.","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":16384},"cost":{"input":5,"output":30,"cache_read":0.5}},"sakana/fugu-max":{"id":"sakana/fugu-max","name":"Fugu Max","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"sakana/fugu-max"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":16384},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"sakana/fugu-ultra"},"z-ai/glm-4.7:thinking":{"id":"z-ai/glm-4.7:thinking","name":"GLM 4.7 Thinking","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65535},"cost":{"input":0.4004,"output":1.9292,"cache_read":0.08008},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-4.6:thinking":{"id":"z-ai/glm-4.6:thinking","name":"GLM 4.6 Thinking","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65535},"cost":{"input":0.5,"output":2,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-4.6"},"z-ai/glm-5:thinking":{"id":"z-ai/glm-5:thinking","name":"GLM 5 Thinking","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":128000},"cost":{"input":0.6,"output":1.9,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-5"},"z-ai/GLM-4.5:thinking":{"id":"z-ai/GLM-4.5:thinking","name":"GLM 4.5 (Thinking)","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.3,"output":1.3,"cache_read":0.15},"canonical_model_id":"zhipuai/glm-4.5"},"z-ai/glm-5.1:thinking":{"id":"z-ai/glm-5.1:thinking","name":"GLM 5.1 Thinking","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":131072},"cost":{"input":0.7,"output":2.5,"cache_read":0.18},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-5.3:thinking":{"id":"z-ai/glm-5.3:thinking","name":"GLM 5.3 Thinking","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-5.3-flash-apex":{"id":"z-ai/glm-5.3-flash-apex","name":"Apex Flash 1","description":"Apex Flash 1 is Cantina Security and Yeta's cybersecurity-focused post-train of GLM 5.3 Flash for code investigation and tool use.","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":32768},"cost":{"input":0.15,"output":0.5,"cache_read":0.075}},"z-ai/glm-5-original":{"id":"z-ai/glm-5-original","name":"GLM 5 Original","description":"GLM-5 is Zhipu's latest flagship model with advanced reasoning and instruction following. Routed directly via Z-AI (Zhipu).","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":128000},"cost":{"input":0.6,"output":1.9,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-5v-turbo:thinking":{"id":"z-ai/glm-5v-turbo:thinking","name":"GLM 5V Turbo Thinking","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":202800,"input":202800,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"z-ai/glm-4.5v:thinking":{"id":"z-ai/glm-4.5v:thinking","name":"GLM 4.5V Thinking","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":16384},"cost":{"input":0.6,"output":1.8,"cache_read":0.3},"canonical_model_id":"zhipuai/glm-4.5v"},"z-ai/glm-4.5v":{"id":"z-ai/glm-4.5v","name":"GLM 4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":16384},"cost":{"input":0.6,"output":1.8,"cache_read":0.3},"canonical_model_id":"zhipuai/glm-4.5v"},"z-ai/glm-5.3-flash-cybersecurity":{"id":"z-ai/glm-5.3-flash-cybersecurity","name":"GLM 5.3 Flash Cybersecurity","description":"GLM 5.3 Flash Cybersecurity is a cybersecurity-focused variant based on the uncensored model, with provider moderation for illegal activities. It supports always-on reasoning, image understanding, tool calling, and a 1,048,576-token context window.","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-09-19","last_updated":"2026-09-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":32768},"cost":{"input":0.15,"output":0.5,"cache_read":0.075}},"z-ai/glm-5.3-uncensored":{"id":"z-ai/glm-5.3-uncensored","name":"GLM 5.3 Uncensored","description":"An experimental uncensored version of the full GLM 5.3 reasoning model for chat, creative writing, coding, and tool use. It may show unusual behavior or inconsistent response quality.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-09-26","last_updated":"2026-09-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":1.25,"output":2.25,"cache_read":0.3}},"z-ai/glm-latest":{"id":"z-ai/glm-latest","name":"GLM Latest","description":"Compatibility alias that routes to the newest thinking GLM model. Currently routes to GLM 5.2 Thinking.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-05-03","last_updated":"2026-05-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2}},"z-ai/glm-4.6v-original":{"id":"z-ai/glm-4.6v-original","name":"GLM 4.6V Original","description":"GLM-4.6V scales its context window to 128k tokens in training, and achieves SoTA performance in visual understanding among models of similar parameter scales. Integrates native Function Calling capabilities, bridging 'visual perception' and 'executable action' for multimodal agents. Direct via Z-AI (Zhipu).","family":"glm","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":24000},"cost":{"input":0.6,"output":0.9,"cache_read":0.3},"canonical_model_id":"zhipuai/glm-4.6v"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM 5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":131072},"cost":{"input":0.7,"output":2.5,"cache_read":0.18},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"GLM 4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65535},"cost":{"input":0.4004,"output":1.9292,"cache_read":0.08008},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5.3-flash-apex-abliterated":{"id":"z-ai/glm-5.3-flash-apex-abliterated","name":"Apex Flash 1 Abliterated","description":"Apex Flash 1 Abliterated is an experimental derivative of Cantina Security and Yeta's Apex Flash 1 with modified refusal behavior.","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":32768},"cost":{"input":0.15,"output":0.5,"cache_read":0.075}},"z-ai/glm-5.3-flash-uncensored":{"id":"z-ai/glm-5.3-flash-uncensored","name":"GLM 5.3 Flash Uncensored","description":"GLM 5.3 Flash Uncensored is an uncensored fine-tune of the efficient 320B mixture-of-experts reasoning model, built for unrestricted chat, creative writing, coding, agentic work, tool use, and long-context tasks.","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-07-29","last_updated":"2026-08-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":32768},"cost":{"input":0.2,"output":0.8,"cache_read":0.07}},"z-ai/glm-4.7-original:thinking":{"id":"z-ai/glm-4.7-original:thinking","name":"GLM 4.7 Original Thinking","description":"GLM-4.7 original with extended thinking capabilities for complex reasoning.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65535},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM 5 Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202800,"input":202800,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"z-ai/glm-4.5":{"id":"z-ai/glm-4.5","name":"GLM 4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.3,"output":1.3,"cache_read":0.15},"canonical_model_id":"zhipuai/glm-4.5"},"z-ai/GLM-4.5-Air:thinking":{"id":"z-ai/GLM-4.5-Air:thinking","name":"GLM 4.5 Air (Thinking)","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":98304},"cost":{"input":0.12,"output":0.8,"cache_read":0.06},"canonical_model_id":"zhipuai/glm-4.5-air"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.1,"output":0.3,"cache_read":0.025},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/GLM-4.6-turbo":{"id":"z-ai/GLM-4.6-turbo","name":"GLM 4.6 Turbo","description":"Fast variant of GLM 4.6 for general chat, coding, and analysis with improved latency and strong reasoning.","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-10-02","last_updated":"2025-10-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"input":204800,"output":131072},"cost":{"input":1,"output":3,"cache_read":0.5}},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.42,"output":1.32,"cache_read":0.078},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-4.7-flash:thinking":{"id":"z-ai/glm-4.7-flash:thinking","name":"GLM 4.7 Flash Thinking","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":128000},"cost":{"input":0.07,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flash"},"z-ai/glm-4.7-flash":{"id":"z-ai/glm-4.7-flash","name":"GLM 4.7 Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":128000},"cost":{"input":0.07,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flash"},"z-ai/glm-4.7-original":{"id":"z-ai/glm-4.7-original","name":"GLM 4.7 Original","description":"GLM-4.7 is a next-gen GLM series text model with stronger reasoning, long-context chat, and reliable tool use. Routed directly via Z-AI (Zhipu).","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65535},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5-original:thinking":{"id":"z-ai/glm-5-original:thinking","name":"GLM 5 Original Thinking","description":"GLM-5 original with extended thinking capabilities for complex reasoning.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":128000},"cost":{"input":0.6,"output":1.9,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM 5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":128000},"cost":{"input":0.6,"output":1.9,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-5.2:thinking":{"id":"z-ai/glm-5.2:thinking","name":"GLM 5.2 Thinking","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.42,"output":1.32,"cache_read":0.078},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/GLM-4.6-turbo:thinking":{"id":"z-ai/GLM-4.6-turbo:thinking","name":"GLM 4.6 Turbo (Thinking)","description":"GLM 4.6 Turbo with thinking mode enabled for enhanced reasoning; shows internal reasoning and supports long context.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2025-10-02","last_updated":"2025-10-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"input":204800,"output":131072},"cost":{"input":1,"output":3,"cache_read":0.5}},"z-ai/glm-4.6":{"id":"z-ai/glm-4.6","name":"GLM 4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":65535},"cost":{"input":0.5,"output":2,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-4.6"},"z-ai/GLM-4.5-Air":{"id":"z-ai/GLM-4.5-Air","name":"GLM 4.5 Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":98304},"cost":{"input":0.12,"output":0.8,"cache_read":0.06},"canonical_model_id":"zhipuai/glm-4.5-air"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.6,"output":1.1,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-5v-turbo":{"id":"z-ai/glm-5v-turbo","name":"GLM 5V Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":202800,"input":202800,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"z-ai/glm-4.6v":{"id":"z-ai/glm-4.6v","name":"GLM 4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":24000},"cost":{"input":0.3,"output":0.9,"cache_read":0.15},"canonical_model_id":"zhipuai/glm-4.6v"},"z-ai/glm-4.6-original":{"id":"z-ai/glm-4.6-original","name":"GLM 4.6 Original","description":"GLM-4.6, Zhipu's flagship text model with 256K context window and advanced reasoning capabilities. Direct via Z-AI (Zhipu).","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":65535},"cost":{"input":0.5,"output":2,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-4.6"},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":98304},"cost":{"input":0.6,"output":2.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"moonshotai/kimi-k2.5:thinking":{"id":"moonshotai/kimi-k2.5:thinking","name":"Kimi K2.5 Thinking","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.375,"output":2,"cache_read":0.07},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.44935,"output":1.892,"cache_read":0.07568},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k2.7-code-highspeed":{"id":"moonshotai/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code High-Speed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":1.9,"output":8,"cache_read":0.32},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshotai/Kimi-K2-Instruct-0905":{"id":"moonshotai/Kimi-K2-Instruct-0905","name":"Kimi K2 0905","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":100352},"cost":{"input":0.4,"output":1.8,"cache_read":0.2},"canonical_model_id":"moonshotai/kimi-k2-0905"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.6712,"output":3.35,"cache_read":0.18},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":943718},"cost":{"input":1.2,"output":5.5,"cache_read":0.2},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2-instruct":{"id":"moonshotai/kimi-k2-instruct","name":"Kimi K2 Instruct","description":"Non-thinking Kimi K2 MoE instruction model for coding and tool-using agents","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":8192},"cost":{"input":0.4,"output":1.8,"cache_read":0.2},"canonical_model_id":"moonshotai/kimi-k2-instruct"},"moonshotai/kimi-latest":{"id":"moonshotai/kimi-latest","name":"Kimi Latest","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-03","last_updated":"2026-05-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":943718},"cost":{"input":1.2,"output":5.5,"cache_read":0.2}},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.375,"output":2,"cache_read":0.07},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/kimi-k2-instruct-0711":{"id":"moonshotai/kimi-k2-instruct-0711","name":"Kimi K2 0711","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.4,"output":1.8,"cache_read":0.2},"canonical_model_id":"moonshotai/kimi-k2-instruct"},"moonshotai/kimi-k2.6:thinking":{"id":"moonshotai/kimi-k2.6:thinking","name":"Kimi K2.6 Thinking","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.44935,"output":1.892,"cache_read":0.07568},"canonical_model_id":"moonshotai/kimi-k2.6"},"nimbz/schattenblume-31b":{"id":"nimbz/schattenblume-31b","name":"Schattenblume 31B","description":"Nimbz's Gemma 4 31B merge for character roleplay, expressive prose, and creative dialogue. It combines Scotoma 2, Giftige-Blume v1, and MeroMero v2. This text-only deployment supports a 65,536-token context window and up to 8,192 output tokens.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"nimbz/gemma-4-giftige-blume-31b-v2":{"id":"nimbz/gemma-4-giftige-blume-31b-v2","name":"Gemma 4 Giftige-Blume V2 31B","description":"Nimbz's second Giftige-Blume merge focuses on character roleplay, creative prose, and stricter prompt adherence. It combines storymaxxed, Equinox, Mero-Artemis, Musica, Ortenzya, Gemopus, and Sprinkle. This text-only Gemma 4 31B deployment supports a 65,536-token context window and up to 8,192 output tokens.","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"mistralai/mistral-medium-3.5:thinking":{"id":"mistralai/mistral-medium-3.5:thinking","name":"Mistral Medium 3.5 Thinking","description":"Mistral Medium 3.5 with reasoning enabled by default (reasoning_effort=high), for complex coding, agentic, and multi-step reasoning prompts.","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":1.5,"output":7.5,"cache_read":0.75}},"mistralai/mistral-saba":{"id":"mistralai/mistral-saba","name":"Mistral Saba","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-02-17","last_updated":"2025-02-17","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":26214},"cost":{"input":0.1989,"output":0.595,"cache_read":0.09945},"canonical_model_id":"mistral/mistral-saba"},"mistralai/ministral-3b-2512":{"id":"mistralai/ministral-3b-2512","name":"Ministral 3B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.1,"output":0.1,"cache_read":0.05},"canonical_model_id":"mistral/ministral-3b-2512"},"mistralai/mistral-large-4":{"id":"mistralai/mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"input":524288,"output":262144},"cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"mistralai/ministral-14b-instruct-2512":{"id":"mistralai/ministral-14b-instruct-2512","name":"Ministral 3 14B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.05},"canonical_model_id":"mistral/ministral-3-14b-instruct-2512"},"mistralai/mistral-medium-3":{"id":"mistralai/mistral-medium-3","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.4,"output":2,"cache_read":0.2},"canonical_model_id":"mistral/mistral-medium-3"},"mistralai/mistral-small-3.1-24b-instruct":{"id":"mistralai/mistral-small-3.1-24b-instruct","name":"Mistral Small 3.1 24B (2503)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":102400},"cost":{"input":0.1,"output":0.3,"cache_read":0.05},"canonical_model_id":"mistral/mistral-small-3.1-24b-instruct"},"mistralai/ministral-8b-2512":{"id":"mistralai/ministral-8b-2512","name":"Ministral 8B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.15,"output":0.15,"cache_read":0.075},"canonical_model_id":"mistral/ministral-8b-2512"},"mistralai/devstral-small-2505":{"id":"mistralai/devstral-small-2505","name":"Mistral Devstral Small 2505","description":"OpenHands+Devstral is 100% local 100% open, and is SOTA for the category on SWE-Bench Verified: 46.8% accuracy.","family":"devstral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-08-02","last_updated":"2025-08-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.06,"output":0.06,"cache_read":0.03}},"mistralai/codestral-2508":{"id":"mistralai/codestral-2508","name":"Codestral 2508","description":"Mistral code model for low-latency completion, fill-in-the-middle, and code generation","family":"codestral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.15},"canonical_model_id":"mistral/codestral-2508"},"mistralai/devstral-2-123b-instruct-2512":{"id":"mistralai/devstral-2-123b-instruct-2512","name":"Devstral 2 123B","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.4,"output":1.4,"cache_read":0.2},"canonical_model_id":"mistral/devstral-2512"},"mistralai/mistral-medium-3.1":{"id":"mistralai/mistral-medium-3.1","name":"Mistral Medium 3.1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-06-30","release_date":"2025-08-12","last_updated":"2025-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.4,"output":2,"cache_read":0.2},"canonical_model_id":"mistral/mistral-medium-3.1"},"mistralai/mistral-nemo-instruct-2407":{"id":"mistralai/mistral-nemo-instruct-2407","name":"Mistral Nemo","description":"12B parameter model with multilingual support.","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.1003,"output":0.1207,"cache_read":0.05015},"canonical_model_id":"mistral/mistral-nemo"},"mistralai/ministral-14b-2512":{"id":"mistralai/ministral-14b-2512","name":"Ministral 14B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.2,"output":0.2,"cache_read":0.1},"canonical_model_id":"mistral/ministral-14b-2512"},"mistralai/mistral-small-4-119b-2603:thinking":{"id":"mistralai/mistral-small-4-119b-2603:thinking","name":"Mistral Small 4 119B Thinking","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.4,"output":1.4,"cache_read":0.2},"canonical_model_id":"mistral/mistral-small-2603"},"mistralai/mistral-medium-3.5":{"id":"mistralai/mistral-medium-3.5","name":"Mistral Medium 3.5","description":"Mistral Medium 3.5 is a 128B dense open-weights flagship model for instruction-following, reasoning, coding, long-horizon agentic work, tool use, structured output, and multimodal prompts. It supports a 256k context window and configurable reasoning effort.","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":1.5,"output":7.5,"cache_read":0.75}},"mistralai/mistral-large":{"id":"mistralai/mistral-large","name":"Mistral Large 2411","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-11-30","release_date":"2024-02-26","last_updated":"2024-02-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":102400},"cost":{"input":2.006,"output":6.001,"cache_read":0.2},"canonical_model_id":"mistral/mistral-large"},"mistralai/mistral-small-3.2-24b-instruct":{"id":"mistralai/mistral-small-3.2-24b-instruct","name":"Mistral Small 3.2 24B (2506)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.2,"output":0.4,"cache_read":0.1},"canonical_model_id":"mistral/mistral-small-3.2-24b-instruct"},"mistralai/mistral-small-4-119b-2603":{"id":"mistralai/mistral-small-4-119b-2603","name":"Mistral Small 4 119B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.4,"output":1.4,"cache_read":0.2},"canonical_model_id":"mistral/mistral-small-2603"},"mistralai/mixtral-8x22b-instruct-v0.1":{"id":"mistralai/mixtral-8x22b-instruct-v0.1","name":"Mixtral 8x22B","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-01-31","release_date":"2024-04-17","last_updated":"2024-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":52428},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mixtral-8x22b-instruct"},"mistralai/mistral-small-24b-instruct-2501":{"id":"mistralai/mistral-small-24b-instruct-2501","name":"Mistral Small 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-01-30","last_updated":"2025-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.1155,"output":0.3465},"canonical_model_id":"mistral/mistral-small-24b-instruct-2501"},"bytedance-seed/seed-2-1-turbo":{"id":"bytedance-seed/seed-2-1-turbo","name":"ByteDance Seed 2.1 Turbo","description":"ByteDance Seed 2.1 Turbo is a multimodal model for coding and long-horizon agent workflows, including end-to-end software delivery and multi-step task execution. It supports text, image, and video input with a 262k context window.","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":235929},"cost":{"input":0.5,"output":2.5,"cache_read":0.25},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"bytedance-seed/seed-2.0-code":{"id":"bytedance-seed/seed-2.0-code","name":"ByteDance Seed 2.0 Code","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.5,"output":3,"cache_read":0.25},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"bytedance-seed/seed-2.0-lite":{"id":"bytedance-seed/seed-2.0-lite","name":"ByteDance Seed 2.0 Lite","description":"Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation","family":"seed","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.25,"output":2,"cache_read":0.125},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"shisa-ai/shisa-v2-llama3.3-70b":{"id":"shisa-ai/shisa-v2-llama3.3-70b","name":"Shisa V2 Llama 3.3 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.5,"output":0.5,"cache_read":0.25}},"shisa-ai/shisa-v2.1-llama3.3-70b":{"id":"shisa-ai/shisa-v2.1-llama3.3-70b","name":"Shisa V2.1 Llama 3.3 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":4096},"cost":{"input":0.5,"output":0.5,"cache_read":0.25}},"NeverSleep/Lumimaid-v0.2-70B":{"id":"NeverSleep/Lumimaid-v0.2-70B","name":"Lumimaid v0.2","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":1,"output":1.5,"cache_read":0.5}},"undi95/remm-slerp-l2-13b":{"id":"undi95/remm-slerp-l2-13b","name":"ReMM SLERP 13B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-06-30","release_date":"2023-07-22","last_updated":"2023-07-22","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":6144,"input":6144,"output":4096},"cost":{"input":0.799,"output":1.207,"cache_read":0.3995},"canonical_model_id":"undi95/remm-slerp-l2-13b"},"reka/reka-edge-2603":{"id":"reka/reka-edge-2603","name":"Reka Edge 2603","description":"Reka's compact multimodal model for visual understanding and tool use, served directly by Reka.","family":"reka","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":16384},"cost":{"input":0.105,"output":0.105},"canonical_model_id":"rekaai/reka-edge"},"reka/reka-flash-3":{"id":"reka/reka-flash-3","name":"Reka Flash 3","description":"Reka's 21B reasoning model for chat and coding, served directly by Reka.","family":"reka","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01-31","release_date":"2025-03-10","last_updated":"2025-03-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":65536},"cost":{"input":0.105,"output":0.21},"canonical_model_id":"rekaai/reka-flash-3"},"NousResearch/hermes-4-405b:thinking":{"id":"NousResearch/hermes-4-405b:thinking","name":"Hermes 4 Large (Thinking)","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"hermes","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.15},"canonical_model_id":"nousresearch/hermes-4-405b"},"NousResearch/hermes-4-405b":{"id":"NousResearch/hermes-4-405b","name":"Hermes 4 Large","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"hermes","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.3,"output":1.2,"cache_read":0.15},"canonical_model_id":"nousresearch/hermes-4-405b"},"NousResearch/hermes-3-llama-3.1-70b":{"id":"NousResearch/hermes-3-llama-3.1-70b","name":"Hermes 3 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-18","last_updated":"2024-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.408,"output":0.408,"cache_read":0.204},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-70b"},"MarinaraSpaghetti/NemoMix-Unleashed-12B":{"id":"MarinaraSpaghetti/NemoMix-Unleashed-12B","name":"NemoMix 12B Unleashed","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"MiniMaxAI/MiniMax-M1-80k":{"id":"MiniMaxAI/MiniMax-M1-80k","name":"MiniMax M1 80K","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":131072},"cost":{"input":0.6052,"output":2.4225,"cache_read":0.3026},"canonical_model_id":"minimax/minimax-m1-80k"},"inclusionai/ling-3.0-flash":{"id":"inclusionai/ling-3.0-flash","name":"Ling 3.0 Flash","description":"Hybrid-linear MoE reasoning model with 124B total and 5.1B active parameters for efficient agents","family":"ling","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash"},"inclusionai/ling-3.0-flash-vl":{"id":"inclusionai/ling-3.0-flash-vl","name":"Ling 3.0 Flash VL","description":"Native multimodal hybrid MoE model for image and video understanding, reasoning, and visual agents","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012},"canonical_model_id":"inclusionai/ling-3.0-flash-vl"},"inclusionai/ling-3.1-flash":{"id":"inclusionai/ling-3.1-flash","name":"Ling 3.1 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.1-flash"},"inclusionai/ling-3.0-flash:thinking":{"id":"inclusionai/ling-3.0-flash:thinking","name":"Ling 3.0 Flash Thinking","description":"Hybrid-linear MoE reasoning model with 124B total and 5.1B active parameters for efficient agents","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.075,"output":0.22,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.0-flash"},"Steelskull/L3.3-MS-Nevoria-70b":{"id":"Steelskull/L3.3-MS-Nevoria-70b","name":"Steelskull Nevoria 70b","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"Steelskull/L3.3-Nevoria-R1-70b":{"id":"Steelskull/L3.3-Nevoria-R1-70b","name":"Steelskull Nevoria R1 70b","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"Steelskull/L3.3-Cu-Mai-R1-70b":{"id":"Steelskull/L3.3-Cu-Mai-R1-70b","name":"Llama 3.3 70B Cu Mai","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"Steelskull/L3.3-Electra-R1-70b":{"id":"Steelskull/L3.3-Electra-R1-70b","name":"Steelskull Electra R1 70b","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.69989,"output":0.69989,"cache_read":0.349945}},"ornith-ai/ornith-1.5-122b":{"id":"ornith-ai/ornith-1.5-122b","name":"Ornith 1.5 122B","description":"Ornith 1.5 122B is an MIT-licensed open-weight model for coding, reasoning, tool use and image understanding. It accepts text and images across a shared 262,144-token context, with up to 32,768 output tokens. Reasoning is enabled by default; JSON object output and streaming are supported.","family":"ornith","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-10-20","last_updated":"2026-10-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.2,"output":0.8,"cache_read":0.02}},"ornith-ai/ornith-1.5-35b-a3b:thinking":{"id":"ornith-ai/ornith-1.5-35b-a3b:thinking","name":"Ornith 1.5 35B Thinking","description":"Ornith 1.5 35B A3B is an open-weight mixture-of-experts model for agentic coding, reasoning, tool use, image understanding, and long-context work. This variant enables thinking by default.","family":"ornith","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-07-29","last_updated":"2026-08-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.05}},"ornith-ai/ornith-1.5-35b-a3b":{"id":"ornith-ai/ornith-1.5-35b-a3b","name":"Ornith 1.5 35B","description":"Ornith 1.5 35B A3B is an open-weight mixture-of-experts model for agentic coding, tool use, image understanding, and long-context work. This variant disables thinking for faster direct responses.","family":"ornith","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-07-29","last_updated":"2026-08-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.05}},"featherless-ai/Qwerky-72B":{"id":"featherless-ai/Qwerky-72B","name":"Qwerky 72B","description":"General-purpose chat model for instruction following, writing, and analysis","family":"qwerky","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-03-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"input":32000,"output":8192},"cost":{"input":0.5,"output":0.5,"cache_read":0.25}},"liquid/lfm-2.5-2.6b":{"id":"liquid/lfm-2.5-2.6b","name":"LFM2.5 2.6B","description":"Liquid AI's compact 2.6B reasoning model for agent workflows, data extraction, RAG, and long-context processing. It supports tool calling and structured output, but Liquid advises against using it for agentic coding. Warning: prompts and responses may be logged and used for model training or service improvement; do not send sensitive data.","family":"liquid","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":32768},"cost":{"input":0.1,"output":0.2,"cache_read":0.05}},"liquid/lfm-2.5-24b-a2b":{"id":"liquid/lfm-2.5-24b-a2b","name":"LFM2.5 24B A2B","description":"Liquid AI's text-only mixture-of-experts model for reasoning and tool-using agents, with 24B total and 2.45B active parameters. It always thinks before answering and supports structured output. Pair it with retrieval for knowledge-heavy tasks. Open weights use the LFM Open License v1.0, with a $10M annual-revenue commercial threshold; hosted API use is subject to separate service terms that permit retention and use of prompts and responses for model training.","family":"liquid","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-10-14","last_updated":"2026-10-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.05,"output":0.7,"cache_read":0.025}},"THUDM/GLM-4-32B-0414":{"id":"THUDM/GLM-4-32B-0414","name":"GLM 4 32B 0414","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-04-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.2,"output":0.2,"cache_read":0.1}},"THUDM/GLM-4-9B-0414":{"id":"THUDM/GLM-4-9B-0414","name":"GLM 4 9B 0414","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-04-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"input":32000,"output":8000},"cost":{"input":0.2,"output":0.2,"cache_read":0.1}},"THUDM/GLM-Z1-9B-0414":{"id":"THUDM/GLM-Z1-9B-0414","name":"GLM Z1 9B 0414","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm-z","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-04-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"input":32000,"output":8000},"cost":{"input":0.2,"output":0.2,"cache_read":0.1}},"poolside/laguna-s-2.1:thinking":{"id":"poolside/laguna-s-2.1:thinking","name":"Laguna S 2.1 Thinking","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.01},"canonical_model_id":"poolside/laguna-s-2.1"},"poolside/laguna-s-2.1":{"id":"poolside/laguna-s-2.1","name":"Laguna S 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.01},"canonical_model_id":"poolside/laguna-s-2.1"},"LatitudeGames/Wayfarer-Large-70B-Llama-3.3":{"id":"LatitudeGames/Wayfarer-Large-70B-Llama-3.3","name":"Llama 3.3 70B Wayfarer","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-02-20","last_updated":"2025-02-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.7,"output":0.7,"cache_read":0.35}},"Salesforce/Llama-xLAM-2-70b-fc-r":{"id":"Salesforce/Llama-xLAM-2-70b-fc-r","name":"Llama-xLAM-2 70B fc-r","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-04-13","last_updated":"2025-04-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":2.5,"output":2.5,"cache_read":1.25}},"thinkingmachines/Inkling-Small:thinking":{"id":"thinkingmachines/Inkling-Small:thinking","name":"Inkling Small Thinking","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"input":524288,"output":32768},"cost":{"input":0.45,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/inkling:thinking":{"id":"thinkingmachines/inkling:thinking","name":"Inkling Thinking","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048000,"input":1048000,"output":32768},"cost":{"input":0.95,"output":4.05,"cache_read":0.16},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/Inkling-Small":{"id":"thinkingmachines/Inkling-Small","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"input":524288,"output":32768},"cost":{"input":0.45,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048000,"input":1048000,"output":32768},"cost":{"input":0.95,"output":4.05,"cache_read":0.16},"canonical_model_id":"thinkingmachines/inkling"},"failspy/Meta-Llama-3-70B-Instruct-abliterated-v3.5":{"id":"failspy/Meta-Llama-3-70B-Instruct-abliterated-v3.5","name":"Llama 3 70B abliterated","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"input":8192,"output":8192},"cost":{"input":0.7,"output":0.7,"cache_read":0.35}},"mlabonne/NeuralDaredevil-8B-abliterated":{"id":"mlabonne/NeuralDaredevil-8B-abliterated","name":"Neural Daredevil 8B abliterated","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"input":8192,"output":8192},"cost":{"input":0.44,"output":0.44,"cache_read":0.22}},"inflatebot/MN-12B-Mag-Mell-R1":{"id":"inflatebot/MN-12B-Mag-Mell-R1","name":"Mag Mell R1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"abliteration-ai/abliterated-model":{"id":"abliteration-ai/abliterated-model","name":"Abliterated Model","description":"Abliteration.ai's multimodal reasoning model supports text and image input, structured output, automatic prompt caching, and a 262K-token context window.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":false,"structured_output":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":262134},"cost":{"input":1,"output":3,"cache_read":0.1}},"abliteration-ai/abliterated-model-large":{"id":"abliteration-ai/abliterated-model-large","name":"Abliterated Model Large","description":"Abliteration.ai's large text reasoning model is derived from GLM-5.2 and supports native tool calling, structured output, automatic prompt caching, and a one-million-token context window.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":999990},"cost":{"input":3,"output":5,"cache_read":0.3}},"abliteration-ai/abliterated-model-large-v2":{"id":"abliteration-ai/abliterated-model-large-v2","name":"Abliterated Model Large V2","description":"Abliteration.ai's default large text reasoning model is derived from GLM-5.3 for harder reasoning and evaluation workloads, with automatic prompt caching and a one-million-token context window.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-08-31","last_updated":"2026-08-31","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":999990},"cost":{"input":3,"output":5,"cache_read":0.3}},"unsloth/gemma-3-27b-it":{"id":"unsloth/gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":96000},"cost":{"input":0.2992,"output":0.2992,"cache_read":0.1496},"canonical_model_id":"google/gemma-3-27b-it"},"unsloth/gemma-3-4b-it":{"id":"unsloth/gemma-3-4b-it","name":"Gemma 3 4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.2006,"output":0.2006,"cache_read":0.1003},"canonical_model_id":"google/gemma-3-4b-it"},"unsloth/gemma-3-12b-it":{"id":"unsloth/gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":16384},"cost":{"input":0.272,"output":0.272,"cache_read":0.136},"canonical_model_id":"google/gemma-3-12b-it"},"aion-labs/aion-3.5":{"id":"aion-labs/aion-3.5","name":"AionLabs: Aion 3.5","description":"A GLM-family collaborative generation model tuned for immersive roleplay and storytelling, with stronger narrative structure, tension, conflict, and nuanced mature themes.","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":3,"output":6,"cache_read":0.75}},"aion-labs/aion-3.0-mini":{"id":"aion-labs/aion-3.0-mini","name":"AionLabs: Aion 3.0 Mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-14","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.7,"output":1.4,"cache_read":0.18},"canonical_model_id":"aion-labs/aion-3.0-mini"},"aion-labs/aion-2.0":{"id":"aion-labs/aion-2.0","name":"AionLabs: Aion-2.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-21","last_updated":"2025-12-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.8,"output":1.6,"cache_read":0.2},"canonical_model_id":"aion-labs/aion-2.0"},"aion-labs/aion-rp-llama-3.1-8b":{"id":"aion-labs/aion-rp-llama-3.1-8b","name":"Llama 3.1 8b (uncensored)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12-31","release_date":"2025-02-04","last_updated":"2025-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.8,"output":1.6,"cache_read":0.4},"canonical_model_id":"aion-labs/aion-rp-llama-3.1-8b"},"aion-labs/aion-3.5-mini":{"id":"aion-labs/aion-3.5-mini","name":"AionLabs: Aion 3.5 Mini","description":"A GLM-family collaborative generation model tuned for immersive roleplay and storytelling, with stronger narrative structure, tension, conflict, and nuanced mature themes.","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.7,"output":1.4,"cache_read":0.18}},"aion-labs/aion-3.0":{"id":"aion-labs/aion-3.0","name":"AionLabs: Aion 3.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-05","last_updated":"2026-05-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":3,"output":6,"cache_read":0.75},"canonical_model_id":"aion-labs/aion-3.0"},"inception/mercury-3":{"id":"inception/mercury-3","name":"Mercury 3","description":"Inception's diffusion reasoning model for chat, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-16","last_updated":"2026-10-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":131072},"cost":{"input":0.1,"output":0.35,"cache_read":0.01},"canonical_model_id":"inception/mercury-3"},"inception/mercury-2.5-preview":{"id":"inception/mercury-2.5-preview","name":"Mercury 2.5 Preview","description":"Mercury 2.5 Preview is Inception's latest and most intelligent diffusion language model. Instead of generating tokens strictly one at a time, it produces and refines multiple tokens in parallel, reaching up to 1,107 tokens per second on standard GPUs. It delivers a 10+ point intelligence gain over Mercury 2, with tunable reasoning, parallel tool calls, schema-aligned JSON output, and a 260K context window. It is built for latency-sensitive production work such as search agents, voice pipelines, customer support, rapid coding iteration, and coding subagents.","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":260000,"input":260000,"output":65536},"cost":{"input":0.04,"output":0.15,"cache_read":0.004}},"Envoid/Llama-3.05-NT-Storybreaker-Ministral-70B":{"id":"Envoid/Llama-3.05-NT-Storybreaker-Ministral-70B","name":"Llama 3.05 Storybreaker Ministral 70b","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"Envoid/Llama-3.05-Nemotron-Tenyxchat-Storybreaker-70B":{"id":"Envoid/Llama-3.05-Nemotron-Tenyxchat-Storybreaker-70B","name":"Nemotron Tenyxchat Storybreaker 70b","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"baseten/Kimi-K2-Instruct-FP4":{"id":"baseten/Kimi-K2-Instruct-FP4","name":"Kimi K2 0711 Instruct FP4","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":131072},"status":"deprecated","cost":{"input":0.4,"output":1.8,"cache_read":0.2}},"VongolaChouko/Starcannon-Unleashed-12B-v1.0":{"id":"VongolaChouko/Starcannon-Unleashed-12B-v1.0","name":"Mistral Nemo Starcannon 12b v1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"openai/gpt-terra-latest":{"id":"openai/gpt-terra-latest","name":"GPT Terra Latest","description":"Compatibility alias that routes to GPT 5.6 Terra, the latest supported GPT Terra model.","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5}},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT 6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT 5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT 4.1 Mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"input":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-6-sol-pro":{"id":"openai/gpt-6-sol-pro","name":"GPT 6 Sol Pro","description":"GPT-6 Sol Pro uses the same underlying model as GPT-6 Sol with Pro reasoning mode enabled for higher-quality responses on complex tasks.","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5}},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT 5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":15,"output":120,"cache_read":1.5},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT 6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"openai/gpt-6-astra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"OpenAI o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":16384},"cost":{"input":0.02,"output":0.1},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-6.1-sol-pro":{"id":"openai/gpt-6.1-sol-pro","name":"GPT 6.1 Sol Pro","description":"GPT-6.1 Sol Pro uses the same underlying model as GPT-6.1 Sol with Pro reasoning mode enabled for higher-quality responses on complex tasks.","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5}},"openai/gpt-latest":{"id":"openai/gpt-latest","name":"GPT Latest","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-29","last_updated":"2026-03-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5}},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT 5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-6-luna-pro":{"id":"openai/gpt-6-luna-pro","name":"GPT 6 Luna Pro","description":"GPT-6 Luna Pro uses the same underlying model as GPT-6 Luna with Pro reasoning mode enabled for higher-quality responses on complex tasks.","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125}},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT 5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT 5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/o1-pro":{"id":"openai/o1-pro","name":"OpenAI o1 Pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2023-09","release_date":"2025-03-19","last_updated":"2025-03-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":150,"output":600,"cache_read":75},"canonical_model_id":"openai/o1-pro"},"openai/gpt-chat-latest":{"id":"openai/gpt-chat-latest","name":"GPT Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-03","last_updated":"2026-05-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5}},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT 5.2 Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT 5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-astra-latest":{"id":"openai/gpt-astra-latest","name":"GPT Astra Latest","description":"Compatibility alias that routes to GPT 6 Astra, the latest supported GPT Astra model.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5}},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT 5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-oss-safeguard-20b":{"id":"openai/gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.075,"output":0.3},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT 5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"openai/o3":{"id":"openai/o3","name":"OpenAI o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":1},"canonical_model_id":"openai/o3"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT 5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT 6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-luna-latest":{"id":"openai/gpt-luna-latest","name":"GPT Luna Latest","description":"Compatibility alias that routes to GPT 5.6 Luna, the latest supported GPT Luna model.","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125}},"openai/gpt-6-astra-pro":{"id":"openai/gpt-6-astra-pro","name":"GPT 6 Astra Pro","description":"GPT 6 Astra in Pro reasoning mode. Uses additional model work for difficult tasks, with higher latency and token usage at the same per-token rates. Reasoning effort remains independently configurable.","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5}},"openai/o3-pro-2025-06-10":{"id":"openai/o3-pro-2025-06-10","name":"OpenAI o3-pro (2025-06-10)","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":22,"output":88,"cache_read":11},"canonical_model_id":"openai/o3-pro"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.03,"output":0.15,"cache_read":0.03},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT 5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT 6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT 5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-5.6-luna-pro":{"id":"openai/gpt-5.6-luna-pro","name":"GPT 5.6 Luna Pro","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT 4.1 Nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"input":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-sol-latest":{"id":"openai/gpt-sol-latest","name":"GPT Sol Latest","description":"Compatibility alias that routes to GPT 5.6 Sol, the latest supported GPT Sol model.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5}},"openai/gpt-4o-2024-11-20":{"id":"openai/gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"openai/gpt-5.6-sol-pro":{"id":"openai/gpt-5.6-sol-pro","name":"GPT 5.6 Sol Pro","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-4o-2024-08-06":{"id":"openai/gpt-4o-2024-08-06","name":"GPT-4o (2024-08-06)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-08-06"},"openai/o3-mini-low":{"id":"openai/o3-mini-low","name":"OpenAI o3-mini (Low)","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/o4-mini":{"id":"openai/o4-mini","name":"OpenAI o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o4-mini"},"openai/gpt-5.1-2025-11-13":{"id":"openai/gpt-5.1-2025-11-13","name":"GPT-5.1 (2025-11-13)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"input":16385,"output":4096},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT 4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"input":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/o1":{"id":"openai/o1","name":"OpenAI o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"openai/gpt-5.1-codex-max":{"id":"openai/gpt-5.1-codex-max","name":"GPT 5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":2.5,"output":20,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.1-codex-max"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT 5.1 Codex Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/o4-mini-high":{"id":"openai/o4-mini-high","name":"OpenAI o4-mini high","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o4-mini"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT 5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT 5.1 Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"openai/o3-mini-high":{"id":"openai/o3-mini-high","name":"OpenAI o3-mini (High)","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT 5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.6-terra-pro":{"id":"openai/gpt-5.6-terra-pro","name":"GPT 5.6 Terra Pro","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":1050000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT 5.4 Nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":400000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-v4-flash-latest":{"id":"deepseek/deepseek-v4-flash-latest","name":"DeepSeek V4 Flash Latest","description":"Compatibility alias that routes to the newest dated DeepSeek V4 Flash release. Currently routes to DeepSeek V4 Flash 0731. ⚠️ This route goes directly to DeepSeek, so privacy and logging guarantees are limited.","family":"deepseek","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-08-02","last_updated":"2026-08-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":393216},"cost":{"input":0.1,"output":0.6,"cache_read":0.003}},"deepseek/deepseek-v4.1-flash-1017:thinking":{"id":"deepseek/deepseek-v4.1-flash-1017:thinking","name":"DeepSeek V4.1 Flash 1017 Thinking","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":393216},"cost":{"input":0.1,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":163000,"input":163000,"output":65536},"cost":{"input":0.2088,"output":0.3096,"cache_read":0.0216},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":393216},"cost":{"input":0.1,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.1452,"output":0.4356,"cache_read":0.00462},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.2156,"output":0.6468,"cache_read":0.00686},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.7,"output":1.4,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash:thinking":{"id":"deepseek/deepseek-v4-flash:thinking","name":"DeepSeek V4 Flash (Thinking)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4-pro:thinking":{"id":"deepseek/deepseek-v4-pro:thinking","name":"DeepSeek V4 Pro (Thinking)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.7,"output":1.4,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash-0731:thinking":{"id":"deepseek/deepseek-v4-flash-0731:thinking","name":"DeepSeek V4 Flash 0731 (Thinking)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.04444,"output":0.13332,"cache_read":0.001414},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-v4-flash-0423:thinking":{"id":"deepseek/deepseek-v4-flash-0423:thinking","name":"DeepSeek V4 Flash 0423 (Thinking)","description":"Initial DeepSeek V4 Flash snapshot for economical reasoning, coding, and million-token agent workloads","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0423"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":384000},"cost":{"input":0.0342,"output":0.1368,"cache_read":0.000684},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4.1-flash:thinking":{"id":"deepseek/deepseek-v4.1-flash:thinking","name":"DeepSeek V4.1 Flash Thinking","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":384000},"cost":{"input":0.0342,"output":0.1368,"cache_read":0.000684},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.04444,"output":0.13332,"cache_read":0.001414},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-latest":{"id":"deepseek/deepseek-latest","name":"DeepSeek Latest","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-03","last_updated":"2026-05-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.1452,"output":0.4356,"cache_read":0.00462}},"deepseek/deepseek-v3.2:thinking":{"id":"deepseek/deepseek-v3.2:thinking","name":"DeepSeek V3.2 Thinking","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":163000,"input":163000,"output":65536},"cost":{"input":0.2088,"output":0.3096,"cache_read":0.0216},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4-pro-0813:thinking":{"id":"deepseek/deepseek-v4-pro-0813:thinking","name":"DeepSeek V4 Pro 0813 Thinking","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.1452,"output":0.4356,"cache_read":0.00462},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-flash-0423":{"id":"deepseek/deepseek-v4-flash-0423","name":"DeepSeek V4 Flash 0423","description":"Initial DeepSeek V4 Flash snapshot for economical reasoning, coding, and million-token agent workloads","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0423"},"anthropic/claude-haiku-4.5:thinking":{"id":"anthropic/claude-haiku-4.5:thinking","name":"Claude Haiku 4.5 Thinking","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-sonnet-4.5:thinking":{"id":"anthropic/claude-sonnet-4.5:thinking","name":"Claude Sonnet 4.5 Thinking","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-sonnet-4.6:thinking":{"id":"anthropic/claude-sonnet-4.6:thinking","name":"Claude Sonnet 4.6 Thinking","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-opus-4.1":{"id":"anthropic/claude-opus-4.1","name":"Claude 4.1 Opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5},"canonical_model_id":"anthropic/claude-opus-4-1"},"anthropic/claude-opus-4.6:thinking:low":{"id":"anthropic/claude-opus-4.6:thinking:low","name":"Claude 4.6 Opus Thinking Low","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-fable-latest":{"id":"anthropic/claude-fable-latest","name":"Claude Fable Latest","description":"Compatibility alias for Claude Fable.","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5}},"anthropic/claude-opus-4.8:thinking":{"id":"anthropic/claude-opus-4.8:thinking","name":"Claude Opus 4.8 Thinking","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-opus-4.6:thinking:medium":{"id":"anthropic/claude-opus-4.6:thinking:medium","name":"Claude 4.6 Opus Thinking Medium","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude 4.7 Opus","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-4.1:thinking":{"id":"anthropic/claude-opus-4.1:thinking","name":"Claude 4.1 Opus Thinking","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5},"canonical_model_id":"anthropic/claude-opus-4-1"},"anthropic/claude-opus-4.6:thinking":{"id":"anthropic/claude-opus-4.6:thinking","name":"Claude 4.6 Opus Thinking","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-opus-4.1:thinking:8192":{"id":"anthropic/claude-opus-4.1:thinking:8192","name":"Claude 4.1 Opus Thinking (8K)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5},"canonical_model_id":"anthropic/claude-opus-4-1"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude 4.6 Opus","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude 4.5 Opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-opus-4.7:thinking":{"id":"anthropic/claude-opus-4.7:thinking","name":"Claude 4.7 Opus Thinking","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-opus-4.1:thinking:1024":{"id":"anthropic/claude-opus-4.1:thinking:1024","name":"Claude 4.1 Opus Thinking (1K)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5},"canonical_model_id":"anthropic/claude-opus-4-1"},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-sonnet-5.5:thinking":{"id":"anthropic/claude-sonnet-5.5:thinking","name":"Claude Sonnet 5.5 Thinking","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-haiku-latest":{"id":"anthropic/claude-haiku-latest","name":"Claude Haiku Latest","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-29","last_updated":"2026-03-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]}},"anthropic/claude-opus-4.6:thinking:max":{"id":"anthropic/claude-opus-4.6:thinking:max","name":"Claude 4.6 Opus Thinking Max","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-4.1:thinking:32768":{"id":"anthropic/claude-opus-4.1:thinking:32768","name":"Claude 4.1 Opus Thinking (32K)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5},"canonical_model_id":"anthropic/claude-opus-4-1"},"anthropic/claude-sonnet-5:thinking":{"id":"anthropic/claude-sonnet-5:thinking","name":"Claude Sonnet 5 Thinking","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-opus-4.5:thinking":{"id":"anthropic/claude-opus-4.5:thinking","name":"Claude 4.5 Opus Thinking","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-opus-latest":{"id":"anthropic/claude-opus-latest","name":"Claude Opus Latest","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-29","last_updated":"2026-03-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5}},"anthropic/claude-sonnet-latest":{"id":"anthropic/claude-sonnet-latest","name":"Claude Sonnet Latest","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-01","last_updated":"2026-03-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5}},"Sao10K/L3.3-70B-Euryale-v2.3":{"id":"Sao10K/L3.3-70B-Euryale-v2.3","name":"Llama 3.3 70B Euryale","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-12-18","last_updated":"2024-12-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":20480,"input":20480,"output":16384},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465},"canonical_model_id":"sao10k/l3.3-euryale-70b"},"Sao10K/L3.1-70B-Hanami-x1":{"id":"Sao10K/L3.1-70B-Hanami-x1","name":"Llama 3.1 70B Hanami","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"Sao10K/L3.1-70B-Euryale-v2.2":{"id":"Sao10K/L3.1-70B-Euryale-v2.2","name":"Llama 3.1 70B Euryale","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-28","last_updated":"2024-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":20480,"input":20480,"output":16384},"cost":{"input":0.306,"output":0.357,"cache_read":0.153},"canonical_model_id":"sao10k/l3.1-euryale-70b"},"Sao10K/L3-8B-Stheno-v3.2":{"id":"Sao10K/L3-8B-Stheno-v3.2","name":"Sao10K Stheno 8b","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-11-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.2006,"output":0.2006,"cache_read":0.1003}},"nanogpt/coding-router:medium":{"id":"nanogpt/coding-router:medium","name":"Coding Router Medium","description":"Automatic model router for matching prompts to suitable backends and budgets","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":0.14,"output":0.28,"cache_read":0.028}},"nanogpt/coding-router":{"id":"nanogpt/coding-router","name":"Coding Router","description":"Automatic model router for matching prompts to suitable backends and budgets","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":1.1,"output":2.2,"cache_read":0.11}},"nanogpt/coding-router:max":{"id":"nanogpt/coding-router:max","name":"Coding Router Max","description":"Automatic model router for matching prompts to suitable backends and budgets","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5}},"nanogpt/coding-router:low":{"id":"nanogpt/coding-router:low","name":"Coding Router Low","description":"Automatic model router for matching prompts to suitable backends and budgets","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":0.14,"output":0.28,"cache_read":0.028}},"nanogpt/coding-router:high":{"id":"nanogpt/coding-router:high","name":"Coding Router High","description":"Automatic model router for matching prompts to suitable backends and budgets","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":128000},"cost":{"input":1.1,"output":2.2,"cache_read":0.11}},"Gryphe/MythoMax-L2-13b":{"id":"Gryphe/MythoMax-L2-13b","name":"MythoMax 13B","description":"Llama 2 merge combining MythoLogic and Huginn for creative writing and roleplay","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2023-08-10","last_updated":"2023-08-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"input":4096,"output":3686},"cost":{"input":0.1003,"output":0.1003,"cache_read":0.05015},"canonical_model_id":"gryphe/mythomax-l2-13b"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Tencent Hy4 Preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":64000},"cost":{"input":0.7506,"output":2.2509,"cache_read":0.0378},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Tencent Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":128000},"cost":{"input":0.066,"output":0.26,"cache_read":0.029},"canonical_model_id":"tencent/hy3"},"ibm-granite/granite-4.2-8b":{"id":"ibm-granite/granite-4.2-8b","name":"Granite 4.2 8B","description":"IBM Granite 4.2 8B is an Apache 2.0-licensed dense model with native step-by-step reasoning and specialized training for agentic work. It can plan before acting, sequence tools, navigate codebases, work in terminals, and verify results across coding, search, mathematics, science, and complex instruction-following tasks.","family":"granite","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-25","last_updated":"2026-08-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":117964},"cost":{"input":0.1,"output":0.15,"cache_read":0.05},"canonical_model_id":"ibm/granite-4.2-8b"},"moondream/moondream3.1-9b-a2b":{"id":"moondream/moondream3.1-9b-a2b","name":"Moondream 3.1 9B A2B","description":"Moondream 3.1 is the successor to Moondream 3 Preview: a distinct 9B mixture-of-experts vision-language checkpoint with 2B active parameters, for visual question answering, captioning, OCR, and visual reasoning. It accepts text or one image with optional reasoning, a 32,000-token context, and up to 4,096 output tokens.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"input":32000,"output":4096},"cost":{"input":0.3,"output":2.5}},"moondream/moondream3-preview":{"id":"moondream/moondream3-preview","name":"Moondream 3 Preview","description":"Moondream 3 Preview is the original 9B mixture-of-experts vision-language checkpoint with 2B active parameters, for visual question answering, captioning, OCR, and visual reasoning. It accepts text or one image with optional reasoning, a 32,000-token context, and up to 4,096 output tokens.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"input":32000,"output":4096},"cost":{"input":0.3,"output":2.5}},"LLM360/K2-Think":{"id":"LLM360/K2-Think","name":"K2-Think","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":32768},"cost":{"input":0.17,"output":0.68,"cache_read":0.085}},"meta-llama/llama-3.1-8b-instruct":{"id":"meta-llama/llama-3.1-8b-instruct","name":"Llama 3.1 8b Instruct","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":16384},"cost":{"input":0.02,"output":0.04,"cache_read":0.0272},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta-llama/llama-3.2-3b-instruct":{"id":"meta-llama/llama-3.2-3b-instruct","name":"Llama 3.2 3b Instruct","description":"Small instruction-tuned Llama model for multilingual dialogue and agentic retrieval","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":8192},"cost":{"input":0.0306,"output":0.0493,"cache_read":0.0153},"canonical_model_id":"meta/llama-3.2-3b-instruct"},"meta-llama/llama-3.3-70b-instruct":{"id":"meta-llama/llama-3.3-70b-instruct","name":"Llama 3.3 70b Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":16384},"cost":{"input":0.05,"output":0.23,"cache_read":0.025},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/llama-4-scout":{"id":"meta-llama/llama-4-scout","name":"Llama 4 Scout","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":328000,"input":328000,"output":65536},"cost":{"input":0.085,"output":0.46,"cache_read":0.0425},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"meta-llama/llama-4-maverick":{"id":"meta-llama/llama-4-maverick","name":"Llama 4 Maverick","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.1875,"output":0.6525,"cache_read":0.05},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"xiaomi/mimo-v2.6-pro-ultraspeed":{"id":"xiaomi/mimo-v2.6-pro-ultraspeed","name":"MiMo V2.6 Pro UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":4.35,"output":8.7,"cache_read":0.036,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2.5-pro":{"id":"xiaomi/mimo-v2.5-pro","name":"MiMo V2.5 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.6-flash-uncensored":{"id":"xiaomi/mimo-v2.6-flash-uncensored","name":"MiMo V2.6 Flash Uncensored","description":"A text-only, lower-refusal MiMo V2.6 Flash finetune served by Heabsy with a 1M-token context window, separate reasoning output, and tool calling.","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.5,"output":1.5,"cache_read":0.1995}},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo V2.6 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"xiaomi/mimo-v2.6-flash":{"id":"xiaomi/mimo-v2.6-flash","name":"MiMo V2.6 Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"xiaomi/mimo-v2.5-pro:thinking":{"id":"xiaomi/mimo-v2.5-pro:thinking","name":"MiMo V2.5 Pro Thinking","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.6-flash-abliterated":{"id":"xiaomi/mimo-v2.6-flash-abliterated","name":"MiMo V2.6 Flash Abliterated","description":"A MiMo V2.6 Flash finetune with refusal-direction ablation, image input, a 1M-token context window, separate reasoning output, and tool calling.","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.5,"output":1.5,"cache_read":0.275}},"xiaomi/mimo-v2.5:thinking":{"id":"xiaomi/mimo-v2.5:thinking","name":"MiMo V2.5 Thinking","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028,"cache_write":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2.6-flash-uncensored:thinking":{"id":"xiaomi/mimo-v2.6-flash-uncensored:thinking","name":"MiMo V2.6 Flash Uncensored Thinking","description":"MiMo V2.6 Flash Uncensored with maximum thinking enabled. Supports a 1M-token context window and tool calling.","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["max"]}],"tool_call":true,"structured_output":false,"release_date":"2026-09-26","last_updated":"2026-09-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.5,"output":1.5,"cache_read":0.1995}},"Doctor-Shotgun/MS3.2-24B-Magnum-Diamond":{"id":"Doctor-Shotgun/MS3.2-24B-Magnum-Diamond","name":"MS3.2 24B Magnum Diamond","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"anthracite-org/magnum-v4-72b":{"id":"anthracite-org/magnum-v4-72b","name":"Magnum v4 72B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":2.006,"output":2.992,"cache_read":1.003},"canonical_model_id":"anthracite-org/magnum-v4-72b"},"anthracite-org/magnum-v2-72b":{"id":"anthracite-org/magnum-v2-72b","name":"Magnum V2 72B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":2.006,"output":2.992,"cache_read":1.003}},"stepfun/step-5-preview":{"id":"stepfun/step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":1000000},"cost":{"input":1,"output":2.7,"cache_read":0.05},"canonical_model_id":"stepfun/step-5-preview"},"stepfun/step-3.7-flash:thinking":{"id":"stepfun/step-3.7-flash:thinking","name":"Step 3.7 Flash Thinking","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"reflection/beam-1.0":{"id":"reflection/beam-1.0","name":"Beam 1.0 501B A23B","description":"Reflection's Apache 2.0 open-weight text model with 501.3B total and 23.4B active parameters, built for coding, reasoning and tool use. It supports a shared 1,048,576-token context window, up to 131,072 output tokens, JSON mode and structured output. Reasoning is always enabled with low, medium, high, xhigh and max effort; NanoGPT defaults to max. Knowledge cutoff: June 30, 2026.","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"elorian/glimpse":{"id":"elorian/glimpse","name":"Elorian Glimpse","description":"Glimpse is Elorian's visual reasoning model for understanding images and video and answering in text. It can analyze charts, count objects, follow events over time, and reason about spatial relationships and physical cause and effect. It supports tool calling and structured output across a shared 262,144-token context window, with up to 256,000 output tokens within the remaining context.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":256000},"cost":{"input":0.63,"output":3.15,"cache_read":0.063}},"GalrionSoftworks/MN-LooseCannon-12B-v1":{"id":"GalrionSoftworks/MN-LooseCannon-12B-v1","name":"MN-LooseCannon-12B-v1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"fireworks/ember-1":{"id":"fireworks/ember-1","name":"Ember 1","description":"Specialized model from Fireworks built on Kimi K3, producing shorter reasoning traces with approximately 40% fewer tokens while maintaining comparable quality","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"fireworks/ember-1"},"TheDrummer/Cydonia-24B-v2":{"id":"TheDrummer/Cydonia-24B-v2","name":"The Drummer Cydonia 24B v2","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-02-17","last_updated":"2025-02-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":0.1003,"output":0.1207,"cache_read":0.05015}},"TheDrummer/Magidonia-24B-v4.3":{"id":"TheDrummer/Magidonia-24B-v4.3","name":"The Drummer Magidonia 24B v4.3","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-12-25","last_updated":"2025-12-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":0.1003,"output":0.1207,"cache_read":0.05015}},"TheDrummer/UnslopNemo-12B-v4.1":{"id":"TheDrummer/UnslopNemo-12B-v4.1","name":"UnslopNemo 12b v4","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"input":8192,"output":26214},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"TheDrummer/Artemis-v1.1":{"id":"TheDrummer/Artemis-v1.1","name":"TheDrummer/Artemis v1.1","description":"TheDrummer's Artemis v1.1 is a Gemma 4 31B fine-tune for creative writing, expressive dialogue, and roleplay, with optional thinking and a 262K context window.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-06","last_updated":"2026-09-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"TheDrummer/Cydonia-24B-v4.3":{"id":"TheDrummer/Cydonia-24B-v4.3","name":"The Drummer Cydonia 24B v4.3","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-12-25","last_updated":"2025-12-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.12,"output":0.15,"cache_read":0.06}},"TheDrummer/Cydonia-24B-v4.1":{"id":"TheDrummer/Cydonia-24B-v4.1","name":"The Drummer Cydonia 24B v4.1","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04-30","release_date":"2025-09-27","last_updated":"2025-09-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":117964},"cost":{"input":0.3,"output":0.5,"cache_read":0.15},"canonical_model_id":"thedrummer/cydonia-24b-v4.1"},"TheDrummer/Anubis-70B-v1":{"id":"TheDrummer/Anubis-70B-v1","name":"Anubis 70B v1","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":16384},"cost":{"input":0.31,"output":0.31,"cache_read":0.155}},"TheDrummer/skyfall-36b-v2":{"id":"TheDrummer/skyfall-36b-v2","name":"TheDrummer Skyfall 36B V2","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06-30","release_date":"2025-03-10","last_updated":"2025-03-10","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":29491},"cost":{"input":0.55,"output":0.8,"cache_read":0.25},"canonical_model_id":"thedrummer/skyfall-36b-v2"},"TheDrummer/orion-26b-a4b-v1.1":{"id":"TheDrummer/orion-26b-a4b-v1.1","name":"Orion 26B A4B v1.1","description":"TheDrummer's Gemma 4 26B A4B fine-tune for creative writing, roleplay, and expressive dialogue. Supports optional thinking with a 65,536-token context window and up to 8,192 output tokens. Text input only.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":false,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"TheDrummer/skyfall-31b-v4.2":{"id":"TheDrummer/skyfall-31b-v4.2","name":"TheDrummer Skyfall 31B v4.2","description":"TheDrummer's Skyfall 31B v4.2 is a Magistral Small finetune for creative writing, roleplay, and dialogue. This text-only deployment uses NVFP4 weights and supports a 16K-token context window.","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":4096},"cost":{"input":0.12,"output":0.15,"cache_read":0.06}},"TheDrummer/Rocinante-12B-v1.1":{"id":"TheDrummer/Rocinante-12B-v1.1","name":"Rocinante 12b","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.408,"output":0.595,"cache_read":0.204}},"TheDrummer/Anubis-70B-v1.1":{"id":"TheDrummer/Anubis-70B-v1.1","name":"Anubis 70B v1.1","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"input":32000,"output":16384},"cost":{"input":0.31,"output":0.31,"cache_read":0.155}},"TheDrummer/Cydonia-24B-v4":{"id":"TheDrummer/Cydonia-24B-v4","name":"The Drummer Cydonia 24B v4","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":0.2006,"output":0.2414,"cache_read":0.1003}},"amazon/nova-2-lite-v1":{"id":"amazon/nova-2-lite-v1","name":"Amazon Nova 2 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":65535},"cost":{"input":0.51,"output":4.25,"cache_read":0.255},"canonical_model_id":"amazon/nova-2-lite-v1"},"amazon/nova-pro-v1":{"id":"amazon/nova-pro-v1","name":"Amazon Nova Pro 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"input":300000,"output":32000},"cost":{"input":0.799,"output":3.196,"cache_read":0.3995},"canonical_model_id":"amazon/nova-pro-v1"},"amazon/nova-lite-v1":{"id":"amazon/nova-lite-v1","name":"Amazon Nova Lite 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"input":300000,"output":5120},"cost":{"input":0.0595,"output":0.238,"cache_read":0.02975},"canonical_model_id":"amazon/nova-lite-v1"},"nothingiisreal/L3.1-70B-Celeste-V0.1-BF16":{"id":"nothingiisreal/L3.1-70B-Celeste-V0.1-BF16","name":"Llama 3.1 70B Celeste v0.1","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"stepfun-ai/step-3.5-flash":{"id":"stepfun-ai/step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.1,"output":0.3,"cache_read":0.05},"canonical_model_id":"stepfun/step-3.5-flash"},"stepfun-ai/step-3.5-flash-2603":{"id":"stepfun-ai/step-3.5-flash-2603","name":"Step 3.5 Flash 2603","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.1,"output":0.3,"cache_read":0.05},"canonical_model_id":"stepfun/step-3.5-flash-2603"},"TEE/nemotron-3.5-lightning":{"id":"TEE/nemotron-3.5-lightning","name":"Nvidia Nemotron 3.5 Lightning TEE","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.08,"output":0.2,"cache_read":0.04},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"TEE/qwen3.8-27b":{"id":"TEE/qwen3.8-27b","name":"Qwen3.8 27B TEE","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":262144},"cost":{"input":0.4,"output":3,"cache_read":0.15},"canonical_model_id":"alibaba/qwen3.8-27b"},"TEE/qwen3.6-35b-a3b":{"id":"TEE/qwen3.6-35b-a3b","name":"Qwen3.6 35B A3B TEE","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":262144},"cost":{"input":0.2,"output":1.27,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"TEE/kimi-k2.6":{"id":"TEE/kimi-k2.6","name":"Kimi K2.6 TEE","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":1.5,"output":5.25,"cache_read":0.375},"canonical_model_id":"moonshotai/kimi-k2.6"},"TEE/gemma4-31b:thinking":{"id":"TEE/gemma4-31b:thinking","name":"Gemma 4 31B Thinking TEE","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.4,"output":1,"cache_read":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"TEE/llama3-3-70b":{"id":"TEE/llama3-3-70b","name":"Llama 3.3 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":1.75,"output":2.75,"cache_read":1.75},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"TEE/deepseek-v3.2":{"id":"TEE/deepseek-v3.2","name":"DeepSeek V3.2 TEE","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"input":164000,"output":65536},"cost":{"input":0.5,"output":1,"cache_read":0.25},"canonical_model_id":"deepseek/deepseek-v3.2"},"TEE/glm-5.1-thinking":{"id":"TEE/glm-5.1-thinking","name":"GLM 5.1 Thinking TEE","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"input":202752,"output":65535},"cost":{"input":0.7,"output":2.5,"cache_read":0.18},"canonical_model_id":"zhipuai/glm-5.1"},"TEE/qwen2.5-vl-72b-instruct":{"id":"TEE/qwen2.5-vl-72b-instruct","name":"Qwen2.5 VL 72B TEE","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-01-26","last_updated":"2025-01-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.7,"output":0.7,"cache_read":0.35},"canonical_model_id":"alibaba/qwen2-5-vl-72b-instruct"},"TEE/glm-5.1":{"id":"TEE/glm-5.1","name":"GLM 5.1 TEE","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"input":202752,"output":65535},"cost":{"input":0.7,"output":2.5,"cache_read":0.18},"canonical_model_id":"zhipuai/glm-5.1"},"TEE/gpt-oss-120b":{"id":"TEE/gpt-oss-120b","name":"GPT-OSS 120B TEE","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":16384},"cost":{"input":2,"output":2,"cache_read":2},"canonical_model_id":"openai/gpt-oss-120b"},"TEE/muse-glimmer-30b":{"id":"TEE/muse-glimmer-30b","name":"Muse Glimmer 30B TEE","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":131072},"cost":{"input":0.35,"output":1.5,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"TEE/kimi-k2.7-code":{"id":"TEE/kimi-k2.7-code","name":"Kimi K2.7 Code TEE","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"TEE/qwen3.5-27b":{"id":"TEE/qwen3.5-27b","name":"Qwen3.5 27B TEE","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.3,"output":2.4,"cache_read":0.15},"canonical_model_id":"alibaba/qwen3.5-27b"},"TEE/kimi-k3":{"id":"TEE/kimi-k3","name":"Kimi K3 TEE","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":65535},"cost":{"input":3,"output":15,"cache_read":1.5},"canonical_model_id":"moonshotai/kimi-k3"},"TEE/qwen3.5-397b-a17b":{"id":"TEE/qwen3.5-397b-a17b","name":"Qwen3.5 397B A17B TEE","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.55,"output":3.5,"cache_read":0.275},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"TEE/glm-5.3-flash":{"id":"TEE/glm-5.3-flash","name":"GLM 5.3 Flash TEE","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"TEE/gemma4-31b":{"id":"TEE/gemma4-31b","name":"Gemma 4 31B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.4,"output":1,"cache_read":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"TEE/glm-5.2":{"id":"TEE/glm-5.2","name":"GLM 5.2 TEE","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.7},"canonical_model_id":"zhipuai/glm-5.2"},"TEE/gemma-4-26b-a4b-uncensored":{"id":"TEE/gemma-4-26b-a4b-uncensored","name":"Gemma 4 26B A4B Uncensored TEE","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2026-05-23","last_updated":"2026-05-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":65536},"cost":{"input":0.15,"output":0.7,"cache_read":0.075}},"TEE/deepseek-v4.1-flash":{"id":"TEE/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash TEE","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":384000},"cost":{"input":0.65,"output":1.45,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"TEE/glm-5.2:thinking":{"id":"TEE/glm-5.2:thinking","name":"GLM 5.2 Thinking TEE","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.7},"canonical_model_id":"zhipuai/glm-5.2"},"TEE/gemma-4-31b-it":{"id":"TEE/gemma-4-31b-it","name":"Gemma 4 31B IT TEE","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":262144},"cost":{"input":0.15,"output":0.46,"cache_read":0.075},"canonical_model_id":"google/gemma-4-31b-it"},"TEE/glm-5.3":{"id":"TEE/glm-5.3","name":"GLM 5.3 TEE","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"TEE/qwen3.6-27b":{"id":"TEE/qwen3.6-27b","name":"Qwen3.6 27B TEE","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.32,"output":2.7,"cache_read":0.16},"canonical_model_id":"alibaba/qwen3.6-27b"},"ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0":{"id":"ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0","name":"Omega Directive 24B Unslop v2.0","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":32768},"cost":{"input":0.5,"output":0.5,"cache_read":0.25}},"perceptron/perceptron-mk1":{"id":"perceptron/perceptron-mk1","name":"Perceptron Mk1","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.15,"output":1.5,"cache_read":0.075},"canonical_model_id":"perceptron/perceptron-mk1"},"perceptron/perceptron-mk1.5":{"id":"perceptron/perceptron-mk1.5","name":"Perceptron Mk1.5","description":"Perceptron Mk1.5 is a multimodal reasoning model for visual analysis and physical agents. It accepts text, images, video, and audio, supports tool calling and structured output, and can return spatial annotations such as points, boxes, polygons, and tracks.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":36864,"input":36864,"output":8192},"cost":{"input":0.15,"output":1.5}},"deepcogito/cogito-v1-preview-qwen-32B":{"id":"deepcogito/cogito-v1-preview-qwen-32B","name":"Cogito v1 Preview Qwen 32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-05-10","last_updated":"2025-05-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":32768},"cost":{"input":1.8,"output":1.8,"cache_read":0.9}},"heabsy/cyberheabsy":{"id":"heabsy/cyberheabsy","name":"CyberHeabsy","description":"A low-refusal text model for authorized security research, roleplay, and creative writing. It preserves tool calling after refusal removal, supports optional reasoning, and has a 1M-token context window.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-09-27","last_updated":"2026-09-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":1.5,"output":4.5,"cache_read":0.275}},"heabsy/qwen3.8-27b-uncensored-fast":{"id":"heabsy/qwen3.8-27b-uncensored-fast","name":"Qwen 3.8 27B Uncensored Fast by Heabsy","description":"Heabsy's NVFP4 build of Qwen3.8-27B by Alibaba Cloud, with refusal removal based on huihui-ai's Huihui-Qwen3.8-27B-abliterated. Quantized and calibrated by Heabsy, with unpublished weights under the Heabsy Model License. Supports text and image input, tools, and optional reasoning capped at 16,384 tokens within the output budget. This named build is separate from Qwen 3.8 27B Uncensored.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"input":524288,"output":235929},"cost":{"input":1,"output":3,"cache_read":0.22}},"minimax/minimax-latest":{"id":"minimax/minimax-latest","name":"MiniMax Latest","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"release_date":"2026-05-03","last_updated":"2026-05-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"input":512000,"output":80000},"cost":{"input":0.23,"output":0.96,"cache_read":0.05}},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"input":204800,"output":131072},"cost":{"input":0.21,"output":0.84,"cache_read":0.042},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m3:thinking":{"id":"minimax/minimax-m3:thinking","name":"MiniMax M3 Thinking","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"input":512000,"output":80000},"cost":{"input":0.23,"output":0.96,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":131072},"cost":{"input":0.33,"output":1.32,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-01":{"id":"minimax/minimax-01","name":"MiniMax 01","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-01-15","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000192,"input":1000192,"output":16384},"cost":{"input":0.1394,"output":1.122,"cache_read":0.0697}},"minimax/minimax-m2-her":{"id":"minimax/minimax-m2-her","name":"MiniMax M2-her","description":"MiniMax M2 variant tuned for conversational and character-driven agent interactions","family":"minimax","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-01-23","last_updated":"2026-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65532,"input":65532,"output":2048},"cost":{"input":0.302,"output":1.207,"cache_read":0.151},"canonical_model_id":"minimax/MiniMax-M2-Her"},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"input":200000,"output":131072},"cost":{"input":0.17,"output":1.53,"cache_read":0.085},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"input":512000,"output":80000},"cost":{"input":0.23,"output":0.96,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.7-turbo":{"id":"minimax/minimax-m2.7-turbo","name":"MiniMax M2.7 Turbo","description":"Efficient MiniMax model for quick assistance, coding, and routine automation","family":"minimax-m2.7","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"input":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.06}},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"input":204800,"output":131072},"cost":{"input":0.27,"output":1.08,"cache_read":0.027},"canonical_model_id":"minimax/MiniMax-M2.5"},"minimax/minimax-m3.1-flash":{"id":"minimax/minimax-m3.1-flash","name":"MiniMax M3.1 Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":524288},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"aisingapore/Gemma-SEA-LION-v4-27B-IT":{"id":"aisingapore/Gemma-SEA-LION-v4-27B-IT","name":"Gemma SEA-LION v4 27B","description":"Gemma 3 27B tuned by AI Singapore for Southeast Asian languages and instruction following","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":128000},"cost":{"input":0.24,"output":0.48},"canonical_model_id":"aisingapore/gemma-sea-lion-v4-27b-it"},"huihui-ai/Qwen2.5-32B-Instruct-abliterated":{"id":"huihui-ai/Qwen2.5-32B-Instruct-abliterated","name":"Qwen 2.5 32B Abliterated","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-01-06","last_updated":"2025-01-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.7,"output":0.7,"cache_read":0.35}},"huihui-ai/Llama-3.3-70B-Instruct-abliterated":{"id":"huihui-ai/Llama-3.3-70B-Instruct-abliterated","name":"Llama 3.3 70B Instruct abliterated","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-08-08","last_updated":"2025-08-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":0.7,"output":0.7,"cache_read":0.35}},"huihui-ai/DeepSeek-R1-Distill-Qwen-32B-abliterated":{"id":"huihui-ai/DeepSeek-R1-Distill-Qwen-32B-abliterated","name":"DeepSeek R1 Qwen Abliterated","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-01-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":1.4,"output":1.4,"cache_read":0.7}},"huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated":{"id":"huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated","name":"DeepSeek R1 Llama 70B Abliterated","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-01-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.7,"output":0.7,"cache_read":0.35}},"soob3123/Veiled-Calla-12B":{"id":"soob3123/Veiled-Calla-12B","name":"Veiled Calla 12B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-04-13","last_updated":"2025-04-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.3,"output":0.3,"cache_read":0.15}},"soob3123/amoral-gemma3-27B-v2":{"id":"soob3123/amoral-gemma3-27B-v2","name":"Amoral Gemma3 27B v2","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-05-23","last_updated":"2025-05-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.3,"output":0.3,"cache_read":0.15}},"upstage/solar-mini4":{"id":"upstage/solar-mini4","name":"Solar Mini 4","description":"Upstage's compact, cost-efficient model for agentic use","family":"solar-mini","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"input":524288,"output":131072},"cost":{"input":0.05,"output":0.2,"cache_read":0.005},"canonical_model_id":"upstage/solar-mini4"},"upstage/solar-pro-3":{"id":"upstage/solar-pro-3","name":"Solar Pro 3","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-07","release_date":"2026-01-26","last_updated":"2026-03-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"input":131072,"output":117964},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"upstage/solar-pro3"},"upstage/solar-mini4:thinking":{"id":"upstage/solar-mini4:thinking","name":"Solar Mini 4 Thinking","description":"Upstage's compact, cost-efficient model for agentic use","family":"solar-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"input":524288,"output":131072},"cost":{"input":0.05,"output":0.2,"cache_read":0.005},"canonical_model_id":"upstage/solar-mini4"},"upstage/solar-pro4":{"id":"upstage/solar-pro4","name":"Solar Pro 4","description":"Upstage's flagship model, specialized for agentic use","family":"solar-pro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"input":524288,"output":131072},"cost":{"input":0.03,"output":0.12,"cache_read":0.006},"canonical_model_id":"upstage/solar-pro4"},"upstage/solar-pro4:thinking":{"id":"upstage/solar-pro4:thinking","name":"Solar Pro 4 Thinking","description":"Upstage's flagship model, specialized for agentic use","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"input":524288,"output":131072},"cost":{"input":0.03,"output":0.12,"cache_read":0.006},"canonical_model_id":"upstage/solar-pro4"},"lightonai/LightOnOCR-2-1B":{"id":"lightonai/LightOnOCR-2-1B","name":"LightOnOCR 2","description":"LightOnOCR 2 hosted by IONOS in Berlin, Germany. Zero data retention.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.1785,"output":0.3465}},"inference-net/schematron-v2-small":{"id":"inference-net/schematron-v2-small","name":"Schematron V2 Small","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":4096},"cost":{"input":0.05,"output":0.23,"cache_read":0.025},"canonical_model_id":"inference-net/schematron-v2-small"},"inference-net/schematron-v2-turbo":{"id":"inference-net/schematron-v2-turbo","name":"Schematron V2 Turbo","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":8192},"cost":{"input":0.03,"output":0.15,"cache_read":0.015},"canonical_model_id":"inference-net/schematron-v2-turbo"},"nbeerbower/gemma4-gutenberg-31b":{"id":"nbeerbower/gemma4-gutenberg-31b","name":"Gemma 4 Gutenberg 31B","description":"nbeerbower's Gemma 4 31B finetune for literary fiction, character interiority, and controlled narrative pacing. Its text decoder was trained with ORPO on 14,816 Athanorlite-DPO preference pairs. This text-only deployment supports a 65,536-token context window and up to 8,192 output tokens.","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"microsoft/wizardlm-2-8x22b":{"id":"microsoft/wizardlm-2-8x22b","name":"WizardLM-2 8x22B","description":"Microsoft WizardLM instruction-tuned Mixtral MoE for multilingual dialogue and complex tasks","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-04-15","last_updated":"2024-04-15","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465},"canonical_model_id":"microsoft/wizardlm-2-8x22b"},"EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.1":{"id":"EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.1","name":"EVA-LLaMA-3.33-70B-v0.1","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":2.006,"output":2.006,"cache_read":1.003}},"EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.0":{"id":"EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.0","name":"EVA Llama 3.33 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":16384},"cost":{"input":2.006,"output":2.006,"cache_read":1.003}},"EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2":{"id":"EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2","name":"EVA-Qwen2.5-72B-v0.2","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.799,"output":0.799,"cache_read":0.3995}},"EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2":{"id":"EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2","name":"EVA-Qwen2.5-32B-v0.2","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.799,"output":0.799,"cache_read":0.3995}},"nvidia/nemotron-3.5-lightning":{"id":"nvidia/nemotron-3.5-lightning","name":"Nvidia Nemotron 3.5 Lightning","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":0.05,"output":0.2,"cache_read":0.01},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/Llama-3.1-Nemotron-70B-Instruct-HF":{"id":"nvidia/Llama-3.1-Nemotron-70B-Instruct-HF","name":"Nvidia Nemotron 70b","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"input":16384,"output":8192},"cost":{"input":0.357,"output":0.408,"cache_read":0.1785},"canonical_model_id":"nvidia/llama-3.1-nemotron-70b-instruct"},"nvidia/nemotron-3-super-120b-a12b":{"id":"nvidia/nemotron-3-super-120b-a12b","name":"Nvidia Nemotron 3 Super 120B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.05,"output":0.25,"cache_read":0.025},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/nemotron-3-ultra-550b-a55b":{"id":"nvidia/nemotron-3-ultra-550b-a55b","name":"Nvidia Nemotron 3 Ultra 550B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":0.5,"output":2.5,"cache_read":0.25},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/nemotron-3-nano-30b-a3b":{"id":"nvidia/nemotron-3-nano-30b-a3b","name":"Nvidia Nemotron 3 Nano 30B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":235929},"cost":{"input":0.17,"output":0.68,"cache_read":0.085},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/nemotron-3.5-super-vl":{"id":"nvidia/nemotron-3.5-super-vl","name":"Nvidia Nemotron 3.5 Super VL","description":"NVIDIA’s open-weight 120B model (12B active) for reasoning, coding, tool use, document and GUI understanding, and video understanding. Text, images and video in; text out. Served in NVFP4 with a 262,144-token shared context window. NanoGPT currently limits responses to 16,384 tokens while separate hosted output limits remain unpublished. OpenMDW 1.1 license.","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.15,"output":0.5,"cache_read":0.02},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/Llama-3.3-Nemotron-Super-49B-v1":{"id":"nvidia/Llama-3.3-Nemotron-Super-49B-v1","name":"Nvidia Nemotron Super 49B","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-18","last_updated":"2025-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"cost":{"input":0.15,"output":0.15,"cache_read":0.075},"canonical_model_id":"nvidia/llama-3.3-nemotron-super-49b-v1"},"nvidia/nemotron-3-super-120b-a12b:thinking":{"id":"nvidia/nemotron-3-super-120b-a12b:thinking","name":"Nvidia Nemotron 3 Super 120B Thinking","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.05,"output":0.25,"cache_read":0.025},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/nemotron-3.5-content-safety":{"id":"nvidia/nemotron-3.5-content-safety","name":"Nemotron 3.5 Content Safety","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.05,"output":0.15,"cache_read":0.025},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"nvidia/nemotron-3-ultra-550b-a55b:thinking":{"id":"nvidia/nemotron-3-ultra-550b-a55b:thinking","name":"Nvidia Nemotron 3 Ultra 550B Thinking","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":0.5,"output":2.5,"cache_read":0.25},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/nemotron-3.5-lightning:thinking":{"id":"nvidia/nemotron-3.5-lightning:thinking","name":"Nvidia Nemotron 3.5 Lightning Thinking","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":0.05,"output":0.2,"cache_read":0.01},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"deepseek-ai/DeepSeek-V3.1:thinking":{"id":"deepseek-ai/DeepSeek-V3.1:thinking","name":"DeepSeek V3.1 Thinking","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.2,"output":0.7,"cache_read":0.1},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek-ai/deepseek-v3.2-exp":{"id":"deepseek-ai/deepseek-v3.2-exp","name":"DeepSeek V3.2 Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"input":163840,"output":65536},"cost":{"input":0.28,"output":0.42,"cache_read":0.14}},"deepseek-ai/DeepSeek-V3.1-Terminus":{"id":"deepseek-ai/DeepSeek-V3.1-Terminus","name":"DeepSeek V3.1 Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2025-08-02","last_updated":"2025-08-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.25,"output":0.7,"cache_read":0.125}},"deepseek-ai/DeepSeek-R1-0528":{"id":"deepseek-ai/DeepSeek-R1-0528","name":"DeepSeek R1 0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"input":163840,"output":32768},"cost":{"input":0.5,"output":2.15,"cache_read":0.35},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek-ai/deepseek-v3.2-exp-thinking":{"id":"deepseek-ai/deepseek-v3.2-exp-thinking","name":"DeepSeek V3.2 Exp Thinking","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2024-01-01","last_updated":"2025-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"input":163840,"output":65536},"cost":{"input":0.28,"output":0.42,"cache_read":0.14}},"deepseek-ai/DeepSeek-V3.1-Terminus:thinking":{"id":"deepseek-ai/DeepSeek-V3.1-Terminus:thinking","name":"DeepSeek V3.1 Terminus (Thinking)","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek-thinking","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"release_date":"2024-01-01","last_updated":"2025-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.25,"output":0.7,"cache_read":0.125}},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.2,"output":0.7,"cache_read":0.1},"canonical_model_id":"deepseek/deepseek-v3.1"},"bytedance/doubao-seed-character":{"id":"bytedance/doubao-seed-character","name":"Doubao Seed Character","description":"ByteDance's character-focused Doubao Seed model for roleplay, persona consistency, dialogue, and creative character interactions. It supports text and image input with a 128k context window. Requests route through ZenMux to ByteDance; ZenMux does not publish a model-API zero-retention or training guarantee, so avoid sensitive data.","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":32768},"cost":{"input":0.15,"output":0.37,"cache_read":0.03,"cache_write":0.003},"canonical_model_id":"bytedance-seed/seed-character"},"bytedance/doubao-seed-2.1-turbo":{"id":"bytedance/doubao-seed-2.1-turbo","name":"Doubao Seed 2.1 Turbo","description":"Fast, lower-cost model in the Doubao Seed 2.1 family for everyday chat, coding assistance, document work, and high-throughput productivity tasks. Supports a 256k context window and up to 128k output tokens. Note: privacy and logging guarantees may be limited.","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":128000},"cost":{"input":0.5,"output":2.5,"cache_read":0.25},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"bytedance/doubao-seed-2.1-pro":{"id":"bytedance/doubao-seed-2.1-pro","name":"Doubao Seed 2.1 Pro","description":"Higher-capability model in the Doubao Seed 2.1 family for agentic coding, long-context analysis, complex instruction following, and productivity workflows. Supports a 256k context window and up to 128k output tokens. Note: privacy and logging guarantees may be limited.","family":"seed","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":128000},"cost":{"input":1,"output":5,"cache_read":0.5},"canonical_model_id":"bytedance-seed/seed-2.1-pro"},"ateron/gemma-4-dark-thoughts-v2-31b":{"id":"ateron/gemma-4-dark-thoughts-v2-31b","name":"Gemma 4 Dark Thoughts V2 31B","description":"Ateron's second Dark Thoughts merge combines Dark Scarlett V2, MeroMero V2, and Scotoma 2 for creative dialogue and extended character roleplay. This text-only Gemma 4 31B deployment supports a 65,536-token context window and up to 8,192 output tokens.","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":false,"structured_output":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"input":65536,"output":8192},"cost":{"input":0.1,"output":0.45,"cache_read":0.05}},"qwen/qwen3.8-27b-uncensored:thinking":{"id":"qwen/qwen3.8-27b-uncensored:thinking","name":"Qwen 3.8 27B Uncensored Thinking","description":"Qwen 3.8 27B Uncensored with thinking enabled for more deliberate creative work, coding, multimodal analysis, tool use, and long-context problem solving.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-07-29","last_updated":"2026-08-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":32768},"cost":{"input":1,"output":3,"cache_read":0.22}},"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":260096,"input":260096,"output":65536},"cost":{"input":0.14,"output":1,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3-30b-a3b":{"id":"qwen/qwen3-30b-a3b","name":"Qwen3 30B A3B","description":"Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":41000,"input":41000,"output":32768},"cost":{"input":0.1,"output":0.3,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":1.475,"output":4.425,"cache_read":0.295,"cache_write":1.84375},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":65536},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3.8-27b-uncensored":{"id":"qwen/qwen3.8-27b-uncensored","name":"Qwen 3.8 27B Uncensored","description":"Qwen 3.8 27B Uncensored is an NVFP4 open-weight multimodal model LoRA-tuned for fewer refusals across chat, coding, reasoning, tool use, and long-context work.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-07-29","last_updated":"2026-08-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":32768},"cost":{"input":1,"output":3,"cache_read":0.22}},"qwen/qwen3-max":{"id":"qwen/qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":1.2002,"output":6.001,"cache_read":0.6001},"canonical_model_id":"alibaba/qwen3-max"},"qwen/qwen3.5-397b-a17b:thinking":{"id":"qwen/qwen3.5-397b-a17b:thinking","name":"Qwen3.5 397B A17B Thinking","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.39,"output":2.34,"cache_read":0.3},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3.7-max:thinking":{"id":"qwen/qwen3.7-max:thinking","name":"Qwen3.7 Max Thinking","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":65536},"cost":{"input":1.475,"output":4.425,"cache_read":0.295,"cache_write":1.84375},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3-14b":{"id":"qwen/qwen3-14b","name":"Qwen 3 14b","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":41000,"input":41000,"output":32768},"cost":{"input":0.08,"output":0.24,"cache_read":0.04}},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1001536,"input":1001536,"output":32768},"cost":{"input":0.15,"output":0.7,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen-max":{"id":"qwen/qwen-max","name":"Qwen 2.5 Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"input":32000,"output":8192},"cost":{"input":1.5997,"output":6.392,"cache_read":0.79985},"canonical_model_id":"alibaba/qwen-max"},"qwen/qwen2.5-coder-32b-instruct":{"id":"qwen/qwen2.5-coder-32b-instruct","name":"Qwen 2.5 Coder 32b","description":"Open coding-focused Qwen model for code generation, repair, and repository reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"input":32000,"output":8192},"cost":{"input":0.2006,"output":0.2006,"cache_read":0.1003},"canonical_model_id":"alibaba/qwen2.5-coder-32b-instruct"},"qwen/qwen3.8-max-prime":{"id":"qwen/qwen3.8-max-prime","name":"Qwen3.8 Max Prime","description":"High-throughput edition of Qwen3.8 Max for coding, professional work, multimodal understanding, and long-running agent workflows","family":"qwen3.8-max","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":131072},"cost":{"input":4,"output":12,"cache_read":0.5},"canonical_model_id":"alibaba/qwen3.8-max-prime"},"qwen/qwen3.8-27b-hemmingway":{"id":"qwen/qwen3.8-27b-hemmingway","name":"Qwen 3.8 27B Hemingway","description":"Qwen 3.8 27B Hemingway is an open-weight NVFP4 multimodal creative finetune for long-form prose, character dialogue, storytelling, and roleplay.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.25,"output":1.5,"cache_read":0.125}},"qwen/qwen3.5-flash:thinking":{"id":"qwen/qwen3.5-flash:thinking","name":"Qwen3.5 Flash Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.5-flash"},"qwen/qwen-long":{"id":"qwen/qwen-long","name":"Qwen Long 10M","description":"Alibaba's huge context window model. Takes in up to 10 million tokens, which is equivalent to dozens of books.","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2024-08-01","last_updated":"2024-08-01","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":10000000,"input":10000000,"output":8192},"cost":{"input":0.1003,"output":0.408,"cache_read":0.05015}},"qwen/qwen3.8-27b-kimiko":{"id":"qwen/qwen3.8-27b-kimiko","name":"Qwen 3.8 27B Kimiko","description":"Qwen 3.8 27B Kimiko is a creative finetune served by Aoru in NVFP4 with a 262,144-token context window. It supports image understanding, tool calling, and optional reasoning.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3.6-35b-a3b":{"id":"qwen/qwen3.6-35b-a3b","name":"Qwen3.6 35B A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.112,"output":0.8,"cache_read":0.056},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.26,"output":2.08,"cache_read":0.103788},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3-coder-next":{"id":"qwen/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.250251,"output":0.80293,"cache_read":0.079764},"canonical_model_id":"alibaba/qwen3-coder-next"},"qwen/qwen3.5-flash":{"id":"qwen/qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.5-flash"},"qwen/qwen3-30b-a3b-instruct-2507":{"id":"qwen/qwen3-30b-a3b-instruct-2507","name":"Qwen3 30B A3B Instruct 2507","description":"Updated compact Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":0.2,"output":0.5,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3-30b-a3b-instruct-2507"},"qwen/qwen3.8-max-0902":{"id":"qwen/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"qwen/qwen-2.5-72b-instruct":{"id":"qwen/qwen-2.5-72b-instruct","name":"Qwen2.5 72B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2025-07-03","last_updated":"2025-07-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":8192},"cost":{"input":0.36,"output":0.4,"cache_read":0.1785}},"qwen/qwen3.8-27b-queen":{"id":"qwen/qwen3.8-27b-queen","name":"Qwen 3.8 27B Queen","description":"Qwen 3.8 27B Queen is an open-weight roleplay finetune with image understanding, tool calling, optional reasoning, and a 262,144-token context window.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-09","last_updated":"2026-09-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"input":524288,"output":32768},"cost":{"input":0.25,"output":1.5,"cache_read":0.125}},"qwen/qwen3.8-27b-antislop":{"id":"qwen/qwen3.8-27b-antislop","name":"Qwen 3.8 27B Antislop","description":"Qwen 3.8 27B Antislop is a creative finetune served by Aoru in NVFP4 with a 262,144-token context window. It supports image understanding, tool calling, and optional reasoning.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3.6-max-preview":{"id":"qwen/qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":245760,"input":245760,"output":65536},"cost":{"input":1.04,"output":6.24,"cache_read":0.52},"canonical_model_id":"alibaba/qwen3.6-max-preview"},"qwen/qwen3.8-27b-synthia":{"id":"qwen/qwen3.8-27b-synthia","name":"Qwen 3.8 27B Synthia","description":"Qwen 3.8 27B Synthia is a text-only roleplay finetune served by Aoru in NVFP4 with a 262,144-token context window.","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3-235b-a22b-instruct-2507":{"id":"qwen/qwen3-235b-a22b-instruct-2507","name":"Qwen 3 235b A22B 2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.0875,"output":0.35,"cache_read":0.0175},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"qwen/qwen3-vl-235b-a22b-thinking":{"id":"qwen/qwen3-vl-235b-a22b-thinking","name":"Qwen3 VL 235B A22B Thinking","description":"Qwen vision-language thinking model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-thinking"},"qwen/qwen3.7-plus:thinking":{"id":"qwen/qwen3.7-plus:thinking","name":"Qwen3.7 Plus Thinking","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"input":983616,"output":65536},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3.7-flash":{"id":"qwen/qwen3.7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":65536},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.038},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen/qwen3-235b-a22b":{"id":"qwen/qwen3-235b-a22b","name":"Qwen 3 235b A22B","description":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.0875,"output":0.35,"cache_read":0.0175},"canonical_model_id":"alibaba/qwen3-235b-a22b"},"qwen/qwen3.8-max":{"id":"qwen/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"input":991000,"output":65536},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen/qwen3.8-27b-fable":{"id":"qwen/qwen3.8-27b-fable","name":"Qwen 3.8 27B Fable","description":"Qwen 3.8 27B Fable is an open-weight multimodal creative finetune for expressive dialogue, long-form storytelling, character work, and roleplay.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-07-29","last_updated":"2026-08-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"input":524288,"output":32768},"cost":{"input":0.25,"output":1.5,"cache_read":0.125}},"qwen/qwen3.8-27b-obliterated:thinking":{"id":"qwen/qwen3.8-27b-obliterated:thinking","name":"Qwen 3.8 27B Obliterated Thinking","description":"Qwen 3.8 27B Obliterated with thinking enabled for more deliberate creative work, coding, multimodal analysis, tool use, and long-context problem solving.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-07-29","last_updated":"2026-08-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"input":524288,"output":32768},"cost":{"input":0.25,"output":1.75,"cache_read":0.225}},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":260096,"input":260096,"output":65536},"cost":{"input":0.25,"output":2,"cache_read":0.135},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":65536},"cost":{"input":0.19,"output":1.16,"cache_read":0.02,"cache_write":0.24},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3.5-omni-plus":{"id":"qwen/qwen3.5-omni-plus","name":"Qwen3.5 Omni Plus","description":"Qwen3.5 Omni Plus is Qwen's stronger general multimodal model. We verified live support for text prompts, images, audio files, and direct video URLs on Alibaba's chat-completions-compatible API. Alibaba describes Plus as a comprehensive evolution of Qwen3 Omni with support for over 10 hours of audio input.","family":"qwen3.5","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-03-30","last_updated":"2026-03-30","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"input":983616,"output":65536}},"qwen/qwen3.5-35b-a3b:thinking":{"id":"qwen/qwen3.5-35b-a3b:thinking","name":"Qwen3.5 35B A3B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":260096,"input":260096,"output":65536},"cost":{"input":0.14,"output":1,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.39,"output":2.34,"cache_read":0.3},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3-coder-30b-a3b-instruct":{"id":"qwen/qwen3-coder-30b-a3b-instruct","name":"Qwen3 Coder 30B A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"qwen/qwen3.6-27b:thinking":{"id":"qwen/qwen3.6-27b:thinking","name":"Qwen3.6 27B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.203,"output":2.24,"cache_read":0.1015},"canonical_model_id":"alibaba/qwen3.6-27b"},"qwen/qwen3-next-80b-a3b-thinking":{"id":"qwen/qwen3-next-80b-a3b-thinking","name":"Qwen3 Next 80B A3B (Thinking)","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":0.15,"output":0.65,"cache_read":0.075},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"qwen/qwen3.8-omni-flash":{"id":"qwen/qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"qwen/qwen-turbo":{"id":"qwen/qwen-turbo","name":"Qwen Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-11-01","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":8192},"cost":{"input":0.04998,"output":0.2006,"cache_read":0.02499},"canonical_model_id":"alibaba/qwen-turbo"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":131072},"cost":{"input":0.14,"output":0.42,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3.8-27b-fimi":{"id":"qwen/qwen3.8-27b-fimi","name":"Qwen 3.8 27B Fimi","description":"Qwen 3.8 27B Fimi is a creative finetune served by Aoru in NVFP4 with a 262,144-token context window. It supports image understanding, tool calling, and optional reasoning.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3.8-27b-novelist":{"id":"qwen/qwen3.8-27b-novelist","name":"Qwen 3.8 27B Novelist","description":"Qwen 3.8 27B Novelist is a creative finetune served by Aoru in NVFP4 with a 262,144-token context window. It supports image understanding, tool calling, and optional reasoning.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3.5-27b:thinking":{"id":"qwen/qwen3.5-27b:thinking","name":"Qwen3.5 27B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":260096,"input":260096,"output":65536},"cost":{"input":0.25,"output":2,"cache_read":0.135},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3-next-80b-a3b-instruct":{"id":"qwen/qwen3-next-80b-a3b-instruct","name":"Qwen3 Next 80B A3B (Instruct)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":1.1,"cache_read":0.07},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"qwen/qwen3.8-27b:thinking":{"id":"qwen/qwen3.8-27b:thinking","name":"Qwen3.8 27B Thinking","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1001536,"input":1001536,"output":32768},"cost":{"input":0.15,"output":0.7,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3-32b":{"id":"qwen/qwen3-32b","name":"Qwen 3 32b","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":41000,"input":41000,"output":32768},"cost":{"input":0.1,"output":0.3,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3-32b"},"qwen/qwen3.7-flash:thinking":{"id":"qwen/qwen3.7-flash:thinking","name":"Qwen3.7 Flash Thinking","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"input":983616,"output":65536},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.038},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen/qwen3-8b":{"id":"qwen/qwen3-8b","name":"Qwen 3 8B","description":"Dense 8B Qwen3 model with switchable thinking for chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":41000,"input":41000,"output":32768},"cost":{"input":0.47,"output":0.47,"cache_read":0.235},"canonical_model_id":"alibaba/qwen3-8b"},"qwen/qwen3.8-27b-cybersecurity":{"id":"qwen/qwen3.8-27b-cybersecurity","name":"Qwen 3.8 27B Cybersecurity","description":"Qwen 3.8 27B Cybersecurity is a cybersecurity-focused variant based on the uncensored model, with provider moderation for illegal activities. It supports optional reasoning, image understanding, tool calling, and a 262,144-token context window.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-19","last_updated":"2026-09-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3-235b-a22b-thinking-2507":{"id":"qwen/qwen3-235b-a22b-thinking-2507","name":"Qwen 3 235b A22B 2507 Thinking","description":"Updated large open Qwen3 MoE thinking model for reasoning, math, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":81920},"cost":{"input":0.3,"output":0.5},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"qwen/qwen3.8-max:thinking":{"id":"qwen/qwen3.8-max:thinking","name":"Qwen3.8 Max Thinking","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"input":991000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen/qwen3.5-9b":{"id":"qwen/qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.05,"output":0.15,"cache_read":0.025},"canonical_model_id":"alibaba/qwen3.5-9b"},"qwen/qwen3.5-plus":{"id":"qwen/qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"input":983616,"output":65536},"cost":{"input":0.4,"output":2.4,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen 3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991808,"input":991808,"output":65536},"cost":{"input":0.325,"output":1.95,"cache_read":0.0325,"cache_write":0.40625},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen/qwen3.8-27b-darkidol":{"id":"qwen/qwen3.8-27b-darkidol","name":"Qwen 3.8 27B DarkIdol","description":"Qwen 3.8 27B DarkIdol is a creative finetune served by Aoru in NVFP4 with a 262,144-token context window. It supports image understanding, tool calling, and optional reasoning.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3-vl-235b-a22b-instruct":{"id":"qwen/qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":131072,"output":32768},"cost":{"input":0.3,"output":1.9,"cache_read":0.15},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"qwen/qwen3.8-27b-kiwen":{"id":"qwen/qwen3.8-27b-kiwen","name":"Qwen 3.8 27B Kiwen","description":"Qwen 3.8 27B Kiwen is a creative finetune served by Aoru in NVFP4 with a 262,144-token context window. It supports image understanding, tool calling, and optional reasoning.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3-max-2026-01-23":{"id":"qwen/qwen3-max-2026-01-23","name":"Qwen3 Max 2026-01-23","description":"Qwen3 Max is Alibaba's flagship Qwen 3 reasoning model with native tool use (web search, web extractor, code interpreter) and a 256K context window.","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-01-26","last_updated":"2026-01-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":32768},"cost":{"input":1.2002,"output":6.001,"cache_read":0.6001}},"qwen/qwen3.8-27b-obliterated":{"id":"qwen/qwen3.8-27b-obliterated","name":"Qwen 3.8 27B Obliterated","description":"Qwen 3.8 27B Obliterated is an open-weight multimodal model LoRA-tuned for fewer refusals across chat, coding, reasoning, tool use, and long-context work.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"release_date":"2026-07-29","last_updated":"2026-08-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"input":524288,"output":32768},"cost":{"input":0.25,"output":1.75,"cache_read":0.225}},"qwen/qwen-plus":{"id":"qwen/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":995904,"input":995904,"output":32768},"cost":{"input":0.3995,"output":1.2002,"cache_read":0.19975},"canonical_model_id":"alibaba/qwen-plus"},"qwen/qwen3.6-35b-a3b:thinking":{"id":"qwen/qwen3.6-35b-a3b:thinking","name":"Qwen3.6 35B A3B Thinking","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":16384},"cost":{"input":0.112,"output":0.8,"cache_read":0.056},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3-coder-flash":{"id":"qwen/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":65536},"cost":{"input":0.3,"output":1.5,"cache_read":0.15},"canonical_model_id":"alibaba/qwen3-coder-flash"},"qwen/qwen3.8-27b-swift":{"id":"qwen/qwen3.8-27b-swift","name":"Qwen 3.8 27B Swift","description":"Qwen 3.8 27B Swift is a finetune for fast responses, served by Aoru in NVFP4 with a 262,144-token context window. It supports image understanding, tool calling, and optional reasoning.","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-10-05","last_updated":"2026-10-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.6,"cache_read":0.05}},"qwen/qwen3.5-122b-a10b:thinking":{"id":"qwen/qwen3.5-122b-a10b:thinking","name":"Qwen3.5 122B A10B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.26,"output":2.08,"cache_read":0.103788},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3.5-omni-flash":{"id":"qwen/qwen3.5-omni-flash","name":"Qwen3.5 Omni Flash","description":"Qwen3.5 Omni Flash is Qwen's fast multimodal model. We verified live support for text prompts, images, audio files, and direct video URLs on Alibaba's chat-completions-compatible API. Alibaba describes Flash as a fully evolved version of Qwen3 Omni with audio input support across 60+ languages.","family":"qwen3.5","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2026-03-30","last_updated":"2026-03-30","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":49152,"input":49152,"output":16384}},"qwen/qwen3.5-plus:thinking":{"id":"qwen/qwen3.5-plus:thinking","name":"Qwen3.5 Plus Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"input":983616,"output":65536},"cost":{"input":0.4,"output":2.4,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen/qwen3-coder":{"id":"qwen/qwen3-coder","name":"Qwen 3 Coder 480B","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"input":262000,"output":65536},"cost":{"input":0.13,"output":0.5,"cache_read":0.065},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":65536},"cost":{"input":0.203,"output":2.24,"cache_read":0.1015},"canonical_model_id":"alibaba/qwen3.6-27b"},"cohere/command-r-plus-08-2024":{"id":"cohere/command-r-plus-08-2024","name":"Cohere: Command R+","description":"Cohere's RAG workhorse for long-context enterprise search and tool use","family":"command-r","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":4096},"cost":{"input":2.856,"output":14.246,"cache_read":1.428},"canonical_model_id":"cohere/command-r-plus-08-2024"},"abacusai/Dracarys-72B-Instruct":{"id":"abacusai/Dracarys-72B-Instruct","name":"Llama 3.1 70B Dracarys 2","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-08-02","last_updated":"2025-08-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":8192},"cost":{"input":0.493,"output":0.493,"cache_read":0.2465}},"pamanseau/OpenReasoning-Nemotron-32B":{"id":"pamanseau/OpenReasoning-Nemotron-32B","name":"OpenReasoning Nemotron 32B","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.05}},"unbiased/pareto-26.10":{"id":"unbiased/pareto-26.10","name":"Pareto 26.10","description":"Pareto 26.10 is Unbiased AI's multimodal composite model for research, coding, and agentic workflows. It accepts text and images, supports tool calling, and provides a 1M-token context window with up to 131K output tokens. This is the stable release of Pareto 26.10. Tool selection is automatic; forced tool selection, structured output, and adjustable reasoning controls are not supported.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":131072},"cost":{"input":0.8,"output":3.2,"cache_read":0.03}},"unbiased/pareto-26.9":{"id":"unbiased/pareto-26.9","name":"Pareto 26.9","description":"Pareto 26.9 is an experimental multimodal composite model from Unbiased AI for research, coding, agentic workflows, and general-purpose tasks. It runs several frontier and open models for each request and returns a single selected or synthesized answer. This experimental release supports text and image input, tool calling, a 262K-token context window, and up to 131K output tokens. Its underlying model mix and behavior may change as it is evaluated. Request content may be retained for abuse prevention and security monitoring for up to 30 days, and is not used for model training without explicit consent.","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.25}},"meganova-ai/manta-pro-1.0":{"id":"meganova-ai/manta-pro-1.0","name":"Manta Pro 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-09-20","last_updated":"2025-12-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"input":65536,"output":32768},"cost":{"input":0.06,"output":0.5,"cache_read":0.03}},"meganova-ai/manta-flash-1.0":{"id":"meganova-ai/manta-flash-1.0","name":"Manta Flash 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-09-20","last_updated":"2025-12-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"input":16384,"output":16384},"cost":{"input":0.02,"output":0.16,"cache_read":0.01}},"meganova-ai/manta-mini-1.0":{"id":"meganova-ai/manta-mini-1.0","name":"Manta Mini 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"release_date":"2025-09-20","last_updated":"2025-12-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"input":8192,"output":8192},"cost":{"input":0.02,"output":0.16,"cache_read":0.01}},"prism-ml/ternary-bonsai-2-27b":{"id":"prism-ml/ternary-bonsai-2-27b","name":"Ternary Bonsai 2 27B","description":"Ternary Bonsai 2 27B is a 27B-parameter reasoning model from PrismML derived from Qwen3.8-27B. It supports coding, mathematics, tool calling, and image understanding with a 262K-token context window.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","xhigh"]}],"tool_call":true,"structured_output":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.075,"output":0.5,"cache_read":0.0375}},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-3.1-pro-preview-high":{"id":"google/gemini-3.1-pro-preview-high","name":"Gemini 3.1 Pro (Preview High)","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":0.1875},"canonical_model_id":"google/gemini-4-argon"},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro (Preview Custom Tools)","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemma-4-26b-a4b-it-cybersecurity":{"id":"google/gemma-4-26b-a4b-it-cybersecurity","name":"Gemma 4 26B A4B Cybersecurity","description":"Gemma 4 26B A4B Cybersecurity is a cybersecurity-focused variant based on the uncensored model, with provider moderation for illegal activities. It supports optional reasoning, image understanding, tool calling, and a 262,144-token context window.","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"release_date":"2026-09-19","last_updated":"2026-09-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1056,"output":0.3344,"cache_read":0.0528}},"google/gemini-pro-latest":{"id":"google/gemini-pro-latest","name":"Gemini Pro Latest","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-29","last_updated":"2026-03-29","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":0.375}},"google/gemma-4-31b-it:thinking":{"id":"google/gemma-4-31b-it:thinking","name":"Gemma 4 31B Thinking","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.1,"output":0.34,"cache_read":0.1},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-3-flash-preview-thinking":{"id":"google/gemini-3-flash-preview-thinking","name":"Gemini 3 Flash Thinking","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-flash-latest":{"id":"google/gemini-flash-latest","name":"Gemini Flash Latest","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-flash-latest"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-3.1-pro-preview-low":{"id":"google/gemini-3.1-pro-preview-low","name":"Gemini 3.1 Pro (Preview Low)","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro (Preview)","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.0765,"output":0.255,"cache_read":0.0425},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-flash-lite-latest":{"id":"google/gemini-flash-lite-latest","name":"Gemini Flash Lite Latest","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.08333},"canonical_model_id":"google/gemini-flash-lite-latest"},"google/gemma4-31b-splituntied":{"id":"google/gemma4-31b-splituntied","name":"Gemma 4 31B Split-Untied","description":"Blazed-Forge's Split-Untied is a text-only Gemma 4 31B community finetune with an untied BF16 output head, built for creative writing, roleplay, expressive dialogue, and tool use.","family":"gemma","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.1,"output":0.3,"cache_read":0.05}},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemma-4-26b-a4b-it:thinking":{"id":"google/gemma-4-26b-a4b-it:thinking","name":"Gemma 4 26B A4B Thinking","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.0765,"output":0.255,"cache_read":0.0425},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/diffusiongemma":{"id":"google/diffusiongemma","name":"DiffusionGemma","description":"DiffusionGemma is a high-speed diffusion-based version of Gemma 4 26B A4B. It supports optional reasoning and a 262,144-token context window.","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","xhigh"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-09","last_updated":"2026-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.05,"output":0.15,"cache_read":0.025},"canonical_model_id":"google/diffusiongemma-26b-a4b-it"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash (Preview)","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-3.5-flash-thinking":{"id":"google/gemini-3.5-flash-thinking","name":"Gemini 3.5 Flash Thinking","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":131072},"cost":{"input":0.14,"output":0.4,"cache_read":0.05},"canonical_model_id":"google/gemma-4-31b-it"},"x-ai/grok-4.5":{"id":"x-ai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"input":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.3},"canonical_model_id":"xai/grok-4.5"},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"input":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"x-ai/grok-latest":{"id":"x-ai/grok-latest","name":"Grok Latest","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"release_date":"2026-05-03","last_updated":"2026-05-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"input":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5}},"x-ai/grok-4.7":{"id":"x-ai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"input":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"x-ai/grok-4.20-multi-agent":{"id":"x-ai/grok-4.20-multi-agent","name":"Grok 4.20 Multi-Agent","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"input":2000000,"output":131072},"cost":{"input":1.25,"output":2.5,"cache_read":0.2}},"x-ai/grok-4.20":{"id":"x-ai/grok-4.20","name":"Grok 4.20","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"input":2000000,"output":131072},"cost":{"input":1.25,"output":2.5,"cache_read":0.2}},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":1000000,"output":900000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"x-ai/grok-build-0.1":{"id":"x-ai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":230400},"cost":{"input":1,"output":2,"cache_read":0.2},"canonical_model_id":"xai/grok-build-0.1"},"arcee-ai/trinity-large-thinking":{"id":"arcee-ai/trinity-large-thinking","name":"Trinity Large Thinking","description":"Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use","family":"trinity","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":80000},"cost":{"input":0.25,"output":0.9,"cache_read":0.125},"canonical_model_id":"arcee-ai/trinity-large-thinking"}}},"dinference":{"id":"dinference","env":["DINFERENCE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.dinference.com/v1","name":"DInference","doc":"https://dinference.com","models":{"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"status":"deprecated","cost":{"input":1.25,"output":3.89},"canonical_model_id":"zhipuai/glm-5.1"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.0675,"output":0.27},"canonical_model_id":"openai/gpt-oss-120b"},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.45,"output":1.65},"canonical_model_id":"zhipuai/glm-4.7"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"status":"deprecated","cost":{"input":1.25,"output":3.89},"canonical_model_id":"zhipuai/glm-5.2"},"minimax-m2.5":{"id":"minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":0.22,"output":0.88},"canonical_model_id":"minimax/MiniMax-M2.5"},"glm-5":{"id":"glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.75,"output":2.4},"canonical_model_id":"zhipuai/glm-5"}}},"aihubmix":{"id":"aihubmix","env":["AIHUBMIX_API_KEY"],"npm":"@aihubmix/ai-sdk-provider","name":"AIHubMix","doc":"https://docs.aihubmix.com","models":{"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.8-flash"},"qwen3.5-35b-a3b":{"id":"qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.0564,"output":0.4512,"tiers":[{"input":0.2254,"output":1.8032,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":1,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.499,"cache_read":0.03,"input_audio":0.9999},"canonical_model_id":"google/gemini-2.5-flash"},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":131072},"cost":{"input":1.69,"output":5.07,"cache_read":0.169,"cache_write":2.1125},"canonical_model_id":"alibaba/qwen3.7-max"},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":1000000},"cost":{"input":2.2,"output":6.6,"cache_read":0.55,"tiers":[{"input":4.4,"output":13.2,"cache_read":1.1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4.4,"output":13.2,"cache_read":1.1}},"canonical_model_id":"xai/grok-4.5"},"deepseek-v4-flash-0731-fast":{"id":"deepseek-v4-flash-0731-fast","name":"DeepSeek V4 Flash 0731 Fast","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.28,"output":1.4,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":64000},"cost":{"input":0.282,"output":1.128,"cache_read":0.0564,"cache_write":0.3525,"tiers":[{"input":0.845,"output":3.38,"cache_read":0.169,"cache_write":1.05625,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.845,"output":3.38,"cache_read":0.169,"cache_write":1.05625}},"canonical_model_id":"alibaba/qwen3.7-plus"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"qwen3-max":{"id":"qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.4508,"output":2.7048,"cache_read":0.09016,"cache_write":0.5635,"tiers":[{"input":0.902,"output":5.412,"cache_read":0.1804,"cache_write":1.1275,"tier":{"type":"context","size":32000}},{"input":1.3522,"output":8.1132,"cache_read":0.27044,"cache_write":1.69025,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3-max"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"gpt-5.5-pro":{"id":"gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1,"input_audio":2},"canonical_model_id":"google/gemini-4-argon"},"mimo-v2.6-pro-ultraspeed":{"id":"mimo-v2.6-pro-ultraspeed","name":"MiMo-V2.6-Pro-UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.8,"output":9.6,"cache_read":0.0384},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"coding-minimax-m2.7-free":{"id":"coding-minimax-m2.7-free","name":"Coding MiniMax M2.7 (Free)","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128100},"cost":{"input":0,"output":0},"canonical_model_id":"minimax/MiniMax-M2.7"},"step-3.7-flash":{"id":"step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.22,"output":1.32,"cache_read":0.044},"canonical_model_id":"stepfun/step-3.7-flash"},"gemini-3.1-pro-preview-customtools":{"id":"gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"claude-opus-4-8-think":{"id":"claude-opus-4-8-think","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.155,"output":0.31,"cache_read":0.0031},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi-mimo-v2.5-pro":{"id":"xiaomi-mimo-v2.5-pro","name":"Xiaomi MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo-v2.5-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-05-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":1.1,"output":3.3,"cache_read":0.22,"tiers":[{"input":2.2,"output":6.6,"cache_read":0.44,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2.2,"output":6.6,"cache_read":0.44}},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5"},"alicloud-deepseek-v4-pro":{"id":"alicloud-deepseek-v4-pro","name":"DeepSeek V4 Pro (Alibaba Cloud)","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":1.69,"output":3.38,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v4-pro"},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"muse-spark-1.3":{"id":"muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.375,"output":4.675,"cache_read":0.165},"canonical_model_id":"meta/muse-spark-1.3"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125,"tiers":[{"input":6.6,"output":24.75,"cache_read":0.66,"cache_write":8.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6.6,"output":24.75,"cache_read":0.66,"cache_write":8.25}},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.48,"output":0.96,"cache_read":0.00384},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.254,"output":1.524},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen3.5-122b-a10b":{"id":"qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.1126,"output":0.9008,"tiers":[{"input":0.2818,"output":2.2544,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen3.8-2.4t-a95b":{"id":"qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"qwen3.5-flash":{"id":"qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.0282,"output":0.282,"cache_read":0.00282,"cache_write":0.03525,"tiers":[{"input":0.1126,"output":1.126,"cache_read":0.01126,"cache_write":0.14075,"tier":{"type":"context","size":128000}},{"input":0.169,"output":1.69,"cache_read":0.0169,"cache_write":0.21125,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.5-flash"},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"coding-minimax-m2.7-highspeed":{"id":"coding-minimax-m2.7-highspeed","name":"Coding MiniMax M2.7 Highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128100},"cost":{"input":0.2,"output":0.2},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"xiaomi-mimo-v2.5-pro-free":{"id":"xiaomi-mimo-v2.5-pro-free","name":"Xiaomi MiMo-V2.5-Pro (free)","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo-v2.5-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-05-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"qwen3.8-flash-free":{"id":"qwen3.8-flash-free","name":"Qwen3.8 Flash (free)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.8-flash"},"zai-glm-5.1":{"id":"zai-glm-5.1","name":"GLM-5.1 (Z.ai)","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"status":"deprecated","cost":{"input":0.845,"output":3.38,"cache_read":0.183112},"canonical_model_id":"zhipuai/glm-5.1"},"gemini-3.1-flash-lite-image":{"id":"gemini-3.1-flash-lite-image","type":"image","name":"Nano Banana 2 Lite","description":"Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":true,"interleaved":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":4096},"cost":{"input":0.25,"output":1.5,"cache_read":0.025},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":64000},"cost":{"input":1.08,"output":3.0888,"cache_read":0.054},"canonical_model_id":"stepfun/step-5-preview"},"qwen3-vl-plus":{"id":"qwen3-vl-plus","name":"Qwen3-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.137,"output":1.37,"cache_read":0,"tiers":[{"input":0.2054,"output":2.054,"tier":{"type":"context","size":32000}},{"input":0.411,"output":4.11,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3-vl-plus"},"xiaomi-mimo-v2.5-free":{"id":"xiaomi-mimo-v2.5-free","name":"Xiaomi MiMo-V2.5 (free)","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo-v2.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-05-13","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.48,"output":0.96,"cache_read":0.00384},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"coding-xiaomi-mimo-v2.5-pro":{"id":"coding-xiaomi-mimo-v2.5-pro","name":"Coding Xiaomi MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo-v2.5-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-05-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.2,"output":0.4,"cache_read":0.0016},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"gemini-2.5-flash-image":{"id":"gemini-2.5-flash-image","type":"image","name":"Nano Banana","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":30,"cache_read":0.03},"canonical_model_id":"google/gemini-2.5-flash-image"},"xiaomi-mimo-v2.5":{"id":"xiaomi-mimo-v2.5","name":"Xiaomi MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo-v2.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-05-13","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.44,"output":2.2,"cache_read":0.088,"tiers":[{"input":0.88,"output":4.4,"cache_read":0.176,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.88,"output":4.4,"cache_read":0.176}},"canonical_model_id":"xiaomi/mimo-v2.5"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.155,"output":0.31,"cache_read":0.0031},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"glm-4.5v":{"id":"glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":16384},"cost":{"input":0.274,"output":1.096,"tiers":[{"input":0.548,"output":1.644,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.5v"},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.6918,"output":2.0754,"cache_read":0.023057694},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"o3":{"id":"o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"qwen3.6-max-preview":{"id":"qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":240000,"output":64000},"status":"deprecated","cost":{"input":1.27,"output":7.61,"cache_read":0.1268,"cache_write":1.585,"tiers":[{"input":2.11,"output":12.67,"cache_read":0.2112,"cache_write":2.64,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.6-max-preview"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"deepseek-v4-flash-vision-exp":{"id":"deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0.155,"output":0.62,"cache_read":0.0031},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"doubao-seed-2-0-mini-260428":{"id":"doubao-seed-2-0-mini-260428","name":"Doubao Seed 2.0 Mini 260428","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.030136,"output":0.30136,"cache_read":0.0060272,"input_audio":0.423,"tiers":[{"input":0.060274,"output":0.60274,"cache_read":0.0120548,"tier":{"type":"context","size":32000}},{"input":0.120548,"output":1.20548,"cache_read":0.0241096,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":1.1,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5"},"claude-opus-4-6-think":{"id":"claude-opus-4-6-think","name":"Claude Opus 4.6 Thinking","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"gpt-image-2":{"id":"gpt-image-2","type":"image","name":"GPT-Image-2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"coding-glm-5.1-free":{"id":"coding-glm-5.1-free","name":"Coding GLM 5.1 (free)","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.1"},"kimi-k2.7-code-highspeed":{"id":"kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":1.9,"output":7.999,"cache_read":0.32167},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"qwen3.8-max-preview":{"id":"qwen3.8-max-preview","name":"Qwen3.8 Max Preview","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-07-19","last_updated":"2026-07-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0.338,"output":1.014,"cache_read":0.0676,"cache_write":0.4225},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":11,"output":55,"cache_read":0.275,"cache_write":13.75},"canonical_model_id":"anthropic/claude-fable-5-1"},"coding-xiaomi-mimo-v2.5":{"id":"coding-xiaomi-mimo-v2.5","name":"Coding Xiaomi MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo-v2.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-05-13","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.08,"output":0.16,"cache_read":0.0016},"canonical_model_id":"xiaomi/mimo-v2.5"},"qwen3.7-flash":{"id":"qwen3.7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"input":991808,"output":131072},"cost":{"input":0.0282,"output":0.1128,"cache_read":0.00564,"cache_write":0.03525,"tiers":[{"input":0.0845,"output":0.338,"cache_read":0.0169,"cache_write":0.105625,"tier":{"type":"context","size":32000}},{"input":0.169,"output":0.676,"cache_read":0.0338,"cache_write":0.21125,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"nemotron-3.5-lightning-free":{"id":"nemotron-3.5-lightning-free","name":"Nemotron 3.5 Lightning 30B A3B (free)","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":131072},"cost":{"input":1.69,"output":5.07,"cache_read":0.169,"cache_write":2.1125},"canonical_model_id":"alibaba/qwen3.8-max"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.95,"output":3.9995,"cache_read":0.160835},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"hy3-preview":{"id":"hy3-preview","name":"Hy3 Preview","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":128000},"status":"deprecated","cost":{"input":0.17,"output":0.566661,"cache_read":0.051},"canonical_model_id":"tencent/hy3-preview"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.499999,"cache_read":0.03,"input_audio":0.3},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"coding-glm-5.2-free":{"id":"coding-glm-5.2-free","name":"GLM-5.2 (Coding, free)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.2"},"ling-3.0-flash-free":{"id":"ling-3.0-flash-free","name":"Ling 3.0 Flash (free)","description":"Hybrid-linear MoE reasoning model with 124B total and 5.1B active parameters for efficient agents","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"inclusionai/ling-3.0-flash"},"qwen3.5-27b":{"id":"qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.0846,"output":0.6768,"tiers":[{"input":0.2536,"output":2.0288,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":65536},"cost":{"input":0.169,"output":1.014,"cache_read":0.0169,"cache_write":0.21125,"tiers":[{"input":0.676,"output":4.056,"cache_read":0.0676,"cache_write":0.845,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.676,"output":4.056,"cache_read":0.0676,"cache_write":0.845}},"canonical_model_id":"alibaba/qwen3.6-flash"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.1644,"output":0.9864,"tiers":[{"input":0.411,"output":2.466,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"doubao-seed-2-0-code-preview":{"id":"doubao-seed-2-0-code-preview","name":"Doubao Seed 2.0 Code Preview","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.4822,"output":2.411,"cache_read":0.09644,"tiers":[{"input":0.72328,"output":3.6164,"cache_read":0.144656,"tier":{"type":"context","size":32000}},{"input":1.4466,"output":7.233,"cache_read":0.28932,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"alicloud-glm-5.1":{"id":"alicloud-glm-5.1","name":"GLM-5.1 (Alibaba Cloud)","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"status":"deprecated","cost":{"input":0.84,"output":3.38,"cache_read":0.169,"cache_write":1.05625},"canonical_model_id":"zhipuai/glm-5.1"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"claude-opus-4-7-think":{"id":"claude-opus-4-7-think","name":"Claude Opus 4.7 Thinking","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"longcat-2.0":{"id":"longcat-2.0","name":"LongCat-2.0","description":"Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.7746,"output":3.0984,"cache_read":0.015492},"canonical_model_id":"meituan/longcat-2.0"},"claude-sonnet-4-5-think":{"id":"claude-sonnet-4-5-think","name":"Claude Sonnet 4.5 Thinking","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125,"tiers":[{"input":6.6,"output":24.75,"cache_read":0.66,"cache_write":8.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6.6,"output":24.75,"cache_read":0.66,"cache_write":8.25}},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.273974,"output":1.095896,"cache_read":0.0547948,"tiers":[{"input":0.547946,"output":2.191784,"cache_read":0.1095892,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.7"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"nemotron-3-ultra-550b-a55b-free":{"id":"nemotron-3-ultra-550b-a55b-free","name":"Nemotron 3 Ultra 550B A55B (free)","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"coding-glm-5.3-free":{"id":"coding-glm-5.3-free","name":"GLM-5.3 (Coding, free)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3"},"qwen3.8-omni-flash":{"id":"qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1126,"output":0.380025,"cache_read":0.014075,"cache_write":0.1759375},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.7-flash"},"gpt-image-2.5-sunburst":{"id":"gpt-image-2.5-sunburst","type":"image","name":"GPT Image 2.5 Sunburst","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-sunburst"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"claude-sonnet-4-6-think":{"id":"claude-sonnet-4-6-think","name":"Claude Sonnet 4.6 Thinking","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"nemotron-3.5-content-safety-free":{"id":"nemotron-3.5-content-safety-free","name":"Nemotron 3.5 Content Safety (free)","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1126,"output":0.380025,"cache_read":0.014075,"cache_write":0.1759375},"canonical_model_id":"alibaba/qwen3.8-flash"},"muse-spark-1.1":{"id":"muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.375,"output":4.675},"canonical_model_id":"meta/muse-spark-1.1"},"muse-spark-1.2":{"id":"muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.375,"output":4.675},"canonical_model_id":"meta/muse-spark-1.2"},"gpt-image-1.5":{"id":"gpt-image-1.5","type":"image","name":"GPT-Image-1.5","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-11-25","last_updated":"2025-11-25","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":32},"canonical_model_id":"openai/gpt-image-1.5"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"doubao-seed-2-0-lite-260428":{"id":"doubao-seed-2-0-lite-260428","name":"Doubao Seed 2.0 Lite 260428","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.09041,"output":0.54246,"cache_read":0.018082,"input_audio":1.35615,"tiers":[{"input":0.135616,"output":0.813696,"cache_read":0.0271232,"tier":{"type":"context","size":32000}},{"input":0.271232,"output":1.627392,"cache_read":0.0542464,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"gpt-image-2.5-flare":{"id":"gpt-image-2.5-flare","type":"image","name":"GPT Image 2.5 Flare","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt-image","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":5,"output":30,"cache_read":1.25},"canonical_model_id":"openai/gpt-image-2.5-flare"},"coding-glm-5.3-flash-free":{"id":"coding-glm-5.3-flash-free","name":"GLM-5.3-Flash (Coding, free)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.11268,"output":0.39438,"cache_read":0.02817},"canonical_model_id":"zhipuai/glm-5.3-flash"},"nemotron-3-super-120b-a12b-free":{"id":"nemotron-3-super-120b-a12b-free","name":"Nemotron 3 Super 120B A12B (free)","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":512000},"cost":{"input":0.288,"output":1.152},"canonical_model_id":"minimax/MiniMax-M3"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.1268,"output":3.9438,"cache_read":0.2817},"canonical_model_id":"zhipuai/glm-5.2"},"nemotron-3-nano-omni-30b-a3b-reasoning-free":{"id":"nemotron-3-nano-omni-30b-a3b-reasoning-free","name":"Nemotron 3 Nano Omni 30B A3B Reasoning (free)","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.5-flash"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.155,"output":0.62,"cache_read":0.0031},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"ox-alpha":{"id":"ox-alpha","name":"Ox Alpha","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"claude-opus-4-5-think":{"id":"claude-opus-4-5-think","name":"Claude Opus 4.5 Thinking","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"gpt-5.1-codex-mini":{"id":"gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"qwen3.5-plus":{"id":"qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1096,"output":0.6576,"cache_read":0.01096,"cache_write":0.137,"tiers":[{"input":0.274,"output":1.644,"cache_read":0.0274,"cache_write":0.3425,"tier":{"type":"context","size":128000}},{"input":0.548,"output":3.288,"cache_read":0.0548,"cache_write":0.685,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.5-plus"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":991000,"output":65536},"cost":{"input":0.282,"output":1.692,"cache_read":0.0282,"cache_write":0.3525,"tiers":[{"input":1.128,"output":6.768,"cache_read":0.1128,"cache_write":1.41,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.128,"output":6.768,"cache_read":0.1128,"cache_write":1.41}},"canonical_model_id":"alibaba/qwen3.6-plus"},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"gemini-3.1-flash-image":{"id":"gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"interleaved":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","video","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.142,"output":0.284,"cache_read":0.0284},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"deep-deepseek-v4-flash":{"id":"deep-deepseek-v4-flash","name":"DeepSeek V4 Flash (DeepSeek)","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0.154,"output":0.308,"cache_read":0.0308},"canonical_model_id":"deepseek/deepseek-v4-flash"},"alicloud-deepseek-v4-flash":{"id":"alicloud-deepseek-v4-flash","name":"DeepSeek V4 Flash (Alibaba Cloud)","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"hy4-preview":{"id":"hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":64000},"cost":{"input":0.845,"output":2.535,"cache_read":0.04225},"canonical_model_id":"tencent/hy4-preview"},"coding-glm-5.1":{"id":"coding-glm-5.1","name":"Coding GLM 5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.06,"output":0.22,"cache_read":0.013},"canonical_model_id":"zhipuai/glm-5.1"},"glm-4.6":{"id":"glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.273974,"output":1.095896,"cache_read":0.0547948,"tiers":[{"input":0.547946,"output":2.191784,"cache_read":0.1095892,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6"},"hy3":{"id":"hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0.1562,"output":0.6248,"cache_read":0.03905},"canonical_model_id":"tencent/hy3"},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"grok-build-0.1":{"id":"grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2},"canonical_model_id":"xai/grok-build-0.1"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.1268,"output":3.9438,"cache_read":0.2817},"canonical_model_id":"zhipuai/glm-5.3"},"glm-5v-turbo":{"id":"glm-5v-turbo","name":"GLM 5 Vision Turbo","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":0.7042,"output":3.09848,"cache_read":0.169008,"tiers":[{"input":0.986,"output":3.662004,"cache_read":0.253402,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5v-turbo"},"gemini-3-pro-image":{"id":"gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"interleaved":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":120},"canonical_model_id":"google/gemini-3-pro-image"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"glm-4.7-flash-free":{"id":"glm-4.7-flash-free","name":"GLM-4.7-Flash (free)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"glm-4.6v":{"id":"glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.137,"output":0.411,"cache_read":0.0274,"tiers":[{"input":0.274,"output":0.822,"cache_read":0.0548,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6v"},"doubao-seed-2-0-pro":{"id":"doubao-seed-2-0-pro","name":"Doubao Seed 2.0 Pro","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.4958,"output":2.479,"cache_read":0.09916,"tiers":[{"input":0.7436,"output":3.718,"cache_read":0.14872,"tier":{"type":"context","size":32000}},{"input":1.4872,"output":7.436,"cache_read":0.29744,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-pro"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":1.5,"output":9,"cache_read":0.15},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"coding-minimax-m2.7":{"id":"coding-minimax-m2.7","name":"Coding MiniMax M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128100},"cost":{"input":0.2,"output":0.2},"canonical_model_id":"minimax/MiniMax-M2.7"},"deep-deepseek-v4-pro":{"id":"deep-deepseek-v4-pro","name":"DeepSeek V4 Pro (DeepSeek)","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0.478,"output":0.956,"cache_read":0.004302},"canonical_model_id":"deepseek/deepseek-v4-pro"},"qwen3.8-max-free":{"id":"qwen3.8-max-free","name":"Qwen3.8 Max (free)","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.8-max"}}},"zenmux":{"id":"zenmux","env":["ZENMUX_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://zenmux.ai/api/v1","name":"ZenMux","doc":"https://docs.zenmux.ai","models":{"volcengine/doubao-seed-2.0-pro":{"id":"volcengine/doubao-seed-2.0-pro","name":"Doubao-Seed-2.0-pro","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-02-14","release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0.45,"output":2.24,"cache_read":0.09,"cache_write":0.0024},"canonical_model_id":"bytedance-seed/seed-2.0-pro"},"volcengine/doubao-seed-code":{"id":"volcengine/doubao-seed-code","name":"Doubao-Seed-Code","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-11-11","last_updated":"2025-11-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0.17,"output":1.12,"cache_read":0.03}},"volcengine/doubao-seed-2.0-code":{"id":"volcengine/doubao-seed-2.0-code","name":"Doubao Seed 2.0 Code","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"seed","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"status":"deprecated","cost":{"input":0.9,"output":4.48},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"volcengine/doubao-seed-1.8":{"id":"volcengine/doubao-seed-1.8","name":"Doubao-Seed-1.8","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-12-28","last_updated":"2025-12-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0.11,"output":0.28,"cache_read":0.02,"cache_write":0.0024},"canonical_model_id":"bytedance-seed/seed-1-8"},"volcengine/doubao-seed-2.0-lite":{"id":"volcengine/doubao-seed-2.0-lite","name":"Doubao-Seed-2.0-lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-02-14","release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0.09,"output":0.51,"cache_read":0.02,"cache_write":0.0024},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"volcengine/doubao-seed-2.0-mini":{"id":"volcengine/doubao-seed-2.0-mini","name":"Doubao-Seed-2.0-mini","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-02-14","release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0.03,"output":0.28,"cache_read":0.01,"cache_write":0.0024},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"z-ai/glm-4.5-air":{"id":"z-ai/glm-4.5-air","name":"GLM 4.5 Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":96000},"cost":{"input":0.1165,"output":0.2911,"cache_read":0.0233,"tiers":[{"input":0.1747,"output":1.1645,"cache_read":0.0349,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.5-air"},"z-ai/glm-4.6v-flash-free":{"id":"z-ai/glm-4.6v-flash-free","name":"GLM 4.6V Flash (Free)","description":"Lightweight GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"tiers":[{"input":0,"output":0,"cache_read":0,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6v-flash"},"z-ai/glm-4.7-flashx":{"id":"z-ai/glm-4.7-flashx","name":"GLM 4.7 FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.0728,"output":0.4367,"cache_read":0.0146},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"z-ai/glm-4.6v-flash":{"id":"z-ai/glm-4.6v-flash","name":"GLM 4.6V FlashX","description":"Lightweight GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.0218,"output":0.2184,"cache_read":0.0044,"tiers":[{"input":0.0437,"output":0.4367,"cache_read":0.0044,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6v-flash"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM 5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.8781,"output":3.5126,"cache_read":0.1903,"tiers":[{"input":1.1709,"output":4.098,"cache_read":0.2927,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"GLM 4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.2911,"output":1.1645,"cache_read":0.0582,"tiers":[{"input":0.5823,"output":2.3291,"cache_read":0.1165,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM 5 Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":0.73,"output":3.19,"cache_read":0.174,"tiers":[{"input":1.02,"output":3.77,"cache_read":0.261,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5-turbo"},"z-ai/glm-4.5":{"id":"z-ai/glm-4.5","name":"GLM 4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":96000},"cost":{"input":0.2911,"output":1.1645,"cache_read":0.0582,"tiers":[{"input":0.5823,"output":2.3291,"cache_read":0.1165,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.5"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.075,"output":0.25,"cache_read":0.015},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-5.3-flashx":{"id":"z-ai/glm-5.3-flashx","name":"GLM 5.3 FlashX","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.375,"output":1.25,"cache_read":0.075},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM 5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.58,"output":2.6,"cache_read":0.14,"tiers":[{"input":0.87,"output":3.18,"cache_read":0.22,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-image":{"id":"z-ai/glm-image","type":"image","name":"GLM-Image","description":"GLM-Image is an image generation model adopts a hybrid autoregressive + diffusion decoder architecture. In general image generation quality, GLM‑Image aligns with mainstream latent diffusion approaches, but it shows significant advantages in text-rendering and knowledge‑intensive generation scenarios. It performs especially well in tasks requiring precise semantic understanding and complex information expression, while maintaining strong capabilities in high‑fidelity and fine‑grained detail generation. In addition to text‑to‑image generation, GLM‑Image also supports a rich set of image‑to‑image tasks including image editing, style transfer, identity‑preserving generation, and multi‑subject consistency.","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-14","last_updated":"2026-01-14","modalities":{"input":["text"],"output":["image"]},"open_weights":true,"limit":{"context":10240,"output":0},"canonical_model_id":"zhipuai/glm-image"},"z-ai/glm-4.6":{"id":"z-ai/glm-4.6","name":"GLM 4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.2911,"output":1.1645,"cache_read":0.0582,"tiers":[{"input":0.5823,"output":2.3291,"cache_read":0.1165,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-5v-turbo":{"id":"z-ai/glm-5v-turbo","name":"GLM 5V Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":0.726,"output":3.1946,"cache_read":0.1743,"tiers":[{"input":1.0165,"output":3.7754,"cache_read":0.2614,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5v-turbo"},"z-ai/glm-4.7-flash-free":{"id":"z-ai/glm-4.7-flash-free","name":"GLM 4.7 Flash (Free)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"z-ai/glm-4.6v":{"id":"z-ai/glm-4.6v","name":"GLM 4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.1456,"output":0.4367,"cache_read":0.0291,"tiers":[{"input":0.2911,"output":0.8734,"cache_read":0.0582,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-4.6v"},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":64000},"status":"deprecated","cost":{"input":0.6,"output":2.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"moonshotai/kimi-k2.7-code-free":{"id":"moonshotai/kimi-k2.7-code-free","name":"Kimi K2.7 Code (Free)","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262140,"output":262140},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 0905","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":64000},"status":"deprecated","cost":{"input":0.6,"output":2.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2-0905"},"moonshotai/kimi-k2.7-code-highspeed":{"id":"moonshotai/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code HighSpeed","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k3-free":{"id":"moonshotai/kimi-k3-free","name":"Kimi K3 (Free)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":64000},"status":"deprecated","cost":{"input":0.58,"output":3.02,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/kimi-k2-thinking-turbo":{"id":"moonshotai/kimi-k2-thinking-turbo","name":"Kimi K2 Thinking Turbo","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":64000},"status":"deprecated","cost":{"input":1.15,"output":8,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2-thinking-turbo"},"kuaishou/kat-coder-pro-v2":{"id":"kuaishou/kat-coder-pro-v2","name":"KAT-Coder-Pro-V2","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-30","last_updated":"2026-03-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":80000},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.06}},"inclusionai/ling-1t":{"id":"inclusionai/ling-1t","name":"Ling-1T","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-10-09","last_updated":"2025-10-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"status":"deprecated","cost":{"input":0.56,"output":2.24,"cache_read":0.11}},"inclusionai/ring-2.6-1t":{"id":"inclusionai/ring-2.6-1t","name":"inclusionAI: Ring-2.6-1T","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-12-31","release_date":"2026-05-07","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":65000},"cost":{"input":0.1318155,"output":1.0984625,"cache_read":0.0263631}},"inclusionai/ring-1t":{"id":"inclusionai/ring-1t","name":"Ring-1T","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-10-12","last_updated":"2025-10-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"status":"deprecated","cost":{"input":0.56,"output":2.24,"cache_read":0.11}},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/gpt-5.5-instant":{"id":"openai/gpt-5.5-instant","name":"GPT-5.5 Instant","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12-01","release_date":"2026-05-05","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5-instant"},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-5-codex":{"id":"openai/gpt-5-codex","name":"GPT-5 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5-codex"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/chat-latest":{"id":"openai/chat-latest","name":"Chat Latest (GPT-5.5 Instant)","description":"Compact GPT model for low-latency assistance and high-volume workloads","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12-01","release_date":"2026-05-05","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5-instant"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1-Codex-Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":100000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.3-chat":{"id":"openai/gpt-5.3-chat","name":"GPT-5.3 Chat","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16380},"status":"deprecated","provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.75,"output":14},"canonical_model_id":"openai/gpt-5.3-chat-latest"},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT-5.1-Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"openai/gpt-5.1-chat":{"id":"openai/gpt-5.1-chat","name":"GPT-5.1 Chat","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"status":"deprecated","provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.12},"canonical_model_id":"openai/gpt-5.1-chat-latest"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2-Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://zenmux.ai/api/v1"},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-v3.2-exp":{"id":"deepseek/deepseek-v3.2-exp","name":"DeepSeek-V3.2-Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163000,"output":64000},"status":"deprecated","cost":{"input":0.22,"output":0.33}},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.22,"output":0.66,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8000},"cost":{"input":0.293,"output":0.4395,"cache_read":0.0293},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-chat":{"id":"deepseek/deepseek-chat","name":"DeepSeek-V3.2 (Non-thinking Mode)","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"status":"deprecated","cost":{"input":0.28,"output":0.42,"cache_read":0.03}},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.075,"output":0.3,"cache_read":0.0015},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"anthropic/claude-3.5-haiku":{"id":"anthropic/claude-3.5-haiku","name":"Claude 3.5 Haiku","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07-31","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.8,"output":4,"cache_read":0.08,"cache_write":1},"canonical_model_id":"anthropic/claude-3-5-haiku-20241022"},"anthropic/claude-opus-4.1":{"id":"anthropic/claude-opus-4.1","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4":{"id":"anthropic/claude-opus-4","name":"Claude Opus 4","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-0"},"anthropic/claude-sonnet-4":{"id":"anthropic/claude-sonnet-4","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-0"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-3.7-sonnet":{"id":"anthropic/claude-3.7-sonnet","name":"Claude 3.7 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-10-31","release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-3-7-sonnet-20250219"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-sonnet-5-free":{"id":"anthropic/claude-sonnet-5-free","name":"Claude Sonnet 5 (Free)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"anthropic/claude-sonnet-5"},"tencent/hy3-preview":{"id":"tencent/hy3-preview","name":"Hy3 preview","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0.172,"output":0.572,"cache_read":0.058,"cache_write":0},"canonical_model_id":"tencent/hy3-preview"},"baidu/ernie-5.0-thinking-preview":{"id":"baidu/ernie-5.0-thinking-preview","name":"ERNIE 5.0","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-01-22","last_updated":"2026-01-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"cost":{"input":0.84,"output":3.37,"tiers":[{"input":1.41,"output":5.62,"tier":{"type":"context","size":32000}}]}},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.4,"output":2,"cache_read":0.08,"tiers":[{"input":0.8,"output":4,"cache_read":0.16,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.8,"output":4,"cache_read":0.16}},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2.5-pro":{"id":"xiaomi/mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":1,"output":3,"cache_read":0.2,"tiers":[{"input":2,"output":6,"cache_read":0.4,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.4}},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"xiaomi/mimo-v2-flash":{"id":"xiaomi/mimo-v2-flash","name":"MiMo-V2-Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-12-01","release_date":"2025-12-16","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.1,"output":0.3,"cache_read":0.01},"canonical_model_id":"xiaomi/mimo-v2-flash"},"xiaomi/mimo-v2-omni":{"id":"xiaomi/mimo-v2-omni","name":"MiMo V2 Omni","description":"MiMo omni model for text, image, video, audio, and agents","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":265000,"output":265000},"status":"deprecated","cost":{"input":0.4,"output":2,"cache_read":0.08},"canonical_model_id":"xiaomi/mimo-v2-omni"},"xiaomi/mimo-v2-pro":{"id":"xiaomi/mimo-v2-pro","name":"MiMo V2 Pro","description":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":256000},"status":"deprecated","cost":{"input":1,"output":3,"cache_read":0.2,"tiers":[{"input":2,"output":6,"cache_read":0.4,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.4}},"canonical_model_id":"xiaomi/mimo-v2-pro"},"stepfun/step-3.7-flash":{"id":"stepfun/step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"stepfun/step-3.5-flash":{"id":"stepfun/step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"stepfun/step-3.5-flash"},"stepfun/step-3.7-flash-free":{"id":"stepfun/step-3.7-flash-free","name":"Step 3.7 Flash (Free)","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"stepfun/step-3.7-flash"},"stepfun/step-3":{"id":"stepfun/step-3","name":"Step-3","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["image","text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":64000},"status":"deprecated","cost":{"input":0.21,"output":0.57}},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131070},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204000,"output":131072},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-m2.5-lightning":{"id":"minimax/minimax-m2.5-lightning","name":"MiniMax M2.5 highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.6,"output":2.4,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5-highspeed"},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax M2","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":512000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"tiers":[{"input":0.6,"output":2.4,"cache_read":0.12,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":0.6,"output":2.4,"cache_read":0.12}},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5"},"minimax/minimax-m3.1-flash":{"id":"minimax/minimax-m3.1-flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":512000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax M2.7 highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131070},"provider":{"npm":"@ai-sdk/anthropic","api":"https://zenmux.ai/api/anthropic/v1"},"cost":{"input":0.611,"output":2.4439,"cache_read":0.0611,"cache_write":0.3819},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"bytedance/doubao-seed-2.0-pro":{"id":"bytedance/doubao-seed-2.0-pro","name":"Doubao-Seed-2.0-pro","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.45,"output":2.24,"cache_read":0.09,"tiers":[{"input":0.67,"output":3.36,"cache_read":0.14,"tier":{"type":"context","size":32000}},{"input":1.34,"output":6.71,"cache_read":0.27,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-pro"},"bytedance/doubao-seed-code":{"id":"bytedance/doubao-seed-code","name":"Doubao-Seed-Code","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-11-11","last_updated":"2025-11-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.17,"output":1.12,"cache_read":0.034,"tiers":[{"input":0.2,"output":1.69,"cache_read":0.034,"tier":{"type":"context","size":32000}},{"input":0.39,"output":2.25,"cache_read":0.034,"tier":{"type":"context","size":128000}}]}},"bytedance/doubao-seed-2.0-code":{"id":"bytedance/doubao-seed-2.0-code","name":"Doubao Seed 2.0 Code","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.45,"output":2.24,"cache_read":0.09,"tiers":[{"input":0.67,"output":3.36,"cache_read":0.14,"tier":{"type":"context","size":32000}},{"input":1.34,"output":6.71,"cache_read":0.27,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"bytedance/doubao-seed-1.8":{"id":"bytedance/doubao-seed-1.8","name":"Doubao-Seed-1.8","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2025-12-28","last_updated":"2025-12-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.11,"output":0.28,"cache_read":0.023,"tiers":[{"input":0.17,"output":2.27,"cache_read":0.023,"tier":{"type":"context","size":32000}},{"input":0.34,"output":3.41,"cache_read":0.023,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-8"},"bytedance/doubao-seed-2.0-lite":{"id":"bytedance/doubao-seed-2.0-lite","name":"Doubao-Seed-2.0-lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.09,"output":0.51,"cache_read":0.02,"tiers":[{"input":0.13,"output":0.76,"cache_read":0.03,"tier":{"type":"context","size":32000}},{"input":0.25,"output":1.51,"cache_read":0.05,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"bytedance/doubao-seed-2.0-mini":{"id":"bytedance/doubao-seed-2.0-mini","name":"Doubao-Seed-2.0-mini","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.03,"output":0.28,"cache_read":0.01,"tiers":[{"input":0.06,"output":0.56,"cache_read":0.02,"tier":{"type":"context","size":32000}},{"input":0.12,"output":1.12,"cache_read":0.03,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":1.25,"output":3.75,"cache_read":0.25,"cache_write":1.5625},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.24,"cache_write":1.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.24,"cache_write":1.5}},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3-max":{"id":"qwen/qwen3-max","name":"Qwen3-Max-Thinking","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":1.2,"output":6},"canonical_model_id":"alibaba/qwen3-max"},"qwen/qwen3-coder-plus":{"id":"qwen/qwen3-coder-plus","name":"Qwen3-Coder-Plus","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"alibaba/qwen3-coder-plus"},"qwen/qwen3.5-flash":{"id":"qwen/qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1020000,"output":1020000},"cost":{"input":0.1,"output":0.4,"cache_read":0.02,"cache_write":0.125},"canonical_model_id":"alibaba/qwen3.5-flash"},"qwen/qwen3.5-plus":{"id":"qwen/qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":2.4,"cache_read":0.08,"cache_write":0.5,"tiers":[{"input":0.5,"output":3,"cache_read":0.1,"cache_write":0.625,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.5,"output":3,"cache_read":0.1,"cache_write":0.625}},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6-Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.5,"output":3,"cache_read":0.1,"cache_write":0.625,"tiers":[{"input":2,"output":6,"cache_read":0.4,"cache_write":2.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.4,"cache_write":2.5}},"canonical_model_id":"alibaba/qwen3.6-plus"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048000,"output":65535},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-3.1-flash-lite-preview":{"id":"google/gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":65530},"status":"deprecated","cost":{"input":0.25,"output":1.5},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048000,"output":65535},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048000,"output":65535},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":1,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048000,"output":65535},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048000,"output":65535},"cost":{"input":1.25,"output":10,"cache_read":0.13,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"sapiens-ai/agnes-1.5-pro":{"id":"sapiens-ai/agnes-1.5-pro","name":"Agnes 1.5 Pro","description":"Flagship model for demanding analysis, coding, and production agent workflows","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-21","last_updated":"2026-03-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"status":"deprecated","cost":{"input":0.16,"output":0.8}},"sapiens-ai/agnes-1.5-lite":{"id":"sapiens-ai/agnes-1.5-lite","name":"Agnes 1.5 Lite","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-26","last_updated":"2026-03-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"status":"deprecated","cost":{"input":0.12,"output":0.6}},"x-ai/grok-imagine-image-2.0":{"id":"x-ai/grok-imagine-image-2.0","type":"image","name":"Grok Imagine Image 2.0","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-08-07","last_updated":"2026-08-07","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":66000,"output":0},"canonical_model_id":"xai/grok-imagine-image-2.0"},"x-ai/grok-4.5":{"id":"x-ai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.5"},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"x-ai/grok-voice-tts-1.0":{"id":"x-ai/grok-voice-tts-1.0","type":"speech","name":"Grok Voice TTS 1.0","description":"Convert text into spoken audio with a single API call. The API supports a rich set of expressive voices, inline speech tags for fine-grained delivery control, and output formats from high-fidelity MP3 to telephony-optimized μ-law.","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":15000,"output":15000},"canonical_model_id":"xai/grok-tts"},"x-ai/grok-voice-stt-1.0":{"id":"x-ai/grok-voice-stt-1.0","type":"transcription","name":"Grok Voice STT 1.0","description":"Speech-to-text model with word-level timestamps, speaker diarization, multichannel audio, and streaming","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":15000,"output":15000},"canonical_model_id":"xai/grok-voice-transcribe-2.0"},"x-ai/grok-4.2-fast-non-reasoning":{"id":"x-ai/grok-4.2-fast-non-reasoning","name":"Grok 4.2 Fast Non Reasoning","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2,"tiers":[{"input":4,"output":12,"cache_read":0.2,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"x-ai/grok-4.2-fast":{"id":"x-ai/grok-4.2-fast","name":"Grok 4.2 Fast","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2,"tiers":[{"input":4,"output":12,"cache_read":0.2,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"x-ai/grok-build-0.1":{"id":"x-ai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2},"canonical_model_id":"xai/grok-build-0.1"}}},"opencode":{"id":"opencode","env":["OPENCODE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://opencode.ai/inference/openai/v1","name":"OpenCode","doc":"https://opencode.ai/v2/docs/console/models/","models":{"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":1.5,"output":7.5,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.8-flash"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"kimi-k2-thinking":{"id":"kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Legacy model retained for compatibility with older integrations","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.4,"output":2.5,"cache_read":0.4},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.07,"output":8.5,"cache_read":0.107},"canonical_model_id":"openai/gpt-5.1"},"big-pickle":{"id":"big-pickle","name":"Big Pickle","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"big-pickle","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-10-17","last_updated":"2025-10-17","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":160000,"output":32000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"muse-wonder-1.0":{"id":"muse-wonder-1.0","name":"Muse Wonder","description":"Muse Wonder is a frontier multimodal reasoning model from Meta for personal and professional agents, long-horizon coding, and computer use. It combines Active Memory, which carries relevant context across conversations, with Active Focus, which keeps it working through blockers toward a goal.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"nemotron-3.5-super-vl-free":{"id":"nemotron-3.5-super-vl-free","name":"Nemotron 3.5 Super VL Free","description":"NVIDIA vision-language reasoning model for text, image, and video understanding and agentic tool use","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"glm-4.7-free":{"id":"glm-4.7-free","name":"GLM-4.7 Free","description":"Legacy model retained for compatibility with older integrations","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-4.7"},"qwen3.6-plus-free":{"id":"qwen3.6-plus-free","name":"Qwen3.6 Plus Free","description":"Legacy model retained for compatibility with older integrations","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"alibaba/qwen3.6-plus"},"longcat-2.0-free":{"id":"longcat-2.0-free","name":"LongCat-2.0 Free","description":"Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"meituan/longcat-2.0"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"minimax-m3-free":{"id":"minimax-m3-free","name":"MiniMax-M3 Free","description":"Legacy model retained for compatibility with older integrations","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"minimax/MiniMax-M3"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"gpt-5.5-pro":{"id":"gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":30,"output":180,"cache_read":30},"canonical_model_id":"openai/gpt-5.5-pro"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":2,"output":10,"cache_read":0.1,"input_audio":2},"canonical_model_id":"google/gemini-4-argon"},"step-5-preview-free":{"id":"step-5-preview-free","name":"Step 5 Preview Free","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"stepfun/step-5-preview"},"muse-spark-1.3-contributor-free":{"id":"muse-spark-1.3-contributor-free","name":"Muse Spark 1.3 Free","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for coding and agentic workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"meta/muse-spark-1.3"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"hy3-preview-free":{"id":"hy3-preview-free","name":"Hy3 preview Free","description":"Legacy model retained for compatibility with older integrations","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"tencent/hy3-preview"},"glm-5.3-flash-free":{"id":"glm-5.3-flash-free","name":"GLM-5.3-Flash Free","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"mimo-v2-flash-free":{"id":"mimo-v2-flash-free","name":"MiMo V2 Flash Free","description":"Legacy model retained for compatibility with older integrations","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12-01","release_date":"2025-12-16","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2-flash"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"space-bunny-free":{"id":"space-bunny-free","name":"Space Bunny Free","description":"Anonymous preview reasoning model for coding, agentic tasks, tool use, and multimodal input","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":524288,"output":524288},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"minimax-m3.1-flash-free":{"id":"minimax-m3.1-flash-free","name":"MiniMax M3.1 Flash Free","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":512000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"muse-spark-1.3":{"id":"muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"gpt-5.4-pro":{"id":"gpt-5.4-pro","name":"GPT-5.4 Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":30,"output":180,"cache_read":30},"canonical_model_id":"openai/gpt-5.4-pro"},"exo-free":{"id":"exo-free","name":"Exo Free","description":"Free-tier preview reasoning model with text and image input and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"temperature":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0}},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"ll-alpha-free":{"id":"ll-alpha-free","name":"LL Alpha Free","description":"Best in class local model","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-06","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"mistral-large-4":{"id":"mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":262144},"provider":{"npm":"@ai-sdk/mistral","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.68,"output":2.09,"cache_read":0.07},"canonical_model_id":"mistral/mistral-large-4"},"gemini-3-flash":{"id":"gemini-3-flash","name":"Gemini 3 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3-flash-preview"},"mimo-v2-omni-free":{"id":"mimo-v2-omni-free","name":"MiMo V2 Omni Free","description":"Legacy model retained for compatibility with older integrations","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text","image","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":64000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2-omni"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"claude-opus-4-1":{"id":"claude-opus-4-1","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"flint-40b-free":{"id":"flint-40b-free","name":"Flint 40B Free","description":"Open-weight mixture-of-experts model with about 2.5B active parameters for local coding agents","family":"flint","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-06","release_date":"2026-10-21","last_updated":"2026-10-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"recursive/flint"},"mimo-v2.5-free":{"id":"mimo-v2.5-free","name":"MiMo V2.5 Free","description":"MiMo omni model for text, image, video, audio, and agents","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"interleaved":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"mimo-v2.6-flash-free":{"id":"mimo-v2.6-flash-free","name":"MiMo-V2.6-Flash Free","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"gpt-5-codex":{"id":"gpt-5-codex","name":"GPT-5 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.07,"output":8.5,"cache_read":0.107},"canonical_model_id":"openai/gpt-5-codex"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"deepseek-v4-flash-vision-exp":{"id":"deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.74,"output":3.48,"cache_read":0.145},"canonical_model_id":"deepseek/deepseek-v4-pro"},"ling-3.0-flash-fin-free":{"id":"ling-3.0-flash-fin-free","name":"Ling 3.0 Flash Fin Free","description":"Finance-enhanced model for financial research, multi-step investment workflows, and long-horizon planning and execution","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-09-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"inclusionai/ling-3.0-flash-fin"},"beam-free":{"id":"beam-free","name":"Beam Free","description":"Sparse mixture-of-experts reasoning model with 23B active parameters for coding and agentic workflows","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"deepseek-v4-flash-free":{"id":"deepseek-v4-flash-free","name":"DeepSeek V4 Flash Free","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"gpt-5.3-codex-spark":{"id":"gpt-5.3-codex-spark","name":"GPT-5.3 Codex Spark","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex-spark","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":128000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex-spark"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"claude-3-5-haiku":{"id":"claude-3-5-haiku","name":"Claude Haiku 3.5","description":"Legacy model retained for compatibility with older integrations","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07-31","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0.8,"output":4,"cache_read":0.08,"cache_write":1},"canonical_model_id":"anthropic/claude-3-5-haiku-20241022"},"minimax-m2.1":{"id":"minimax-m2.1","name":"MiniMax-M2.1","description":"Legacy model retained for compatibility with older integrations","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.1},"canonical_model_id":"minimax/MiniMax-M2.1"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"north-mini-code-free":{"id":"north-mini-code-free","name":"North Mini Code Free","description":"Cohere coding model for practical software engineering and agentic edits","family":"north","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-09-23","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0},"canonical_model_id":"cohere/north-mini-code-1-0"},"hy3-free":{"id":"hy3-free","name":"Hy3 Free","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":190000,"input":192000,"output":64000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"tencent/hy3"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"laguna-s-2.1-free":{"id":"laguna-s-2.1-free","name":"Laguna S 2.1 Free","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"poolside/laguna-s-2.1"},"nemotron-3.5-lightning-free":{"id":"nemotron-3.5-lightning-free","name":"Nemotron 3.5 Lightning Free","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"glm-5-free":{"id":"glm-5-free","name":"GLM-5 Free","description":"Legacy model retained for compatibility with older integrations","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"zhipuai/glm-5"},"minimax-m2.5-free":{"id":"minimax-m2.5-free","name":"MiniMax-M2.5 Free","description":"Legacy model retained for compatibility with older integrations","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"minimax/MiniMax-M2.5"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"minimax-m2.1-free":{"id":"minimax-m2.1-free","name":"MiniMax-M2.1 Free","description":"Legacy model retained for compatibility with older integrations","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"minimax/MiniMax-M2.1"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"ling-3.0-flash-free":{"id":"ling-3.0-flash-free","name":"Ling-3.0-flash Free","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"inclusionai/ling-3.0-flash"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"gemini-3-pro":{"id":"gemini-3-pro","name":"Gemini 3 Pro","description":"Legacy model retained for compatibility with older integrations","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3-pro-preview"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"fledge-alpha-free":{"id":"fledge-alpha-free","name":"Fledge Alpha Free","description":"Free-tier preview reasoning model with text and image input and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"temperature":true,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0}},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Legacy model retained for compatibility with older integrations","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.6,"output":2.2,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-4.7"},"deepseek-v4.1-flash-free":{"id":"deepseek-v4.1-flash-free","name":"DeepSeek V4.1 Flash Free","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"kimi-k2.5-free":{"id":"kimi-k2.5-free","name":"Kimi K2.5 Free","description":"Legacy model retained for compatibility with older integrations","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"moonshotai/kimi-k2.5"},"x-preview-f-free":{"id":"x-preview-f-free","name":"Ox Alpha Free (Unlimited)","description":"Stealth reasoning model for coding, agentic tasks, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-08-21","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"longcat-2.5-preview-free":{"id":"longcat-2.5-preview-free","name":"LongCat 2.5 Preview Free","description":"Meituan's multimodal reasoning model for coding and agent tasks, with image understanding and a 1M-token context window","family":"longcat","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"meituan/longcat-2.5-preview"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":1.5,"output":7.5,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.6-flash"},"claude-sonnet-4":{"id":"claude-sonnet-4","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-0"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":1.5,"output":7.5,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.7-flash"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"muse-spark-1.2":{"id":"muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"ling-3.0-tiny-free":{"id":"ling-3.0-tiny-free","name":"Ling-3.0-tiny Free","description":"Compact MoE model for responsive agents, instruction following, and multi-turn conversations","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0}},"grok-code":{"id":"grok-code","name":"Grok Code Fast 1","description":"Legacy model retained for compatibility with older integrations","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"xai/grok-code-fast-1"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"nemotron-3-ultra-free":{"id":"nemotron-3-ultra-free","name":"Nemotron 3 Ultra Free","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":128000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"ling-3.1-flash-free":{"id":"ling-3.1-flash-free","name":"Ling 3.1 Flash Free","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"inclusionai/ling-3.1-flash"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"ring-2.6-1t-free":{"id":"ring-2.6-1t-free","name":"Ring 2.6 1T Free","description":"Legacy model retained for compatibility with older integrations","family":"ring-1t-free","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-06","release_date":"2026-05-08","last_updated":"2026-05-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":66000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0}},"nemotron-3-super-free":{"id":"nemotron-3-super-free","name":"Nemotron 3 Super Free","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"gpt-5.1-codex-max":{"id":"gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex-max"},"gpt-5.1-codex-mini":{"id":"gpt-5.1-codex-mini","name":"GPT-5.1 Codex Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"qwen3.5-plus":{"id":"qwen3.5-plus","name":"Qwen3.5 Plus","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"alibaba/qwen3.5-plus"},"minimax-m2.5":{"id":"minimax-m2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.5"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.07,"output":8.5,"cache_read":0.107},"canonical_model_id":"openai/gpt-5"},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"provider":{"npm":"@ai-sdk/anthropic","api":"https://opencode.ai/inference/anthropic/v1"},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625},"canonical_model_id":"alibaba/qwen3.6-plus"},"glm-5":{"id":"glm-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"mimo-v2-pro-free":{"id":"mimo-v2-pro-free","name":"MiMo V2 Pro Free","description":"Legacy model retained for compatibility with older integrations","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":64000},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2-pro"},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.07,"output":8.5,"cache_read":0.107},"canonical_model_id":"openai/gpt-5.1-codex"},"ling-2.6-flash-free":{"id":"ling-2.6-flash-free","name":"Ling 2.6 Flash Free","description":"Legacy model retained for compatibility with older integrations","family":"ling-flash-free","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262100,"output":32800},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0}},"glm-4.6":{"id":"glm-4.6","name":"GLM-4.6","description":"Legacy model retained for compatibility with older integrations","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.6,"output":2.2,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-4.6"},"trinity-large-preview-free":{"id":"trinity-large-preview-free","name":"Trinity Large Preview","description":"Legacy model retained for compatibility with older integrations","family":"trinity","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-01-27","last_updated":"2026-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0},"canonical_model_id":"arcee-ai/trinity-large-preview"},"grok-build-0.1":{"id":"grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1,"output":2,"cache_read":0.2},"canonical_model_id":"xai/grok-build-0.1"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"cat-jam-free":{"id":"cat-jam-free","name":"Cat Jam Free","description":"A text-only reasoning model for coding and agentic workflows, with tool calling and structured JSON outputs.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0}},"muse-spark-1.2-contributor-free":{"id":"muse-spark-1.2-contributor-free","name":"Muse Spark 1.2 Free","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"meta/muse-spark-1.2"},"kimi-k2":{"id":"kimi-k2","name":"Kimi K2","description":"Legacy model retained for compatibility with older integrations","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.4,"output":2.5,"cache_read":0.4},"canonical_model_id":"moonshotai/kimi-k2-0905"},"gemini-3.1-pro":{"id":"gemini-3.1-pro","name":"Gemini 3.1 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"provider":{"npm":"@ai-sdk/google","api":"https://opencode.ai/inference/google/v1beta"},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"qwen3-coder":{"id":"qwen3-coder","name":"Qwen3 Coder","description":"Legacy model retained for compatibility with older integrations","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","provider":{"api":"https://opencode.ai/inference/openai/v1"},"cost":{"input":0.45,"output":1.8},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"}}},"snowflake-cortex":{"id":"snowflake-cortex","env":["SNOWFLAKE_ACCOUNT","SNOWFLAKE_CORTEX_PAT"],"npm":"@ai-sdk/openai-compatible","api":"https://${SNOWFLAKE_ACCOUNT}.snowflakecomputing.com/api/v2/cortex/v1","name":"Snowflake Cortex","doc":"https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-rest-api","models":{"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"openai-gpt-5.2":{"id":"openai-gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.18},"canonical_model_id":"openai/gpt-5.2"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"openai-gpt-5":{"id":"openai-gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"input":272000,"output":8192},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5"},"openai-gpt-4.1":{"id":"openai-gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai-gpt-5.6-terra":{"id":"openai-gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"openai-gpt-5.6-sol":{"id":"openai-gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"openai-gpt-5.5":{"id":"openai-gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"status":"beta","cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"openai-gpt-5.6-luna":{"id":"openai-gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai-gpt-5-mini":{"id":"openai-gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"input":272000,"output":8192},"cost":{"input":0.25,"output":2,"cache_read":0.03},"canonical_model_id":"openai/gpt-5-mini"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"status":"beta","cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"deepseek-r1":{"id":"deepseek-r1","name":"DeepSeek-R1","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"canonical_model_id":"deepseek/deepseek-r1"},"openai-gpt-5-nano":{"id":"openai-gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"input":272000,"output":8192},"cost":{"input":0.06,"output":0.44,"cache_read":0.01},"canonical_model_id":"openai/gpt-5-nano"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"mistral-large2":{"id":"mistral-large2","name":"Mistral Large 2407","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-03-31","release_date":"2024-07-24","last_updated":"2024-07-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/mistral-large-2407"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"openai-gpt-5.4":{"id":"openai-gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai-gpt-5.1":{"id":"openai-gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"input":272000,"output":8192},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5.1"},"gemini-3.1-pro":{"id":"gemini-3.1-pro","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"beta","canonical_model_id":"google/gemini-3.1-pro-preview"},"snowflake-llama3.3-70b":{"id":"snowflake-llama3.3-70b","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"canonical_model_id":"meta/llama-3.3-70b-instruct"}}},"edenai":{"id":"edenai","env":["EDENAI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.edenai.run/v3","name":"Eden AI","doc":"https://docs.edenai.co","models":{"deepseek-v4-flash-vision-exp":{"id":"deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"meta/muse-wonder-1.0":{"id":"meta/muse-wonder-1.0","name":"Muse Wonder (Meta)","description":"Muse Wonder is a frontier multimodal reasoning model from Meta for personal and professional agents, long-horizon coding, and computer use. It combines Active Memory, which carries relevant context across conversations, with Active Focus, which keeps it working through blockers toward a goal.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3 (Meta)","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1 (Meta)","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2 (Meta)","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"azure/gpt-5.2-codex":{"id":"azure/gpt-5.2-codex","name":"GPT-5.2 Codex (Azure)","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"azure/gpt-5.1-codex-max":{"id":"azure/gpt-5.1-codex-max","name":"GPT-5.1 Codex Max (Azure)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex-max"},"azure/gpt-5.1-codex-mini":{"id":"azure/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini (Azure)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"azure/gpt-5.1-codex":{"id":"azure/gpt-5.1-codex","name":"GPT-5.1 Codex (Azure)","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"fireworks_ai/gpt-oss-120b":{"id":"fireworks_ai/gpt-oss-120b","name":"GPT OSS 120B (Fireworks AI)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.014},"canonical_model_id":"openai/gpt-oss-120b"},"fireworks_ai/accounts/fireworks/models/gpt-oss-120b":{"id":"fireworks_ai/accounts/fireworks/models/gpt-oss-120b","name":"GPT OSS 120B (Fireworks AI)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"openai/gpt-oss-120b"},"fireworks_ai/accounts/fireworks/models/inkling":{"id":"fireworks_ai/accounts/fireworks/models/inkling","name":"Inkling (Fireworks AI)","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"fireworks_ai/accounts/fireworks/models/ember-1":{"id":"fireworks_ai/accounts/fireworks/models/ember-1","name":"Ember-1 (Fireworks AI)","description":"Specialized model from Fireworks built on Kimi K3, producing shorter reasoning traces with approximately 40% fewer tokens while maintaining comparable quality","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"fireworks/ember-1"},"fireworks_ai/accounts/fireworks/models/muse-spark-open":{"id":"fireworks_ai/accounts/fireworks/models/muse-spark-open","name":"Muse Spark Open (Fireworks AI)","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"greenpt/qwen3-235b-a22b-instruct-2507":{"id":"greenpt/qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B-A22B Instruct 2507 (Greenpt)","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":250000,"output":16384},"cost":{"input":1.01106,"output":3.03318},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"greenpt/gpt-oss-120b":{"id":"greenpt/gpt-oss-120b","name":"GPT OSS 120B (Greenpt)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.22468,"output":0.78638},"canonical_model_id":"openai/gpt-oss-120b"},"greenpt/llama-3.3-70b-instruct":{"id":"greenpt/llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct (Greenpt)","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":1.23574,"output":1.23574},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"greenpt/deepseek-v4.1-flash":{"id":"greenpt/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Greenpt)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.247148,"output":1.23574,"cache_read":0.012357},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"greenpt/deepseek-v4-flash-0731":{"id":"greenpt/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731 (Greenpt)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.157276,"output":0.39319,"cache_read":0.044936},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"flexai/gpt-oss-20b":{"id":"flexai/gpt-oss-20b","name":"GPT OSS 20B (FlexAI)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.02,"output":0.1,"cache_read":0.003},"canonical_model_id":"openai/gpt-oss-20b"},"flexai/DeepSeek-V4-Flash-0731":{"id":"flexai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731 (FlexAI)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.119,"output":0.238,"cache_read":0.0178},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"flexai/gpt-oss-120b":{"id":"flexai/gpt-oss-120b","name":"GPT OSS 120B (FlexAI)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.05,"output":0.25,"cache_read":0.0075},"canonical_model_id":"openai/gpt-oss-120b"},"flexai/Muse-Glimmer-30B":{"id":"flexai/Muse-Glimmer-30B","name":"Muse Glimmer 30B (FlexAI)","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.045},"canonical_model_id":"meta/muse-glimmer-30b"},"flexai/Step-3.7-Flash":{"id":"flexai/Step-3.7-Flash","name":"Step 3.7 Flash (FlexAI)","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.03},"canonical_model_id":"stepfun/step-3.7-flash"},"flexai/DeepSeek-V4-Flash-0731@eu":{"id":"flexai/DeepSeek-V4-Flash-0731@eu","name":"DeepSeek V4 Flash 0731 (FlexAI, EU)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.119,"output":0.238,"cache_read":0.0178},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"flexai/DeepSeek-V4.1-Flash":{"id":"flexai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash (FlexAI)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.103,"output":0.715,"cache_read":0.0154},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"mistral/devstral-2512":{"id":"mistral/devstral-2512","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/devstral-2512"},"mistral/mistral-medium-latest":{"id":"mistral/mistral-medium-latest","name":"Mistral Medium (latest)","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"mistral/mistral-medium-latest"},"mistral/magistral-medium-latest":{"id":"mistral/magistral-medium-latest","name":"Magistral Medium (latest)","description":"Mistral reasoning model for transparent analysis, math, and complex decisions","family":"magistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-03-17","last_updated":"2025-03-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":16384},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"mistral/magistral-medium-latest"},"mistral/ministral-3b-2512":{"id":"mistral/ministral-3b-2512","name":"Ministral 3 3B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":104857},"cost":{"input":0.1,"output":0.1,"cache_read":0.01},"canonical_model_id":"mistral/ministral-3b-2512"},"mistral/mistral-small-latest":{"id":"mistral/mistral-small-latest","name":"Mistral Small (latest)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"mistral/mistral-small-latest"},"mistral/mistral-medium-2604":{"id":"mistral/mistral-medium-2604","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"mistral/mistral-medium-2604"},"mistral/mistral-large-latest":{"id":"mistral/mistral-large-latest","name":"Mistral Large (latest)","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mistral-large-latest"},"mistral/devstral-medium-latest":{"id":"mistral/devstral-medium-latest","name":"Devstral 2 (latest)","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/devstral-medium-latest"},"mistral/mistral-medium-3":{"id":"mistral/mistral-medium-3","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":104857},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/mistral-medium-3"},"mistral/ministral-8b-2512":{"id":"mistral/ministral-8b-2512","name":"Ministral 3 8B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.15,"output":0.15,"cache_read":0.015},"canonical_model_id":"mistral/ministral-8b-2512"},"mistral/mistral-small-2603":{"id":"mistral/mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"mistral/mistral-small-2603"},"mistral/codestral-2508":{"id":"mistral/codestral-2508","name":"Codestral 25.08","description":"Mistral code model for low-latency completion, fill-in-the-middle, and code generation","family":"codestral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":8192},"cost":{"input":0.3,"output":0.9,"cache_read":0.03},"canonical_model_id":"mistral/codestral-2508"},"mistral/ministral-14b-2512":{"id":"mistral/ministral-14b-2512","name":"Ministral 3 14B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.2,"output":0.2,"cache_read":0.02},"canonical_model_id":"mistral/ministral-14b-2512"},"mistral/voxtral-small-latest":{"id":"mistral/voxtral-small-latest","name":"Voxtral Small (latest)","description":"Instruct model with native audio input for speech understanding and tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32000},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"mistral/voxtral-small-latest"},"mistral/mistral-large-2407":{"id":"mistral/mistral-large-2407","name":"Mistral Large 2407","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-03-31","release_date":"2024-07-24","last_updated":"2024-07-24","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":104857},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mistral-large-2407"},"mistral/codestral-latest":{"id":"mistral/codestral-latest","name":"Codestral (latest)","description":"Mistral code model for completions, refactors, and developer IDE workflows","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-05-29","last_updated":"2025-01-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":4096},"cost":{"input":0.3,"output":0.9,"cache_read":0.03},"canonical_model_id":"mistral/codestral-latest"},"mistral/mistral-medium-2505":{"id":"mistral/mistral-medium-2505","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/mistral-medium-2505"},"mistral/mistral-medium-3-5":{"id":"mistral/mistral-medium-3-5","name":"Mistral Medium 3.5","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":209715},"cost":{"input":1.5,"output":7.5,"cache_read":0.15},"canonical_model_id":"mistral/mistral-medium-3-5"},"mistral/mistral-large-2512":{"id":"mistral/mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"mistral/mistral-large-2512"},"groq/openai/gpt-oss-20b":{"id":"groq/openai/gpt-oss-20b","name":"GPT OSS 20B (Groq)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-oss-20b"},"groq/openai/gpt-oss-safeguard-20b":{"id":"groq/openai/gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B (Groq)","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"groq/openai/gpt-oss-120b":{"id":"groq/openai/gpt-oss-120b","name":"GPT OSS 120B (Groq)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-oss-120b"},"groq/meta-llama/llama-prompt-guard-2-22m":{"id":"groq/meta-llama/llama-prompt-guard-2-22m","name":"Llama Prompt Guard 2 22M (Groq)","description":"22M-parameter (DeBERTa-xsmall) English classifier that detects prompt injection and jailbreak attempts","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":512},"cost":{"input":0.03,"output":0.03,"cache_read":0.015},"canonical_model_id":"meta/llama-prompt-guard-2-22m"},"groq/meta-llama/llama-prompt-guard-2-86m":{"id":"groq/meta-llama/llama-prompt-guard-2-86m","name":"Llama Prompt Guard 2 86M (Groq)","description":"86M-parameter (mDeBERTa-base) multilingual classifier that detects prompt injection and jailbreak attempts","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":512},"cost":{"input":0.04,"output":0.04,"cache_read":0.02},"canonical_model_id":"meta/llama-prompt-guard-2-86m"},"cerebras/gpt-oss-120b":{"id":"cerebras/gpt-oss-120b","name":"GPT OSS 120B (Cerebras)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.35,"output":0.75,"cache_read":0.35},"canonical_model_id":"openai/gpt-oss-120b"},"tensorx/moonshotai/kimi-k2.5":{"id":"tensorx/moonshotai/kimi-k2.5","name":"Kimi K2.5 (TensorX)","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":2.8,"cache_read":0.125},"canonical_model_id":"moonshotai/kimi-k2.5"},"tensorx/deepseek/deepseek-r1-0528":{"id":"tensorx/deepseek/deepseek-r1-0528","name":"DeepSeek R1 0528 (TensorX)","description":"May 2025 DeepSeek R1 update with improved mathematics, coding, and reasoning","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":164000,"output":32768},"cost":{"input":0.66,"output":2.6,"cache_read":0.165},"canonical_model_id":"deepseek/deepseek-r1-0528"},"tensorx/deepseek/deepseek-v4.1-flash-1017":{"id":"tensorx/deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (TensorX)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.5,"output":1.5,"cache_read":0.125},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"tensorx/deepseek/deepseek-v4-pro-0813":{"id":"tensorx/deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813 (TensorX)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":2,"output":4,"cache_read":0.5},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"tensorx/deepseek/deepseek-v4.1-flash":{"id":"tensorx/deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (TensorX)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.5,"output":1.5,"cache_read":0.125},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"tensorx/deepseek/deepseek-v4-flash-0731":{"id":"tensorx/deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731 (TensorX)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.25,"output":0.3,"cache_read":0.0625},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"together_ai/thinkingmachines/Inkling":{"id":"together_ai/thinkingmachines/Inkling","name":"Inkling (Together AI)","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":1048576},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"together_ai/openai/gpt-oss-120b":{"id":"together_ai/openai/gpt-oss-120b","name":"GPT OSS 120B (Together AI)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"together_ai/meta-models/Muse-Glimmer-30B":{"id":"together_ai/meta-models/Muse-Glimmer-30B","name":"Muse Glimmer 30B (Together AI)","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.35,"output":1.5,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"together_ai/reflection/Beam-1.0-501B-A23B":{"id":"together_ai/reflection/Beam-1.0-501B-A23B","name":"Beam 1.0 501B A23B (Together AI)","description":"Sparse mixture-of-experts reasoning model with 23B active parameters for coding and agentic workflows","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.4,"output":1.2,"cache_read":0.04},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"together_ai/deepseek-ai/DeepSeek-V4.1-Flash-1017":{"id":"together_ai/deepseek-ai/DeepSeek-V4.1-Flash-1017","name":"DeepSeek V4.1 Flash 1017 (Together AI)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"together_ai/deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"together_ai/deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731 (Together AI)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"together_ai/deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"together_ai/deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813 (Together AI)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"together_ai/deepseek-ai/DeepSeek-V4.1-Flash":{"id":"together_ai/deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash (Together AI)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"openai/gpt-mini-latest":{"id":"openai/gpt-mini-latest","name":"GPT Mini Latest (GPT-5.6 Luna)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5-search-api":{"id":"openai/gpt-5-search-api","type":"chat","name":"GPT-5 Search API","description":"GPT-5 model trained for web search in the Chat Completions API","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2025-10-14","last_updated":"2025-10-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":272000,"output":100000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5-search-api"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-latest":{"id":"openai/gpt-latest","name":"GPT Latest (GPT-6 Astra)","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-4":{"id":"openai/gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8191,"output":8192},"cost":{"input":30,"output":60},"canonical_model_id":"openai/gpt-4"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/o1-pro":{"id":"openai/o1-pro","name":"o1-pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2025-03-19","last_updated":"2025-03-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":150,"output":600},"canonical_model_id":"openai/o1-pro"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-pro-latest":{"id":"openai/gpt-pro-latest","name":"GPT Pro Latest (GPT-5.6 Sol)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4o-2024-11-20":{"id":"openai/gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"openai/gpt-4o-2024-08-06":{"id":"openai/gpt-4o-2024-08-06","name":"GPT-4o (2024-08-06)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-08-06"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5-turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/o3-pro":{"id":"openai/o3-pro","name":"o3-pro","description":"High-effort o3 tier for difficult technical reasoning and careful answers","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","pdf","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":20,"output":80},"canonical_model_id":"openai/o3-pro"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2 Pro","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-chat":{"id":"deepseek/deepseek-chat","name":"DeepSeek Chat","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-chat"},"anthropic/claude-sonnet-5-5":{"id":"anthropic/claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4-8":{"id":"anthropic/claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-fable-latest":{"id":"anthropic/claude-fable-latest","name":"Claude Fable Latest (Claude Fable 5.1)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-haiku-5-5":{"id":"anthropic/claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-5-5":{"id":"anthropic/claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-fable-5-1":{"id":"anthropic/claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-opus-4-5-20251101":{"id":"anthropic/claude-opus-4-5-20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-haiku-latest":{"id":"anthropic/claude-haiku-latest","name":"Claude Haiku Latest (Claude Haiku 5.5)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-4-5":{"id":"anthropic/claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-opus-latest":{"id":"anthropic/claude-opus-latest","name":"Claude Opus Latest (Claude Opus 5.5)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-sonnet-latest":{"id":"anthropic/claude-sonnet-latest","name":"Claude Sonnet Latest (Claude Sonnet 5.5)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"ovhcloud/gpt-oss-20b":{"id":"ovhcloud/gpt-oss-20b","name":"GPT OSS 20B (OVHcloud)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.05,"output":0.18},"canonical_model_id":"openai/gpt-oss-20b"},"ovhcloud/gpt-oss-120b":{"id":"ovhcloud/gpt-oss-120b","name":"GPT OSS 120B (OVHcloud)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.09,"output":0.47},"canonical_model_id":"openai/gpt-oss-120b"},"tencent/step-5-preview":{"id":"tencent/step-5-preview","name":"Step 5 Preview (Tencent)","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":1000000},"cost":{"input":1,"output":2.7,"cache_read":0.05},"canonical_model_id":"stepfun/step-5-preview"},"tencent/deepseek-v4-pro-0813":{"id":"tencent/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813 (Tencent)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"tencent/deepseek-v4.1-flash":{"id":"tencent/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Tencent)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 preview (Tencent)","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3 (Tencent)","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0.132,"output":0.528,"cache_read":0.033},"canonical_model_id":"tencent/hy3"},"moonshot/kimi-k2.6":{"id":"moonshot/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshot/kimi-k2.7-code-highspeed":{"id":"moonshot/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshot/kimi-k2.7-code":{"id":"moonshot/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshot/kimi-k3":{"id":"moonshot/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"vertex/gemini-3.8-flash":{"id":"vertex/gemini-3.8-flash","name":"Gemini 3.8 Flash (Vertex AI)","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.8-flash"},"vertex/gemini-3.1-flash-lite":{"id":"vertex/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite (Vertex AI)","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.083333,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"vertex/gemini-4-argon":{"id":"vertex/gemini-4-argon","name":"Gemini 4 Argon (Vertex AI)","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":4,"output":20,"reasoning":20,"cache_read":0.2,"cache_write":0.375,"input_audio":4},"canonical_model_id":"google/gemini-4-argon"},"vertex/gemini-3.8-flash@eu":{"id":"vertex/gemini-3.8-flash@eu","name":"Gemini 3.8 Flash (Vertex AI, EU)","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.8-flash"},"vertex/gemini-4-argon@us":{"id":"vertex/gemini-4-argon@us","name":"Gemini 4 Argon (Vertex AI, US)","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":4,"output":20,"reasoning":20,"cache_read":0.2,"cache_write":0.375,"input_audio":4},"canonical_model_id":"google/gemini-4-argon"},"vertex/gemini-pro-latest":{"id":"vertex/gemini-pro-latest","name":"Gemini Pro Latest (Gemini 4 Argon, Vertex AI)","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":4,"output":20,"reasoning":20,"cache_read":0.2,"cache_write":0.375,"input_audio":4},"canonical_model_id":"google/gemini-4-argon"},"vertex/gemini-3.8-flash@us":{"id":"vertex/gemini-3.8-flash@us","name":"Gemini 3.8 Flash (Vertex AI, US)","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.8-flash"},"vertex/gemini-3.1-flash-lite-image":{"id":"vertex/gemini-3.1-flash-lite-image","type":"image","name":"Nano Banana 2 Lite (Vertex AI)","description":"Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":4096},"cost":{"input":0.25,"output":1.5},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"vertex/gemini-2.5-flash-image":{"id":"vertex/gemini-2.5-flash-image","type":"image","name":"Nano Banana (Vertex AI)","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash-image"},"vertex/gemini-3.1-flash-lite@us":{"id":"vertex/gemini-3.1-flash-lite@us","name":"Gemini 3.1 Flash Lite (Vertex AI, US)","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.083333,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"vertex/gemini-4-argon@eu":{"id":"vertex/gemini-4-argon@eu","name":"Gemini 4 Argon (Vertex AI, EU)","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":4,"output":20,"reasoning":20,"cache_read":0.2,"cache_write":0.375,"input_audio":4},"canonical_model_id":"google/gemini-4-argon"},"vertex/gemini-flash-latest":{"id":"vertex/gemini-flash-latest","name":"Gemini Flash Latest (Gemini 3.8 Flash, Vertex AI)","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.8-flash"},"vertex/gemini-3.1-flash-lite@eu":{"id":"vertex/gemini-3.1-flash-lite@eu","name":"Gemini 3.1 Flash Lite (Vertex AI, EU)","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.083333,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"vertex/gemini-3.5-flash-lite@eu":{"id":"vertex/gemini-3.5-flash-lite@eu","name":"Gemini 3.5 Flash Lite (Vertex AI, EU)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333,"input_audio":0.3},"canonical_model_id":"google/gemini-3.5-flash-lite"},"vertex/gemini-3.5-flash-lite":{"id":"vertex/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite (Vertex AI)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333,"input_audio":0.3},"canonical_model_id":"google/gemini-3.5-flash-lite"},"vertex/gemini-nano-banana-2.1":{"id":"vertex/gemini-nano-banana-2.1","type":"image","name":"Nano Banana 2.1 (Vertex AI)","description":"Image model for prompt-driven generation, multi-turn editing, and visual design workflows","family":"gemini-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"google/gemini-nano-banana-2.1"},"vertex/gemini-3.1-pro-preview":{"id":"vertex/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview (Vertex AI)","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"input_audio":2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":0.25}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"vertex/gemini-3.6-flash":{"id":"vertex/gemini-3.6-flash","name":"Gemini 3.6 Flash (Vertex AI)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"vertex/gemini-3.7-flash":{"id":"vertex/gemini-3.7-flash","name":"Gemini 3.7 Flash (Vertex AI)","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.7-flash"},"vertex/gemini-3.7-flash@eu":{"id":"vertex/gemini-3.7-flash@eu","name":"Gemini 3.7 Flash (Vertex AI, EU)","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.7-flash"},"vertex/gemini-3.5-flash@eu":{"id":"vertex/gemini-3.5-flash@eu","name":"Gemini 3.5 Flash (Vertex AI, EU)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"reasoning":9,"cache_read":0.15,"cache_write":0.083333,"input_audio":3},"canonical_model_id":"google/gemini-3.5-flash"},"vertex/gemini-3.5-flash":{"id":"vertex/gemini-3.5-flash","name":"Gemini 3.5 Flash (Vertex AI)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"reasoning":9,"cache_read":0.15,"cache_write":0.083333,"input_audio":3},"canonical_model_id":"google/gemini-3.5-flash"},"vertex/gemini-3.6-flash@eu":{"id":"vertex/gemini-3.6-flash@eu","name":"Gemini 3.6 Flash (Vertex AI, EU)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"vertex/gemini-3.1-flash-image":{"id":"vertex/gemini-3.1-flash-image","type":"image","name":"Nano Banana 2 (Vertex AI)","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["image","text"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image"},"vertex/gemini-3-flash-preview":{"id":"vertex/gemini-3-flash-preview","name":"Gemini 3 Flash Preview (Vertex AI)","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"reasoning":3,"cache_read":0.05,"cache_write":0.083333,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"vertex/gemini-3.6-flash@us":{"id":"vertex/gemini-3.6-flash@us","name":"Gemini 3.6 Flash (Vertex AI, US)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"vertex/gemini-3.5-flash-lite@us":{"id":"vertex/gemini-3.5-flash-lite@us","name":"Gemini 3.5 Flash Lite (Vertex AI, US)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333,"input_audio":0.3},"canonical_model_id":"google/gemini-3.5-flash-lite"},"vertex/gemini-3-pro-image":{"id":"vertex/gemini-3-pro-image","type":"image","name":"Nano Banana Pro (Vertex AI)","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"input_audio":2},"canonical_model_id":"google/gemini-3-pro-image"},"vertex/gemini-3.5-flash@us":{"id":"vertex/gemini-3.5-flash@us","name":"Gemini 3.5 Flash (Vertex AI, US)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"reasoning":9,"cache_read":0.15,"cache_write":0.083333,"input_audio":3},"canonical_model_id":"google/gemini-3.5-flash"},"vertex/gemini-3.7-flash@us":{"id":"vertex/gemini-3.7-flash@us","name":"Gemini 3.7 Flash (Vertex AI, US)","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.7-flash"},"deepinfra/nemotron-3-super-120b-a12b":{"id":"deepinfra/nemotron-3-super-120b-a12b","name":"Nemotron 3 Super 120B A12B (Deep Infra)","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.085,"output":0.4},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"deepinfra/nemotron-3-ultra-550b-a55b":{"id":"deepinfra/nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B (Deep Infra)","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"cost":{"input":0.5,"output":2.2,"cache_read":0.1},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"deepinfra/moonshotai/Kimi-K2.5":{"id":"deepinfra/moonshotai/Kimi-K2.5","name":"Kimi K2.5 (Deep Infra)","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.45,"output":2.25,"cache_read":0.07},"canonical_model_id":"moonshotai/kimi-k2.5"},"deepinfra/moonshotai/Kimi-K2-Instruct":{"id":"deepinfra/moonshotai/Kimi-K2-Instruct","name":"Kimi K2 Instruct (Deep Infra)","description":"Non-thinking Kimi K2 MoE instruction model for coding and tool-using agents","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":2},"canonical_model_id":"moonshotai/kimi-k2-instruct"},"deepinfra/mistralai/Mistral-Small-24B-Instruct-2501":{"id":"deepinfra/mistralai/Mistral-Small-24B-Instruct-2501","name":"Mistral Small 3 (Deep Infra)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-01-30","last_updated":"2025-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.05,"output":0.08},"canonical_model_id":"mistral/mistral-small-24b-instruct-2501"},"deepinfra/mistralai/Mixtral-8x7B-Instruct-v0.1":{"id":"deepinfra/mistralai/Mixtral-8x7B-Instruct-v0.1","name":"Mixtral 8x7B Instruct v0.1 (Deep Infra)","description":"Open sparse mixture-of-experts Mistral instruction model","family":"mistral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2023-12-11","last_updated":"2023-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.4,"output":0.4},"canonical_model_id":"mistral/mixtral-8x7b-instruct-v0.1"},"deepinfra/NousResearch/Hermes-3-Llama-3.1-405B":{"id":"deepinfra/NousResearch/Hermes-3-Llama-3.1-405B","name":"Hermes 3 405B Instruct (Deep Infra)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-16","last_updated":"2024-08-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":1,"output":1},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-405b"},"deepinfra/NousResearch/Hermes-3-Llama-3.1-70B":{"id":"deepinfra/NousResearch/Hermes-3-Llama-3.1-70B","name":"Hermes 3 70B Instruct (Deep Infra)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-18","last_updated":"2024-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.7,"output":0.7},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-70b"},"deepinfra/thinkingmachines/Inkling":{"id":"deepinfra/thinkingmachines/Inkling","name":"Inkling (Deep Infra)","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":1048576},"cost":{"input":0.95,"output":4.05,"cache_read":0.16},"canonical_model_id":"thinkingmachines/inkling"},"deepinfra/thinkingmachines/Inkling-Small":{"id":"deepinfra/thinkingmachines/Inkling-Small","name":"Inkling Small (Deep Infra)","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":1048576},"cost":{"input":0.45,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"deepinfra/openai/gpt-oss-20b":{"id":"deepinfra/openai/gpt-oss-20b","name":"GPT OSS 20B (Deep Infra)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.03,"output":0.14},"canonical_model_id":"openai/gpt-oss-20b"},"deepinfra/openai/gpt-oss-120b":{"id":"deepinfra/openai/gpt-oss-120b","name":"GPT OSS 120B (Deep Infra)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.037,"output":0.17},"canonical_model_id":"openai/gpt-oss-120b"},"deepinfra/Gryphe/MythoMax-L2-13b":{"id":"deepinfra/Gryphe/MythoMax-L2-13b","name":"MythoMax L2 13B (Deep Infra)","description":"Llama 2 merge combining MythoLogic and Huginn for creative writing and roleplay","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2023-08-10","last_updated":"2023-08-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":3200},"cost":{"input":0.4,"output":0.4},"canonical_model_id":"gryphe/mythomax-l2-13b"},"deepinfra/tencent/Hy4-preview":{"id":"deepinfra/tencent/Hy4-preview","name":"Hy4 preview (Deep Infra)","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"deepinfra/tencent/Hy3":{"id":"deepinfra/tencent/Hy3","name":"Hy3 (Deep Infra)","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"cost":{"input":0.13,"output":0.53,"cache_read":0.033},"canonical_model_id":"tencent/hy3"},"deepinfra/ibm-granite/granite-4.2-8b":{"id":"deepinfra/ibm-granite/granite-4.2-8b","name":"Granite 4.2 8B (Deep Infra)","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"granite","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-25","last_updated":"2026-08-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.06,"output":0.25,"cache_read":0.015},"canonical_model_id":"ibm/granite-4.2-8b"},"deepinfra/meta-models/Muse-Glimmer-30B":{"id":"deepinfra/meta-models/Muse-Glimmer-30B","name":"Muse Glimmer 30B (Deep Infra)","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"deepinfra/meta-llama/Llama-3.2-3B-Instruct":{"id":"deepinfra/meta-llama/Llama-3.2-3B-Instruct","name":"Llama 3.2 3B Instruct (Deep Infra)","description":"Small instruction-tuned Llama model for multilingual dialogue and agentic retrieval","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.02,"output":0.02},"canonical_model_id":"meta/llama-3.2-3b-instruct"},"deepinfra/meta-llama/Llama-3.2-11B-Vision-Instruct":{"id":"deepinfra/meta-llama/Llama-3.2-11B-Vision-Instruct","name":"Llama-3.2-11B-Vision-Instruct (Deep Infra)","description":"Open multimodal Llama model for image understanding, captioning, and visual QA","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.345,"output":0.345},"canonical_model_id":"meta/llama-3.2-11b-vision-instruct"},"deepinfra/meta-llama/Llama-3.3-70B-Instruct":{"id":"deepinfra/meta-llama/Llama-3.3-70B-Instruct","name":"Llama-3.3-70B-Instruct (Deep Infra)","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.1,"output":0.32},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"deepinfra/meta-llama/Llama-Guard-4-12B":{"id":"deepinfra/meta-llama/Llama-Guard-4-12B","name":"Llama Guard 4 12B (Deep Infra)","description":"Multimodal Llama 4-based safety classifier for moderating prompts and responses","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.18,"output":0.18},"canonical_model_id":"meta/llama-guard-4-12b"},"deepinfra/meta-llama/Llama-Guard-3-8B":{"id":"deepinfra/meta-llama/Llama-Guard-3-8B","name":"Llama-Guard-3-8B (Deep Infra)","description":"Llama 3.1-based safety classifier for moderating prompts and model responses","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.055,"output":0.055},"canonical_model_id":"meta/llama-guard-3-8b"},"deepinfra/zai-org/GLM-4.7-Flash":{"id":"deepinfra/zai-org/GLM-4.7-Flash","name":"GLM-4.7-Flash (Deep Infra)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.06,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flash"},"deepinfra/stepfun-ai/Step-3.5-Flash":{"id":"deepinfra/stepfun-ai/Step-3.5-Flash","name":"Step 3.5 Flash (Deep Infra)","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.09,"output":0.3,"cache_read":0.02},"canonical_model_id":"stepfun/step-3.5-flash"},"deepinfra/stepfun-ai/Step-3.7-Flash":{"id":"deepinfra/stepfun-ai/Step-3.7-Flash","name":"Step 3.7 Flash (Deep Infra)","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"deepinfra/stepfun-ai/Step-5-Preview":{"id":"deepinfra/stepfun-ai/Step-5-Preview","name":"Step 5 Preview (Deep Infra)","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1000000},"cost":{"input":1,"output":2.7,"cache_read":0.1},"canonical_model_id":"stepfun/step-5-preview"},"deepinfra/microsoft/WizardLM-2-8x22B":{"id":"deepinfra/microsoft/WizardLM-2-8x22B","name":"WizardLM 2 8x22B (Deep Infra)","description":"Microsoft WizardLM instruction-tuned Mixtral MoE for multilingual dialogue and complex tasks","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-04-15","last_updated":"2024-04-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":8000},"cost":{"input":0.48,"output":0.48},"canonical_model_id":"microsoft/wizardlm-2-8x22b"},"deepinfra/nvidia/Nemotron-3-Nano-30B-A3B":{"id":"deepinfra/nvidia/Nemotron-3-Nano-30B-A3B","name":"Nemotron 3 Nano 30B A3B (Deep Infra)","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.05,"output":0.2,"cache_read":0.025},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"deepinfra/nvidia/Llama-3.1-Nemotron-70B-Instruct":{"id":"deepinfra/nvidia/Llama-3.1-Nemotron-70B-Instruct","name":"Llama 3.1 Nemotron 70B Instruct (Deep Infra)","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.6,"output":0.6},"canonical_model_id":"nvidia/llama-3.1-nemotron-70b-instruct"},"deepinfra/deepseek-ai/DeepSeek-V4.1-Flash-1017":{"id":"deepinfra/deepseek-ai/DeepSeek-V4.1-Flash-1017","name":"DeepSeek V4.1 Flash 1017 (Deep Infra)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.2,"output":0.6,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepinfra/deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepinfra/deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731 (Deep Infra)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.06,"output":0.18,"cache_read":0.015},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepinfra/deepseek-ai/DeepSeek-R1-0528":{"id":"deepinfra/deepseek-ai/DeepSeek-R1-0528","name":"DeepSeek R1 0528 (Deep Infra)","description":"May 2025 DeepSeek R1 update with improved mathematics, coding, and reasoning","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.5,"output":2.15,"cache_read":0.35},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepinfra/deepseek-ai/DeepSeek-R1":{"id":"deepinfra/deepseek-ai/DeepSeek-R1","name":"DeepSeek-R1 (Deep Infra)","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.7,"output":2.4},"canonical_model_id":"deepseek/deepseek-r1"},"deepinfra/deepseek-ai/DeepSeek-V3":{"id":"deepinfra/deepseek-ai/DeepSeek-V3","name":"DeepSeek-V3 (Deep Infra)","description":"Open DeepSeek MoE chat model for coding, math, and general reasoning","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":8192},"cost":{"input":0.32,"output":0.89},"canonical_model_id":"deepseek/deepseek-v3"},"deepinfra/deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepinfra/deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813 (Deep Infra)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.3,"output":2.6,"cache_read":0.1},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepinfra/deepseek-ai/DeepSeek-V4.1-Flash":{"id":"deepinfra/deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash (Deep Infra)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.2,"output":0.6,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepinfra/deepseek-ai/DeepSeek-V3-0324":{"id":"deepinfra/deepseek-ai/DeepSeek-V3-0324","name":"DeepSeek V3 0324 (Deep Infra)","description":"March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.24,"output":0.9,"cache_read":0.135},"canonical_model_id":"deepseek/deepseek-v3-0324"},"deepinfra/Qwen/Qwen3-235B-A22B-Thinking-2507":{"id":"deepinfra/Qwen/Qwen3-235B-A22B-Thinking-2507","name":"Qwen3 235B-A22B Thinking 2507 (Deep Infra)","description":"Updated large open Qwen3 MoE thinking model for reasoning, math, coding, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":81920},"cost":{"input":0.23,"output":2.3,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"deepinfra/Qwen/Qwen3-Next-80B-A3B-Instruct":{"id":"deepinfra/Qwen/Qwen3-Next-80B-A3B-Instruct","name":"Qwen3-Next 80B-A3B Instruct (Deep Infra)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.09,"output":1.1},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"deepinfra/Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"deepinfra/Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B-A22B Instruct 2507 (Deep Infra)","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.09,"output":0.55},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"deepinfra/Qwen/Qwen3-VL-30B-A3B-Instruct":{"id":"deepinfra/Qwen/Qwen3-VL-30B-A3B-Instruct","name":"Qwen3 VL 30B A3B Instruct (Deep Infra)","description":"Open vision-language MoE model for image and video understanding, documents, and visual agents","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"deepinfra/Qwen/Qwen3-VL-235B-A22B-Instruct":{"id":"deepinfra/Qwen/Qwen3-VL-235B-A22B-Instruct","name":"Qwen3 VL 235B A22B Instruct (Deep Infra)","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":0.88,"cache_read":0.11},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"deepinfra/Qwen/Qwen3-Coder-480B-A35B-Instruct":{"id":"deepinfra/Qwen/Qwen3-Coder-480B-A35B-Instruct","name":"Qwen3-Coder 480B-A35B Instruct (Deep Infra)","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":1.6},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"deepinfra/Qwen/Qwen3-Max":{"id":"deepinfra/Qwen/Qwen3-Max","name":"Qwen3 Max (Deep Infra)","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"cost":{"input":1.2,"output":6,"cache_read":0.24},"canonical_model_id":"alibaba/qwen3-max"},"deepinfra/Qwen/Qwen2.5-7B-Instruct":{"id":"deepinfra/Qwen/Qwen2.5-7B-Instruct","name":"Qwen2.5 7B Instruct (Deep Infra)","description":"Small open Qwen2.5 instruction model for chat, structured output, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-09-19","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0.04,"output":0.1},"canonical_model_id":"alibaba/qwen2.5-7b-instruct"},"deepinfra/ByteDance/Seed-2.0-mini":{"id":"deepinfra/ByteDance/Seed-2.0-mini","name":"Seed 2.0 Mini (Deep Infra)","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.1,"output":0.4,"cache_read":0.02},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"deepinfra/ByteDance/Seed-2.0-code":{"id":"deepinfra/ByteDance/Seed-2.0-code","name":"Seed 2.0 Code (Deep Infra)","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":131072},"cost":{"input":0.5,"output":3,"cache_read":0.1},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"deepinfra/google/gemma-3-27b-it":{"id":"deepinfra/google/gemma-3-27b-it","name":"Gemma 3 27B IT (Deep Infra)","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.08,"output":0.16},"canonical_model_id":"google/gemma-3-27b-it"},"deepinfra/google/gemma-3-4b-it":{"id":"deepinfra/google/gemma-3-4b-it","name":"Gemma 3 4B IT (Deep Infra)","description":"Open multimodal Gemma instruction model for efficient text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.1},"canonical_model_id":"google/gemma-3-4b-it"},"deepinfra/google/gemma-3-12b-it":{"id":"deepinfra/google/gemma-3-12b-it","name":"Gemma 3 12B IT (Deep Infra)","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.15},"canonical_model_id":"google/gemma-3-12b-it"},"infomaniak/mistralai/Ministral-3-14B-Instruct-2512":{"id":"infomaniak/mistralai/Ministral-3-14B-Instruct-2512","name":"Ministral 3 14B (Infomaniak)","description":"Open vision-language model for efficient local deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":100000,"output":262144},"cost":{"input":0.33702,"output":0.44936},"canonical_model_id":"mistral/ministral-3-14b-instruct-2512"},"amazon/qwen.qwen3-coder-next":{"id":"amazon/qwen.qwen3-coder-next","name":"Qwen3 Coder Next (Amazon Bedrock)","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.5,"output":1.2},"canonical_model_id":"alibaba/qwen3-coder-next"},"amazon/mistral.voxtral-small-24b-2507@us":{"id":"amazon/mistral.voxtral-small-24b-2507@us","name":"Voxtral Small 24B 2507 (Amazon Bedrock, US)","description":"Open audio-language model for speech transcription, audio understanding, and voice-driven tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/voxtral-small-24b-2507"},"amazon/moonshot.kimi-k2-thinking":{"id":"amazon/moonshot.kimi-k2-thinking","name":"Kimi K2 Thinking (Amazon Bedrock)","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":262144},"cost":{"input":0.6,"output":2.5},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"amazon/amazon.nova-pro-v1:0@us":{"id":"amazon/amazon.nova-pro-v1:0@us","name":"Nova Pro (US)","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.8,"output":3.2,"cache_read":0.2},"canonical_model_id":"amazon/nova-pro"},"amazon/zai.glm-4.7-flash@us":{"id":"amazon/zai.glm-4.7-flash@us","name":"GLM-4.7-Flash (Amazon Bedrock, US)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131072},"cost":{"input":0.07,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"amazon/mistral.pixtral-large-2502-v1:0@us":{"id":"amazon/mistral.pixtral-large-2502-v1:0@us","name":"Pixtral Large (25.02) (Amazon Bedrock, US)","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-08","last_updated":"2025-04-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/pixtral-large-2502"},"amazon/mistral.voxtral-mini-3b-2507":{"id":"amazon/mistral.voxtral-mini-3b-2507","name":"Voxtral Mini 3B 2507 (Amazon Bedrock)","description":"Open audio-language model for speech transcription, audio understanding, and voice-driven tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.04,"output":0.04},"canonical_model_id":"mistral/voxtral-mini-3b-2507"},"amazon/amazon.nova-pro-v1:0":{"id":"amazon/amazon.nova-pro-v1:0","name":"Nova Pro","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.8,"output":3.2,"cache_read":0.2},"canonical_model_id":"amazon/nova-pro"},"amazon/google.gemma-3-27b-it":{"id":"amazon/google.gemma-3-27b-it","name":"Gemma 3 27B IT (Amazon Bedrock)","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.23,"output":0.38},"canonical_model_id":"google/gemma-3-27b-it"},"amazon/zai.glm-4.7-flash":{"id":"amazon/zai.glm-4.7-flash","name":"GLM-4.7-Flash (Amazon Bedrock)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131072},"cost":{"input":0.07,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"amazon/mistral.voxtral-small-24b-2507":{"id":"amazon/mistral.voxtral-small-24b-2507","name":"Voxtral Small 24B 2507 (Amazon Bedrock)","description":"Open audio-language model for speech transcription, audio understanding, and voice-driven tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/voxtral-small-24b-2507"},"amazon/google.gemma-3-27b-it@us":{"id":"amazon/google.gemma-3-27b-it@us","name":"Gemma 3 27B IT (Amazon Bedrock, US)","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.23,"output":0.38},"canonical_model_id":"google/gemma-3-27b-it"},"amazon/amazon.nova-lite-v1:0@us":{"id":"amazon/amazon.nova-lite-v1:0@us","name":"Nova Lite (US)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.06,"output":0.24,"cache_read":0.015},"canonical_model_id":"amazon/nova-lite"},"amazon/mistral.voxtral-mini-3b-2507@us":{"id":"amazon/mistral.voxtral-mini-3b-2507@us","name":"Voxtral Mini 3B 2507 (Amazon Bedrock, US)","description":"Open audio-language model for speech transcription, audio understanding, and voice-driven tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.04,"output":0.04},"canonical_model_id":"mistral/voxtral-mini-3b-2507"},"amazon/google.gemma-3-12b-it":{"id":"amazon/google.gemma-3-12b-it","name":"Gemma 3 12B IT (Amazon Bedrock)","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.09,"output":0.29},"canonical_model_id":"google/gemma-3-12b-it"},"amazon/google.gemma-3-4b-it@us":{"id":"amazon/google.gemma-3-4b-it@us","name":"Gemma 3 4B IT (Amazon Bedrock, US)","description":"Open multimodal Gemma instruction model for efficient text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.04,"output":0.08},"canonical_model_id":"google/gemma-3-4b-it"},"amazon/google.gemma-3-4b-it":{"id":"amazon/google.gemma-3-4b-it","name":"Gemma 3 4B IT (Amazon Bedrock)","description":"Open multimodal Gemma instruction model for efficient text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.04,"output":0.08},"canonical_model_id":"google/gemma-3-4b-it"},"amazon/moonshotai.kimi-k2.5":{"id":"amazon/moonshotai.kimi-k2.5","name":"Kimi K2.5 (Amazon Bedrock)","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3},"canonical_model_id":"moonshotai/kimi-k2.5"},"amazon/mistral.pixtral-large-2502-v1:0":{"id":"amazon/mistral.pixtral-large-2502-v1:0","name":"Pixtral Large (25.02) (Amazon Bedrock)","description":"Mistral vision-language model for image understanding and multimodal chat","family":"pixtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-08","last_updated":"2025-04-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/pixtral-large-2502"},"amazon/openai.gpt-oss-safeguard-20b@us":{"id":"amazon/openai.gpt-oss-safeguard-20b@us","name":"GPT OSS Safeguard 20B (Amazon Bedrock, US)","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.07,"output":0.2},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"amazon/amazon.nova-micro-v1:0":{"id":"amazon/amazon.nova-micro-v1:0","name":"Nova Micro","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.035,"output":0.14,"cache_read":0.00875},"canonical_model_id":"amazon/nova-micro"},"amazon/google.gemma-3-12b-it@us":{"id":"amazon/google.gemma-3-12b-it@us","name":"Gemma 3 12B IT (Amazon Bedrock, US)","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.09,"output":0.29},"canonical_model_id":"google/gemma-3-12b-it"},"amazon/openai.gpt-oss-safeguard-20b":{"id":"amazon/openai.gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B (Amazon Bedrock)","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":131072},"cost":{"input":0.07,"output":0.2},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"amazon/amazon.nova-micro-v1:0@us":{"id":"amazon/amazon.nova-micro-v1:0@us","name":"Nova Micro (US)","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":10000},"cost":{"input":0.035,"output":0.14,"cache_read":0.00875},"canonical_model_id":"amazon/nova-micro"},"amazon/amazon.nova-lite-v1:0":{"id":"amazon/amazon.nova-lite-v1:0","name":"Nova Lite","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-12-03","last_updated":"2024-12-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":10000},"cost":{"input":0.06,"output":0.24,"cache_read":0.015},"canonical_model_id":"amazon/nova-lite"},"databricks/databricks-qwen3-next-80b-a3b-instruct@eu":{"id":"databricks/databricks-qwen3-next-80b-a3b-instruct@eu","name":"Qwen3-Next 80B-A3B Instruct (Databricks, EU)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.15001,"output":1.20001,"cache_read":0.015001,"cache_write":0.15001},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"databricks/databricks-gpt-oss-20b":{"id":"databricks/databricks-gpt-oss-20b","name":"GPT OSS 20B (Databricks)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.07,"output":0.30002,"cache_read":0.007,"cache_write":0.07},"canonical_model_id":"openai/gpt-oss-20b"},"databricks/databricks-qwen3-next-80b-a3b-instruct":{"id":"databricks/databricks-qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct (Databricks)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.15001,"output":1.20001,"cache_read":0.015001,"cache_write":0.15001},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"databricks/databricks-deepseek-v4-pro-0813":{"id":"databricks/databricks-deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813 (Databricks)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.32,"output":3.959999,"cache_read":0.132,"cache_write":1.31999},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"databricks/databricks-gpt-oss-120b@eu":{"id":"databricks/databricks-gpt-oss-120b@eu","name":"GPT OSS 120B (Databricks, EU)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15001,"output":0.59997,"cache_read":0.015001,"cache_write":0.15001},"canonical_model_id":"openai/gpt-oss-120b"},"databricks/databricks-inkling":{"id":"databricks/databricks-inkling","name":"Inkling (Databricks)","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":1048576},"cost":{"input":1,"output":4.05,"cache_read":0.1,"cache_write":1.00002},"canonical_model_id":"thinkingmachines/inkling"},"databricks/databricks-deepseek-v4-flash-0731":{"id":"databricks/databricks-deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731 (Databricks)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.014,"cache_write":0.14},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"databricks/databricks-gpt-oss-20b@eu":{"id":"databricks/databricks-gpt-oss-20b@eu","name":"GPT OSS 20B (Databricks, EU)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.07,"output":0.30002,"cache_read":0.007,"cache_write":0.07},"canonical_model_id":"openai/gpt-oss-20b"},"databricks/databricks-gpt-oss-120b":{"id":"databricks/databricks-gpt-oss-120b","name":"GPT OSS 120B (Databricks)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15001,"output":0.59997,"cache_read":0.015001,"cache_write":0.15001},"canonical_model_id":"openai/gpt-oss-120b"},"ionos/openai/gpt-oss-120b":{"id":"ionos/openai/gpt-oss-120b","name":"GPT OSS 120B (IONOS)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.16851,"output":0.73021},"canonical_model_id":"openai/gpt-oss-120b"},"ionos/meta-llama/Llama-3.3-70B-Instruct":{"id":"ionos/meta-llama/Llama-3.3-70B-Instruct","name":"Llama-3.3-70B-Instruct (IONOS)","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.73021,"output":0.73021},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"ionos/Qwen/Qwen3-Coder-Next":{"id":"ionos/Qwen/Qwen3-Coder-Next","name":"Qwen3 Coder Next (IONOS)","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":65536},"cost":{"input":0.16851,"output":0.89872},"canonical_model_id":"alibaba/qwen3-coder-next"},"perplexityai/sonar-reasoning-pro":{"id":"perplexityai/sonar-reasoning-pro","name":"Sonar Reasoning Pro","description":"Web-grounded Sonar for multi-step research questions that need cited reasoning","family":"sonar-reasoning","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":2,"output":8},"canonical_model_id":"perplexity/sonar-reasoning-pro"},"perplexityai/sonar-deep-research":{"id":"perplexityai/sonar-deep-research","name":"Sonar Deep Research","description":"Sonar search model for autonomous research and citation-backed long-form reports","family":"sonar","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-02-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":2,"output":8,"reasoning":3},"canonical_model_id":"perplexity/sonar-deep-research"},"perplexityai/sonar-pro":{"id":"perplexityai/sonar-pro","name":"Sonar Pro","description":"Deeper Sonar search model with broader retrieval and stronger synthesis","family":"sonar-pro","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":3,"output":15},"canonical_model_id":"perplexity/sonar-pro"},"perplexityai/sonar":{"id":"perplexityai/sonar","name":"Sonar","description":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","family":"sonar","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":127072,"output":4096},"cost":{"input":1,"output":1},"canonical_model_id":"perplexity/sonar"},"minimax/MiniMax-M2":{"id":"minimax/MiniMax-M2","name":"MiniMax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/MiniMax-M3":{"id":"minimax/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/MiniMax-M2.7":{"id":"minimax/MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/MiniMax-M3.1-Flash":{"id":"minimax/MiniMax-M3.1-Flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":512000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"minimax/MiniMax-M2.1":{"id":"minimax/MiniMax-M2.1","name":"MiniMax-M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/MiniMax-M2.5":{"id":"minimax/MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"nebius/openai/gpt-oss-120b":{"id":"nebius/openai/gpt-oss-120b","name":"GPT OSS 120B (Nebius)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6,"cache_read":0.15},"canonical_model_id":"openai/gpt-oss-120b"},"nebius/meta-models/muse-spark-open":{"id":"nebius/meta-models/muse-spark-open","name":"Muse Spark Open (Nebius)","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":1.25},"canonical_model_id":"meta/muse-spark-open"},"nebius/nvidia/nemotron-3-super-120b-a12b":{"id":"nebius/nvidia/nemotron-3-super-120b-a12b","name":"Nemotron 3 Super 120B A12B (Nebius)","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":0.9,"cache_read":0.3},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nebius/nvidia/Nemotron-3-Ultra-550b-a55b":{"id":"nebius/nvidia/Nemotron-3-Ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B (Nebius)","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"cost":{"input":1,"output":3,"cache_read":1},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nebius/deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"nebius/deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731 (Nebius)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.14},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"nebius/deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"nebius/deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813 (Nebius)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":979000,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":1.32},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"nebius/deepseek-ai/DeepSeek-V4.1-Flash":{"id":"nebius/deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash (Nebius)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048000,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.3},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"nebius/Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"nebius/Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B-A22B Instruct 2507 (Nebius)","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.2,"output":0.6,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"nebius/Qwen/Qwen3-30B-A3B-Instruct-2507":{"id":"nebius/Qwen/Qwen3-30B-A3B-Instruct-2507","name":"Qwen3 30B-A3B Instruct 2507 (Nebius)","description":"Updated compact Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.1,"output":0.3,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3-30b-a3b-instruct-2507"},"nebius/google/gemma-3-27b-it":{"id":"nebius/google/gemma-3-27b-it","name":"Gemma 3 27B IT (Nebius)","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":110000,"output":131072},"cost":{"input":0.1,"output":0.3,"cache_read":0.1},"canonical_model_id":"google/gemma-3-27b-it"},"xai/grok-4-fast-non-reasoning":{"id":"xai/grok-4-fast-non-reasoning","name":"X-Ai/Grok-4-Fast-Non-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05,"tiers":[{"input":0.4,"output":1,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"xai/grok-4-fast-non-reasoning"},"xai/grok-4.5":{"id":"xai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"xai/grok-4.6":{"id":"xai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"xai/grok-4.20-0309-non-reasoning":{"id":"xai/grok-4.20-0309-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"xai/grok-latest":{"id":"xai/grok-latest","name":"Grok Latest (Grok 4.7)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"xai/grok-4.7":{"id":"xai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"xai/grok-4.3":{"id":"xai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"xai/grok-4.20-0309-reasoning":{"id":"xai/grok-4.20-0309-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"xai/grok-build-0.1":{"id":"xai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}},"canonical_model_id":"xai/grok-build-0.1"},"zai/glm-5.1":{"id":"zai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai/glm-4.7":{"id":"zai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"zai/glm-5-turbo":{"id":"zai/glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"zai/glm-5.3-flash":{"id":"zai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai/glm-5.2":{"id":"zai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai/glm-5":{"id":"zai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"zai/glm-4.6":{"id":"zai/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"zai/glm-5.3":{"id":"zai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"zai/glm-5v-turbo":{"id":"zai/glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"zai/glm-4.6v":{"id":"zai/glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.05},"canonical_model_id":"zhipuai/glm-4.6v"},"greenference/gpt-oss-20b":{"id":"greenference/gpt-oss-20b","name":"GPT OSS 20B (Greenference)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.009,"output":0.045,"cache_read":0.0045},"canonical_model_id":"openai/gpt-oss-20b"},"greenference/deepseek-v4.1-flash":{"id":"greenference/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Greenference)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.22,"output":0.88,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":32768},"cost":{"input":0.5,"output":3,"cache_read":0.1,"cache_write":0.625},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen-max":{"id":"qwen/qwen-max","name":"Qwen Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":1.6,"output":6.4,"cache_read":0.32},"canonical_model_id":"alibaba/qwen-max"},"qwen/qwen3.8-2.4t-a95b":{"id":"qwen/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"qwen/qwen3.8-max-0902":{"id":"qwen/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"qwen/deepseek-v4.1-flash-1017":{"id":"qwen/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (Alibaba)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"qwen/deepseek-v4-pro-0813":{"id":"qwen/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813 (Alibaba)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.066},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"qwen/qwen3.8-max":{"id":"qwen/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen/qwen-mt-plus":{"id":"qwen/qwen-mt-plus","name":"Qwen MT Plus","description":"Flagship Qwen translation model with terminology control and support for 92 languages","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-24","last_updated":"2025-07-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4096,"output":8192},"cost":{"input":2.46,"output":7.37},"canonical_model_id":"alibaba/qwen-mt-plus"},"qwen/qwen3.8-omni-flash":{"id":"qwen/qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/deepseek-v4.1-flash":{"id":"qwen/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Alibaba)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen/deepseek-v4-flash-0731":{"id":"qwen/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731 (Alibaba)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.22,"output":0.66,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"qwen/qwen3-coder-flash":{"id":"qwen/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":1.5,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"alibaba/qwen3-coder-flash"},"cohere/command-r7b-12-2024":{"id":"cohere/command-r7b-12-2024","name":"Command R7B","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-12-13","last_updated":"2024-12-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":132000,"output":4000},"cost":{"input":0.0375,"output":0.15},"canonical_model_id":"cohere/command-r7b-12-2024"},"cohere/command-r-plus-08-2024":{"id":"cohere/command-r-plus-08-2024","name":"Command R+","description":"Cohere's RAG workhorse for long-context enterprise search and tool use","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-r-plus-08-2024"},"cohere/command-r-08-2024":{"id":"cohere/command-r-08-2024","name":"Command R","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"cohere/command-r-08-2024"},"cohere/c4ai-aya-expanse-32b":{"id":"cohere/c4ai-aya-expanse-32b","name":"Aya Expanse 32B","description":"Open multilingual model optimized for generation across 23 languages","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-10-24","last_updated":"2024-10-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"cohere/c4ai-aya-expanse-32b"},"cohere/command-a-03-2025":{"id":"cohere/command-a-03-2025","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":288000,"output":8000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a-03-2025"},"scaleway/gemma-3-27b-it":{"id":"scaleway/gemma-3-27b-it","name":"Gemma 3 27B IT (Scaleway)","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":40000,"output":131072},"cost":{"input":0.287125,"output":0.57425},"canonical_model_id":"google/gemma-3-27b-it"},"scaleway/qwen3-235b-a22b-instruct-2507":{"id":"scaleway/qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B-A22B Instruct 2507 (Scaleway)","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":250000,"output":16384},"cost":{"input":0.84255,"output":2.52765},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"scaleway/gpt-oss-120b":{"id":"scaleway/gpt-oss-120b","name":"GPT OSS 120B (Scaleway)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.16851,"output":0.67404},"canonical_model_id":"openai/gpt-oss-120b"},"scaleway/qwen3-coder-30b-a3b-instruct":{"id":"scaleway/qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct (Scaleway)","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":65536},"cost":{"input":0.2245,"output":0.898},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"scaleway/llama-3.3-70b-instruct":{"id":"scaleway/llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct (Scaleway)","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":1.01106,"output":1.01106},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"scaleway/deepseek-v4.1-flash":{"id":"scaleway/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Scaleway)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":384000},"cost":{"input":0.67404,"output":2.02212,"cache_read":0.134808},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"scaleway/deepseek-v4-flash-0731":{"id":"scaleway/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731 (Scaleway)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":384000},"cost":{"input":0.44936,"output":0.89872,"cache_read":0.089872},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"cloudflare/@cf/meta/llama-3.2-3b-instruct":{"id":"cloudflare/@cf/meta/llama-3.2-3b-instruct","name":"Llama 3.2 3B Instruct (Cloudflare)","description":"Small instruction-tuned Llama model for multilingual dialogue and agentic retrieval","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":80000,"output":131072},"cost":{"input":0.0509,"output":0.335},"canonical_model_id":"meta/llama-3.2-3b-instruct"},"cloudflare/@cf/meta/llama-3.2-1b-instruct":{"id":"cloudflare/@cf/meta/llama-3.2-1b-instruct","name":"Llama 3.2 1B Instruct (Cloudflare)","description":"Compact instruction-tuned Llama model for multilingual dialogue and on-device assistants","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":60000,"output":32000},"cost":{"input":0.027,"output":0.201},"canonical_model_id":"meta/llama-3.2-1b-instruct"},"cloudflare/@cf/meta/llama-guard-3-8b":{"id":"cloudflare/@cf/meta/llama-guard-3-8b","name":"Llama-Guard-3-8B (Cloudflare)","description":"Llama 3.1-based safety classifier for moderating prompts and model responses","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.484,"output":0.03},"canonical_model_id":"meta/llama-guard-3-8b"},"cloudflare/@cf/mistralai/mistral-small-3.1-24b-instruct":{"id":"cloudflare/@cf/mistralai/mistral-small-3.1-24b-instruct","name":"Mistral Small 3.1 24B (Cloudflare)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":102400},"cost":{"input":0.351,"output":0.555},"canonical_model_id":"mistral/mistral-small-3.1-24b-instruct"},"cloudflare/@cf/openai/gpt-oss-20b":{"id":"cloudflare/@cf/openai/gpt-oss-20b","name":"GPT OSS 20B (Cloudflare)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.2,"output":0.3},"canonical_model_id":"openai/gpt-oss-20b"},"cloudflare/@cf/openai/gpt-oss-120b":{"id":"cloudflare/@cf/openai/gpt-oss-120b","name":"GPT OSS 120B (Cloudflare)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.35,"output":0.75},"canonical_model_id":"openai/gpt-oss-120b"},"cloudflare/@cf/zai-org/glm-4.7-flash":{"id":"cloudflare/@cf/zai-org/glm-4.7-flash","name":"GLM-4.7-Flash (Cloudflare)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.0605,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"cloudflare/@cf/aisingapore/gemma-sea-lion-v4-27b-it":{"id":"cloudflare/@cf/aisingapore/gemma-sea-lion-v4-27b-it","name":"Gemma-SEA-LION-v4-27B-IT (Cloudflare)","description":"Gemma 3 27B tuned by AI Singapore for Southeast Asian languages and instruction following","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.351,"output":0.555},"canonical_model_id":"aisingapore/gemma-sea-lion-v4-27b-it"},"cloudflare/@cf/deepseek-ai/deepseek-v4.1-flash-1017":{"id":"cloudflare/@cf/deepseek-ai/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (Cloudflare)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"cloudflare/@cf/deepseek-ai/deepseek-v4-pro-0813":{"id":"cloudflare/@cf/deepseek-ai/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813 (Cloudflare)","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"cloudflare/@cf/deepseek-ai/deepseek-v4-flash-0731":{"id":"cloudflare/@cf/deepseek-ai/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731 (Cloudflare)","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"cloudflare/@cf/qwen/qwen2.5-coder-32b-instruct":{"id":"cloudflare/@cf/qwen/qwen2.5-coder-32b-instruct","name":"Qwen2.5-Coder-32B-Instruct (Cloudflare)","description":"Open coding-focused Qwen model for code generation, repair, and repository reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0.66,"output":1},"canonical_model_id":"alibaba/qwen2.5-coder-32b-instruct"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.083333,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":4,"output":20,"reasoning":20,"cache_read":0.2,"cache_write":0.375,"input_audio":4},"canonical_model_id":"google/gemini-4-argon"},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"input_audio":2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemini-3.1-flash-lite-preview":{"id":"google/gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.083333,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"google/gemini-pro-latest":{"id":"google/gemini-pro-latest","name":"Gemini Pro Latest (Gemini 4 Argon)","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":4,"output":20,"reasoning":20,"cache_read":0.2,"cache_write":0.375,"input_audio":4},"canonical_model_id":"google/gemini-4-argon"},"google/gemini-3.1-flash-lite-image":{"id":"google/gemini-3.1-flash-lite-image","type":"image","name":"Nano Banana 2 Lite","description":"Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":4096},"cost":{"input":0.25,"output":1.5},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"google/gemini-3-pro-image-preview":{"id":"google/gemini-3-pro-image-preview","type":"image","name":"Nano Banana Pro Preview","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"input_audio":2},"canonical_model_id":"google/gemini-3-pro-image-preview"},"google/gemini-2.5-flash-image":{"id":"google/gemini-2.5-flash-image","type":"image","name":"Nano Banana","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["audio","image","text","video"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash-image"},"google/gemini-flash-latest":{"id":"google/gemini-flash-latest","name":"Gemini Flash Latest (Gemini 3.8 Flash)","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.8-flash"},"google/deep-research-max-preview-04-2026":{"id":"google/deep-research-max-preview-04-2026","name":"Deep Research Max Preview","description":"Maximum-comprehensiveness agentic researcher for multi-step investigation, synthesis, and cited reports","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12},"canonical_model_id":"google/deep-research-max-preview-04-2026"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333,"input_audio":0.3},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-3.1-flash-image-preview":{"id":"google/gemini-3.1-flash-image-preview","type":"image","name":"Nano Banana 2 Preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["image","text"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":65536},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image-preview"},"google/gemini-nano-banana-2.1":{"id":"google/gemini-nano-banana-2.1","type":"image","name":"Nano Banana 2.1","description":"Image model for prompt-driven generation, multi-turn editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["image","text"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"google/gemini-nano-banana-2.1"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"input_audio":2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"reasoning":7.5,"cache_read":0.15,"cache_write":0.083333,"input_audio":1.5},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"reasoning":9,"cache_read":0.15,"cache_write":0.083333,"input_audio":3},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-3.1-flash-image":{"id":"google/gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["image","text"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"reasoning":3,"cache_read":0.05,"cache_write":0.083333,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-robotics-er-2-preview":{"id":"google/gemini-robotics-er-2-preview","name":"Gemini Robotics-ER 2 Preview","description":"Embodied reasoning model for robotics: spatial understanding, video progress, and multi-robot orchestration","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","pdf","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":1,"output":5,"reasoning":5,"cache_read":0.1,"input_audio":1},"canonical_model_id":"google/gemini-robotics-er-2-preview"},"google/deep-research-preview-04-2026":{"id":"google/deep-research-preview-04-2026","name":"Gemini Deep Research Preview","description":"Agentic model for autonomous multi-step research, synthesis, and cited reports","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12},"canonical_model_id":"google/deep-research-preview-04-2026"},"google/gemini-3-pro-image":{"id":"google/gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"input_audio":2},"canonical_model_id":"google/gemini-3-pro-image"}}},"google-vertex-anthropic":{"id":"google-vertex-anthropic","env":["GOOGLE_VERTEX_PROJECT","GOOGLE_VERTEX_LOCATION","GOOGLE_APPLICATION_CREDENTIALS"],"npm":"@ai-sdk/google-vertex/anthropic","name":"Vertex (Anthropic)","doc":"https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude","models":{"claude-opus-4@20250514":{"id":"claude-opus-4@20250514","name":"Claude Opus 4","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-20250514"},"claude-opus-5-5@default":{"id":"claude-opus-5-5@default","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"claude-fable-5@default":{"id":"claude-fable-5@default","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"claude-fable-5-1@default":{"id":"claude-fable-5-1@default","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"claude-sonnet-5@default":{"id":"claude-sonnet-5@default","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"claude-opus-5@default":{"id":"claude-opus-5@default","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"claude-haiku-4-5@20251001":{"id":"claude-haiku-4-5@20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"claude-haiku-5-5@default":{"id":"claude-haiku-5-5@default","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-opus-4-6@default":{"id":"claude-opus-4-6@default","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"claude-opus-4-8@default":{"id":"claude-opus-4-8@default","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"claude-opus-4-5@20251101":{"id":"claude-opus-4-5@20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"claude-sonnet-4@20250514":{"id":"claude-sonnet-4@20250514","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"claude-opus-4-1@20250805":{"id":"claude-opus-4-1@20250805","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"claude-sonnet-4-6@default":{"id":"claude-sonnet-4-6@default","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"claude-sonnet-4-5@20250929":{"id":"claude-sonnet-4-5@20250929","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"claude-sonnet-5-5@default":{"id":"claude-sonnet-5-5@default","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"claude-opus-4-7@default":{"id":"claude-opus-4-7@default","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"}}},"qihang-ai":{"id":"qihang-ai","env":["QIHANG_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.qhaigc.net/v1","name":"QiHang","doc":"https://www.qhaigc.net/docs","models":{"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.09,"output":0.71,"tiers":[{"input":0.09,"output":0.71,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":0.09,"output":0.71}},"canonical_model_id":"google/gemini-2.5-flash"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5-Mini","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":0.04,"output":0.29},"canonical_model_id":"openai/gpt-5-mini"},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":0.14,"output":1.14},"canonical_model_id":"openai/gpt-5.2-codex"},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":0.25,"output":2},"canonical_model_id":"openai/gpt-5.2"},"claude-opus-4-5-20251101":{"id":"claude-opus-4-5-20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":0.71,"output":3.57},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"claude-sonnet-4-5-20250929":{"id":"claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":0.43,"output":2.14},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":0.14,"output":0.71},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.07,"output":0.43,"tiers":[{"input":0.07,"output":0.43,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":0.07,"output":0.43}},"canonical_model_id":"google/gemini-3-flash-preview"},"gemini-3-pro-preview":{"id":"gemini-3-pro-preview","name":"Gemini 3 Pro Preview","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65000},"cost":{"input":0.57,"output":3.43},"canonical_model_id":"google/gemini-3-pro-preview"}}},"minimax":{"id":"minimax","env":["MINIMAX_API_KEY"],"npm":"@ai-sdk/anthropic","api":"https://api.minimax.io/anthropic/v1","name":"MiniMax (minimax.io)","doc":"https://platform.minimax.io/docs/guides/quickstart","models":{"MiniMax-M2":{"id":"MiniMax-M2","name":"MiniMax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3,"output":1.2}},"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":524288},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"tiers":[{"input":0.6,"output":2.4,"cache_read":0.12,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":0.6,"output":2.4,"cache_read":0.12}}},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375}},"MiniMax-M2.5-highspeed":{"id":"MiniMax-M2.5-highspeed","name":"MiniMax-M2.5-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.6,"output":2.4,"cache_read":0.03,"cache_write":0.375}},"MiniMax-M3.1-Flash":{"id":"MiniMax-M3.1-Flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"MiniMax-M2.1":{"id":"MiniMax-M2.1","name":"MiniMax-M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375}},"MiniMax-M2.7-highspeed":{"id":"MiniMax-M2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375}}}},"jalapeno":{"id":"jalapeno","env":["JALAPENO_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.jalapeno-cloud.ai/v1","name":"Jalapeno Cloud","doc":"https://www.jalapeno-cloud.ai/docs/","models":{"Kimi-K3":{"id":"Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"Kimi-K2.7-Code":{"id":"Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":271360,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"Qwen3-Next-80B-A3B-Thinking":{"id":"Qwen3-Next-80B-A3B-Thinking","name":"Qwen3-Next 80B-A3B (Thinking)","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.15,"output":1.5},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"Qwen3.5-27B":{"id":"Qwen3.5-27B","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":2.4},"canonical_model_id":"alibaba/qwen3.5-27b"},"Qwen3-Next-80B-A3B-Instruct":{"id":"Qwen3-Next-80B-A3B-Instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":129024,"output":32768},"status":"deprecated","cost":{"input":0.15,"output":1.5},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"Kimi-K2.5":{"id":"Kimi-K2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":180224},"cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"DeepSeek-V4-Pro":{"id":"DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":1.6,"output":3.38,"cache_read":0.14},"canonical_model_id":"deepseek/deepseek-v4-pro"},"Qwen3.5-35B-A3B":{"id":"Qwen3.5-35B-A3B","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"GLM-5.2":{"id":"GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"Qwen3-VL-235B-A22B-Instruct":{"id":"Qwen3-VL-235B-A22B-Instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":129024,"output":32768},"status":"deprecated","cost":{"input":0.3,"output":1.5},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"Qwen3.5-397B-A17B":{"id":"Qwen3.5-397B-A17B","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"Hy3":{"id":"Hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"input":192000,"output":128000},"status":"deprecated","cost":{"input":0.14,"output":0.58},"canonical_model_id":"tencent/hy3"},"DeepSeek-V4-Flash":{"id":"DeepSeek-V4-Flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"Qwen3.5-122B-A10B":{"id":"Qwen3.5-122B-A10B","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":3.2},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"Qwen3-VL-235B-A22B-Thinking":{"id":"Qwen3-VL-235B-A22B-Thinking","name":"Qwen3 VL 235B A22B Thinking","description":"Qwen vision-language thinking model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.98,"output":3.95},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-thinking"},"GLM-5.1":{"id":"GLM-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1.38,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"}}},"openreason":{"id":"openreason","env":["OPENREASON_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.openreason.app/v1","name":"OpenReason","doc":"https://openreason.app/docs","models":{"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.0022,"output":4.22},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.1055,"output":0.422},"canonical_model_id":"openai/gpt-oss-120b"},"deepseek-ai/deepseek-v4-flash-0731":{"id":"deepseek-ai/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.1371,"output":0.2743},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"}}},"umans-ai":{"id":"umans-ai","env":["UMANS_AI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.code.umans.ai/v1","name":"Umans AI","doc":"https://app.umans.ai/offers/code/docs/orgs","models":{"umans-coder":{"id":"umans-coder","name":"Umans Coder","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131071},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"umans-deepseek-v4-flash-0731":{"id":"umans-deepseek-v4-flash-0731","name":"DeepSeek V4 Flash","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393215},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"umans-glm-5.3":{"id":"umans-glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131071},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"umans-flash":{"id":"umans-flash","name":"Umans Flash","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.15,"output":1,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"umans-glm-5.3-flash":{"id":"umans-glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131071},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"umans-deepseek-v4.1-flash":{"id":"umans-deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393215},"cost":{"input":0.15,"output":0.6,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"umans-kimi-k3":{"id":"umans-kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"umans-deepseek-v4.1-flash-1017":{"id":"umans-deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393215},"cost":{"input":0.15,"output":0.6,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"}}},"xiaomi-token-plan-cn":{"id":"xiaomi-token-plan-cn","env":["XIAOMI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://token-plan-cn.xiaomimimo.com/v1","name":"Xiaomi Token Plan (China)","doc":"https://platform.xiaomimimo.com/#/docs","models":{"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5"},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"mimo-v2-tts":{"id":"mimo-v2-tts","type":"speech","name":"MiMo-V2-TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":8192},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2-tts"},"mimo-v2.5-tts-voicedesign":{"id":"mimo-v2.5-tts-voicedesign","type":"speech","name":"MiMo-V2.5-TTS-VoiceDesign","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.5-tts-voicedesign"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"mimo-v2.5-tts":{"id":"mimo-v2.5-tts","type":"speech","name":"MiMo-V2.5-TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.5-tts"},"mimo-v2-pro":{"id":"mimo-v2-pro","name":"MiMo-V2-Pro","description":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2-pro"},"mimo-v2.5-tts-voiceclone":{"id":"mimo-v2.5-tts-voiceclone","type":"speech","name":"MiMo-V2.5-TTS-VoiceClone","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.5-tts-voiceclone"}}},"helicone":{"id":"helicone","env":["HELICONE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://ai-gateway.helicone.ai/v1","name":"Helicone","doc":"https://helicone.ai/models","models":{"claude-3.5-haiku":{"id":"claude-3.5-haiku","name":"Anthropic: Claude 3.5 Haiku","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":0.8,"output":4,"cache_read":0.08,"cache_write":1},"canonical_model_id":"anthropic/claude-3-5-haiku-20241022"},"chatgpt-4o-latest":{"id":"chatgpt-4o-latest","name":"OpenAI ChatGPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2024-08-14","last_updated":"2024-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":5,"output":20,"cache_read":2.5}},"qwen3-30b-a3b":{"id":"qwen3-30b-a3b","name":"Qwen3 30B A3B","description":"Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":41000,"output":41000},"cost":{"input":0.08,"output":0.29},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"deepseek-reasoner":{"id":"deepseek-reasoner","name":"DeepSeek Reasoner","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":64000},"cost":{"input":0.56,"output":1.68,"cache_read":0.07},"canonical_model_id":"deepseek/deepseek-reasoner"},"kimi-k2-thinking":{"id":"kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":262144},"cost":{"input":0.48,"output":2},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Google Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.3,"output":2.5,"cache_read":0.075,"cache_write":0.3},"canonical_model_id":"google/gemini-2.5-flash"},"gpt-5.1":{"id":"gpt-5.1","name":"OpenAI GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"grok-4-fast-non-reasoning":{"id":"grok-4-fast-non-reasoning","name":"xAI Grok 4 Fast Non-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4-fast-non-reasoning"},"llama-3.1-8b-instruct":{"id":"llama-3.1-8b-instruct","name":"Meta Llama 3.1 8B Instruct","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":0.02,"output":0.05},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"OpenAI GPT-4.1 Mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"gpt-5-pro":{"id":"gpt-5-pro","name":"OpenAI: GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":272000,"output":32768},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"o3-mini":{"id":"o3-mini","name":"OpenAI o3 Mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2023-10","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"OpenAI GPT-OSS 20b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.05,"output":0.2},"canonical_model_id":"openai/gpt-oss-20b"},"gpt-5-chat-latest":{"id":"gpt-5-chat-latest","name":"OpenAI GPT-5 Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2024-09","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":272000,"output":16384},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5-chat-latest"},"gpt-5-mini":{"id":"gpt-5-mini","name":"OpenAI GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"o1-mini":{"id":"o1-mini","name":"OpenAI: o1-mini","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":65536},"cost":{"input":1.1,"output":4.4,"cache_read":0.55}},"sonar-reasoning":{"id":"sonar-reasoning","name":"Perplexity Sonar Reasoning","description":"Web-grounded reasoning model for multi-step research and cited answers","family":"sonar-reasoning","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-01-27","last_updated":"2025-01-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":127000,"output":4096},"cost":{"input":1,"output":5}},"mistral-small":{"id":"mistral-small","name":"Mistral Small 3.2","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.075,"output":0.2},"canonical_model_id":"mistral/mistral-small-2506"},"claude-4.5-opus":{"id":"claude-4.5-opus","name":"Anthropic: Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"mistral-large-2411":{"id":"mistral-large-2411","name":"Mistral-Large","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-11-18","last_updated":"2024-11-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":2,"output":6},"canonical_model_id":"mistral/mistral-large-2411"},"gpt-5-nano":{"id":"gpt-5-nano","name":"OpenAI GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"kimi-k2-0905":{"id":"kimi-k2-0905","name":"Kimi K2 (09/05)","description":"September 2025 Kimi K2 instruction update with a 256K context and improved agentic coding","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.5,"output":2,"cache_read":0.4},"canonical_model_id":"moonshotai/kimi-k2-0905"},"claude-opus-4-1":{"id":"claude-opus-4-1","name":"Anthropic: Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"temperature":true,"knowledge":"2025-08","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"claude-opus-4-1-20250805":{"id":"claude-opus-4-1-20250805","name":"Anthropic: Claude Opus 4.1 (20250805)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"temperature":true,"knowledge":"2025-08","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"deepseek-v3.1-terminus":{"id":"deepseek-v3.1-terminus","name":"DeepSeek V3.1 Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-22","last_updated":"2025-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.27,"output":1,"cache_read":0.216}},"claude-4.5-haiku":{"id":"claude-4.5-haiku","name":"Anthropic: Claude 4.5 Haiku","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-10","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"deepseek-v3":{"id":"deepseek-v3","name":"DeepSeek V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.56,"output":1.68,"cache_read":0.07}},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":65536},"cost":{"input":0.27,"output":0.41},"canonical_model_id":"deepseek/deepseek-v3.2"},"gpt-5-codex":{"id":"gpt-5-codex","name":"OpenAI: GPT-5 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5-codex"},"o3":{"id":"o3","name":"OpenAI o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-06","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"grok-4-1-fast-reasoning":{"id":"grok-4-1-fast-reasoning","name":"xAI Grok 4.1 Fast Reasoning","description":"xAI's fast agentic tool-calling model with a 2M context window and built-in reasoning","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast-reasoning"},"kimi-k2-0711":{"id":"kimi-k2-0711","name":"Kimi K2 (07/11)","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.57,"output":2.3},"canonical_model_id":"moonshotai/kimi-k2"},"grok-4-fast-reasoning":{"id":"grok-4-fast-reasoning","name":"xAI: Grok 4 Fast Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4-fast-reasoning"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"OpenAI GPT-OSS 120b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.04,"output":0.16},"canonical_model_id":"openai/gpt-oss-120b"},"llama-guard-4":{"id":"llama-guard-4","name":"Meta Llama Guard 4 12B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":1024},"cost":{"input":0.21,"output":0.21}},"sonar-reasoning-pro":{"id":"sonar-reasoning-pro","name":"Perplexity Sonar Reasoning Pro","description":"Web-grounded Sonar for multi-step research questions that need cited reasoning","family":"sonar-reasoning","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":127000,"output":4096},"cost":{"input":2,"output":8},"canonical_model_id":"perplexity/sonar-reasoning-pro"},"qwen3-235b-a22b-thinking":{"id":"qwen3-235b-a22b-thinking","name":"Qwen3 235B A22B Thinking","description":"Updated large open Qwen3 MoE thinking model for reasoning, math, coding, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":81920},"cost":{"input":0.3,"output":2.9},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"qwen3-coder-30b-a3b-instruct":{"id":"qwen3-coder-30b-a3b-instruct","name":"Qwen3 Coder 30B A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Meta Llama 3.3 70B Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16400},"cost":{"input":0.13,"output":0.39},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"grok-code-fast-1":{"id":"grok-code-fast-1","name":"xAI Grok Code Fast 1","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":10000},"cost":{"input":0.2,"output":1.5,"cache_read":0.02},"canonical_model_id":"xai/grok-code-fast-1"},"claude-opus-4":{"id":"claude-opus-4","name":"Anthropic: Claude Opus 4","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-0"},"gpt-4o":{"id":"gpt-4o","name":"OpenAI GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-05","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"claude-sonnet-4":{"id":"claude-sonnet-4","name":"Anthropic: Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-0"},"gemma-3-12b-it":{"id":"gemma-3-12b-it","name":"Google Gemma 3 12B","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-12","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.05,"output":0.1},"canonical_model_id":"google/gemma-3-12b-it"},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"OpenAI GPT-4.1 Nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"gpt-4o-mini":{"id":"gpt-4o-mini","name":"OpenAI GPT-4o-mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"llama-3.1-8b-instruct-turbo":{"id":"llama-3.1-8b-instruct-turbo","name":"Meta Llama 3.1 8B Instruct Turbo","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.02,"output":0.03},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"sonar-deep-research":{"id":"sonar-deep-research","name":"Perplexity Sonar Deep Research","description":"Sonar search model for autonomous research and citation-backed long-form reports","family":"sonar","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-02-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":127000,"output":4096},"cost":{"input":2,"output":8},"canonical_model_id":"perplexity/sonar-deep-research"},"gpt-4.1-mini-2025-04-14":{"id":"gpt-4.1-mini-2025-04-14","name":"OpenAI GPT-4.1 Mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"qwen2.5-coder-7b-fast":{"id":"qwen2.5-coder-7b-fast","name":"Qwen2.5 Coder 7B fast","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-09","release_date":"2024-09-15","last_updated":"2024-09-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":8192},"cost":{"input":0.03,"output":0.09}},"grok-3-mini":{"id":"grok-3-mini","name":"xAI Grok 3 Mini","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2024-06-01","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":0.3,"output":0.5,"cache_read":0.075}},"claude-sonnet-4-5-20250929":{"id":"claude-sonnet-4-5-20250929","name":"Anthropic: Claude Sonnet 4.5 (20250929)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"qwen3-next-80b-a3b-instruct":{"id":"qwen3-next-80b-a3b-instruct","name":"Qwen3 Next 80B A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":16384},"cost":{"input":0.14,"output":1.4},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"claude-3.7-sonnet":{"id":"claude-3.7-sonnet","name":"Anthropic: Claude 3.7 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02","release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-3-7-sonnet-20250219"},"o4-mini":{"id":"o4-mini","name":"OpenAI o4 Mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-06","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Anthropic: Claude 4.5 Haiku (20251001)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-10","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"qwen3-32b":{"id":"qwen3-32b","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":40960},"cost":{"input":0.29,"output":0.59},"canonical_model_id":"alibaba/qwen3-32b"},"mistral-nemo":{"id":"mistral-nemo","name":"Mistral Nemo","description":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16400},"cost":{"input":20,"output":40},"canonical_model_id":"mistral/mistral-nemo"},"gemma2-9b-it":{"id":"gemma2-9b-it","name":"Google Gemma 2","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-06","release_date":"2024-06-25","last_updated":"2024-06-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0.01,"output":0.03}},"llama-prompt-guard-2-22m":{"id":"llama-prompt-guard-2-22m","name":"Meta Llama Prompt Guard 2 22M","description":"22M-parameter (DeBERTa-xsmall) English classifier that detects prompt injection and jailbreak attempts","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-10","release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":2},"cost":{"input":0.01,"output":0.01},"canonical_model_id":"meta/llama-prompt-guard-2-22m"},"claude-3.5-sonnet-v2":{"id":"claude-3.5-sonnet-v2","name":"Anthropic: Claude 3.5 Sonnet v2","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-3-5-sonnet-20241022"},"gpt-4.1":{"id":"gpt-4.1","name":"OpenAI GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"o1":{"id":"o1","name":"OpenAI: o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"temperature":false,"knowledge":"2025-01","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"deepseek-r1-distill-llama-70b":{"id":"deepseek-r1-distill-llama-70b","name":"DeepSeek R1 Distill Llama 70B","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-01-20","last_updated":"2025-01-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":0.03,"output":0.13}},"deepseek-tng-r1t2-chimera":{"id":"deepseek-tng-r1t2-chimera","name":"DeepSeek TNG R1T2 Chimera","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek-thinking","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-02","last_updated":"2025-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":130000,"output":163840},"cost":{"input":0.3,"output":1.2}},"gpt-5.1-codex-mini":{"id":"gpt-5.1-codex-mini","name":"OpenAI: GPT-5.1 Codex Mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"claude-3-haiku-20240307":{"id":"claude-3-haiku-20240307","name":"Anthropic: Claude 3 Haiku","description":"Legacy model retained for compatibility with older integrations","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-03","release_date":"2024-03-13","last_updated":"2024-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":4096},"cost":{"input":0.25,"output":1.25,"cache_read":0.03,"cache_write":0.3},"canonical_model_id":"anthropic/claude-3-haiku-20240307"},"gpt-5":{"id":"gpt-5","name":"OpenAI GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"o3-pro":{"id":"o3-pro","name":"OpenAI o3 Pro","description":"High-effort o3 tier for difficult technical reasoning and careful answers","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2024-06","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":20,"output":80},"canonical_model_id":"openai/o3-pro"},"llama-3.3-70b-versatile":{"id":"llama-3.3-70b-versatile","name":"Meta Llama 3.3 70B Versatile","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32678},"cost":{"input":0.59,"output":0.79},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"qwen3-vl-235b-a22b-instruct":{"id":"qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.3,"output":1.5},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"grok-4":{"id":"grok-4","name":"xAI Grok 4","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-09","last_updated":"2024-07-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":3,"output":15,"cache_read":0.75}},"grok-4-1-fast-non-reasoning":{"id":"grok-4-1-fast-non-reasoning","name":"xAI Grok 4.1 Fast Non-Reasoning","description":"xAI's fast agentic tool-calling model with a 2M context window; non-reasoning variant for low-latency responses","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast"},"hermes-2-pro-llama-3-8b":{"id":"hermes-2-pro-llama-3-8b","name":"Hermes 2 Pro Llama 3 8B","description":"Llama 3 fine-tune optimized for function calling, structured JSON responses, and dialogue","family":"hermes","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-05","release_date":"2024-04-30","last_updated":"2024-04-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.14,"output":0.14},"canonical_model_id":"nousresearch/hermes-2-pro-llama-3-8b"},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"OpenAI: GPT-5.1 Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"glm-4.6":{"id":"glm-4.6","name":"Zai GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.45,"output":1.5},"canonical_model_id":"zhipuai/glm-4.6"},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"Google Gemini 2.5 Flash Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.1,"output":0.4,"cache_read":0.025,"cache_write":0.1},"canonical_model_id":"google/gemini-2.5-flash-lite"},"gpt-5.1-chat-latest":{"id":"gpt-5.1-chat-latest","name":"OpenAI GPT-5.1 Chat","description":"Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"temperature":false,"knowledge":"2025-01","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-chat-latest"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Google Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.3125,"cache_write":1.25},"canonical_model_id":"google/gemini-2.5-pro"},"sonar-pro":{"id":"sonar-pro","name":"Perplexity Sonar Pro","description":"Deeper Sonar search model with broader retrieval and stronger synthesis","family":"sonar-pro","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":4096},"cost":{"input":3,"output":15},"canonical_model_id":"perplexity/sonar-pro"},"llama-4-scout":{"id":"llama-4-scout","name":"Meta Llama 4 Scout 17B 16E","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.08,"output":0.3},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"claude-4.5-sonnet":{"id":"claude-4.5-sonnet","name":"Anthropic: Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"sonar":{"id":"sonar","name":"Perplexity Sonar","description":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","family":"sonar","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":127000,"output":4096},"cost":{"input":1,"output":1},"canonical_model_id":"perplexity/sonar"},"gemini-3-pro-preview":{"id":"gemini-3-pro-preview","name":"Google Gemini 3 Pro Preview","description":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-11","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"google/gemini-3-pro-preview"},"llama-prompt-guard-2-86m":{"id":"llama-prompt-guard-2-86m","name":"Meta Llama Prompt Guard 2 86M","description":"86M-parameter (mDeBERTa-base) multilingual classifier that detects prompt injection and jailbreak attempts","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-10","release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":512,"output":2},"cost":{"input":0.01,"output":0.01},"canonical_model_id":"meta/llama-prompt-guard-2-86m"},"llama-4-maverick":{"id":"llama-4-maverick","name":"Meta Llama 4 Maverick 17B 128E","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"ernie-4.5-21b-a3b-thinking":{"id":"ernie-4.5-21b-a3b-thinking","name":"Baidu Ernie 4.5 21B A3B Thinking","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"ernie","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-03","release_date":"2025-03-16","last_updated":"2025-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8000},"cost":{"input":0.07,"output":0.28}},"llama-3.1-8b-instant":{"id":"llama-3.1-8b-instant","name":"Meta Llama 3.1 8B Instant","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32678},"cost":{"input":0.05,"output":0.08},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"grok-3":{"id":"grok-3","name":"xAI Grok 3","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2024-06-01","last_updated":"2024-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.75}},"qwen3-coder":{"id":"qwen3-coder","name":"Qwen3 Coder 480B A35B Instruct Turbo","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.22,"output":0.95},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"}}},"llama":{"id":"llama","env":["LLAMA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.llama.com/compat/v1/","name":"Llama","doc":"https://llama.developer.meta.com/docs/models","models":{"llama-4-maverick-17b-128e-instruct-fp8":{"id":"llama-4-maverick-17b-128e-instruct-fp8","name":"Llama-4-Maverick-17B-128E-Instruct-FP8","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"groq-llama-4-maverick-17b-128e-instruct":{"id":"groq-llama-4-maverick-17b-128e-instruct","name":"Groq-Llama-4-Maverick-17B-128E-Instruct","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"cerebras-llama-4-maverick-17b-128e-instruct":{"id":"cerebras-llama-4-maverick-17b-128e-instruct","name":"Cerebras-Llama-4-Maverick-17B-128E-Instruct","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"llama-3.3-8b-instruct":{"id":"llama-3.3-8b-instruct","name":"Llama-3.3-8B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"llama-4-scout-17b-16e-instruct-fp8":{"id":"llama-4-scout-17b-16e-instruct-fp8","name":"Llama-4-Scout-17B-16E-Instruct-FP8","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"cerebras-llama-4-scout-17b-16e-instruct":{"id":"cerebras-llama-4-scout-17b-16e-instruct","name":"Cerebras-Llama-4-Scout-17B-16E-Instruct","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-4-scout-17b-instruct"}}},"bailing":{"id":"bailing","env":["BAILING_API_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"https://api.tbox.cn/api/llm/v1","name":"Bailing","doc":"https://alipaytbox.yuque.com/sxs0ba/ling/intro","models":{"Ring-1T":{"id":"Ring-1T","name":"Ring-1T","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"ring","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-10","last_updated":"2025-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.57,"output":2.29}},"Ling-1T":{"id":"Ling-1T","name":"Ling-1T","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"ling","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-10","last_updated":"2025-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.57,"output":2.29}}}},"upstage":{"id":"upstage","env":["UPSTAGE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.upstage.ai/v1/solar","name":"Upstage","doc":"https://developers.upstage.ai/docs/apis/chat","models":{"solar-mini4":{"id":"solar-mini4","name":"Solar Mini 4","description":"Upstage's compact, cost-efficient model for agentic use","family":"solar-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"upstage/solar-mini4"},"solar-mini":{"id":"solar-mini","name":"solar-mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"solar-mini","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2024-06-12","last_updated":"2025-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":4096},"cost":{"input":0.15,"output":0.15}},"solar-pro2":{"id":"solar-pro2","name":"solar-pro2","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":8192},"cost":{"input":0.25,"output":0.25},"canonical_model_id":"upstage/solar-pro2"},"solar-pro4":{"id":"solar-pro4","name":"Solar Pro 4","description":"Upstage's flagship model, specialized for agentic use","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"upstage/solar-pro4"},"solar-pro3":{"id":"solar-pro3","name":"solar-pro3","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2026-01-26","last_updated":"2026-03-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.25,"output":0.25},"canonical_model_id":"upstage/solar-pro3"}}},"nebius":{"id":"nebius","env":["NEBIUS_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.tokenfactory.nebius.com/v1","name":"Nebius Token Factory","doc":"https://docs.tokenfactory.nebius.com/","models":{"moonshotai/Kimi-K3":{"id":"moonshotai/Kimi-K3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":8000},"cost":{"input":3,"output":15,"cache_read":3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/Kimi-K2.7-Code":{"id":"moonshotai/Kimi-K2.7-Code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8000},"cost":{"input":0.95,"output":4,"cache_read":0.95},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"NousResearch/Hermes-4-405B":{"id":"NousResearch/Hermes-4-405B","name":"Hermes-4-405B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"hermes","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":120000,"output":8192},"cost":{"input":1,"output":3,"reasoning":3,"cache_read":1},"canonical_model_id":"nousresearch/hermes-4-405b"},"MiniMaxAI/MiniMax-M3":{"id":"MiniMaxAI/MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.3,"output":1.2,"cache_read":0.3},"canonical_model_id":"minimax/MiniMax-M3"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"gpt-oss-120b","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"input":124000,"output":8192},"cost":{"input":0.15,"output":0.6,"reasoning":0.6,"cache_read":0.15},"canonical_model_id":"openai/gpt-oss-120b"},"meta-models/muse-spark-open":{"id":"meta-models/muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.25,"output":4.25,"cache_read":1.25},"canonical_model_id":"meta/muse-spark-open"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.4,"output":4.4,"cache_read":1.4},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":1024000},"cost":{"input":1.4,"output":4.4,"cache_read":1.4},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":1024000},"cost":{"input":0.15,"output":0.5,"cache_read":0.15},"canonical_model_id":"zhipuai/glm-5.3-flash"},"reflection/beam":{"id":"reflection/beam","name":"Beam 1.0 501B A23B","description":"Sparse mixture-of-experts reasoning model with 23B active parameters for coding and agentic workflows","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.25,"output":1,"cache_read":0.25},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"nvidia/Nemotron-3_5-Lightning":{"id":"nvidia/Nemotron-3_5-Lightning","name":"Nemotron 3.5 Lightning 30B A3B","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.06,"output":0.24,"cache_read":0.06},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/Nemotron-3.5-Super-VL":{"id":"nvidia/Nemotron-3.5-Super-VL","name":"Nemotron 3.5 Super VL 120B A12B","description":"NVIDIA vision-language reasoning model for coding, visual understanding, and agentic tool use","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.3,"output":0.9,"cache_read":0.3},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/nemotron-3-super-120b-a12b":{"id":"nvidia/nemotron-3-super-120b-a12b","name":"Nemotron-3-Super-120B-A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.3},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/Nemotron-3-Ultra-550b-a55b":{"id":"nvidia/Nemotron-3-Ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1,"output":3,"cache_read":1},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"deepseek-ai/DeepSeek-V4-Flash-0731":{"id":"deepseek-ai/DeepSeek-V4-Flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":1024000},"cost":{"input":0.14,"output":0.28,"cache_read":0.14},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek-ai/DeepSeek-V4-Pro":{"id":"deepseek-ai/DeepSeek-V4-Pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.75,"output":3.5,"cache_read":1.75},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/DeepSeek-V4-Pro-0813":{"id":"deepseek-ai/DeepSeek-V4-Pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":979000,"output":979000},"cost":{"input":1.32,"output":3.96,"cache_read":1.32},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek-ai/DeepSeek-V4.1-Flash":{"id":"deepseek-ai/DeepSeek-V4.1-Flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048000,"output":1048000},"cost":{"input":0.3,"output":1.2,"cache_read":0.3},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.2,"output":0.6,"cache_read":0.2},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"Qwen/Qwen3.8-27B":{"id":"Qwen/Qwen3.8-27B","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.45,"output":3,"cache_read":0.45},"canonical_model_id":"alibaba/qwen3.8-27b"},"Qwen/Qwen3-30B-A3B-Instruct-2507":{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507","name":"Qwen3-30B-A3B-Instruct-2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":262144,"output":8192},"cost":{"input":0.1,"output":0.3,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3-30b-a3b-instruct-2507"},"Qwen/Qwen3.5-397B-A17B":{"id":"Qwen/Qwen3.5-397B-A17B","name":"Qwen3.5-397B-A17B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07","release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":250000,"output":8192},"cost":{"input":0.6,"output":3.6,"cache_read":0.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"Qwen/Qwen3-Embedding-8B":{"id":"Qwen/Qwen3-Embedding-8B","type":"embedding","name":"Qwen3-Embedding-8B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2025-10","release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"input":40960,"output":0},"cost":{"input":0.01,"output":0},"canonical_model_id":"alibaba/qwen3-embedding-8b"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma-3-27b-it","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":110000,"input":100000,"output":8192},"cost":{"input":0.1,"output":0.3,"cache_read":0.1},"canonical_model_id":"google/gemma-3-27b-it"}}},"xai":{"id":"xai","env":["XAI_API_KEY"],"npm":"@ai-sdk/xai","name":"xAI","doc":"https://docs.x.ai/docs/models","models":{"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"grok-imagine-image":{"id":"grok-imagine-image","type":"image","name":"Grok Imagine Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-28","last_updated":"2026-01-28","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":16000,"output":0},"canonical_model_id":"xai/grok-imagine-image"},"grok-4.20-0309-non-reasoning":{"id":"grok-4.20-0309-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"grok-4.7":{"id":"grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"grok-4.20-multi-agent-0309":{"id":"grok-4.20-multi-agent-0309","name":"Grok 4.20 Multi-Agent","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"grok-imagine-video":{"id":"grok-imagine-video","type":"video","name":"Grok Imagine Video","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-01-28","last_updated":"2026-01-28","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":1024,"output":0}},"grok-imagine-image-quality":{"id":"grok-imagine-image-quality","type":"image","name":"Grok Imagine Image Quality","description":"Higher-fidelity Grok Imagine image model for prompt-driven generation, editing, and visual design workflows","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-03","last_updated":"2026-04-03","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":16000,"output":0},"status":"deprecated","canonical_model_id":"xai/grok-imagine-image-quality"},"grok-imagine-video-1.5-lite":{"id":"grok-imagine-video-1.5-lite","type":"video","name":"Grok Imagine Video 1.5 Lite","description":"Cost-efficient video model for text-to-video and image-to-video generation","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-10-01","last_updated":"2026-10-01","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":1024,"output":0},"canonical_model_id":"xai/grok-imagine-video-1.5-lite"},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"xAI's Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"grok-4.20-0309-reasoning":{"id":"grok-4.20-0309-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"grok-imagine-video-1.5":{"id":"grok-imagine-video-1.5","type":"video","name":"Grok Imagine Video 1.5","description":"Video model for text-to-video, image-to-video, and reference-to-video generation with optional voice references","family":"grok","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-05-30","last_updated":"2026-05-30","modalities":{"input":["text","image","audio"],"output":["video"]},"open_weights":false,"limit":{"context":1024,"output":0},"canonical_model_id":"xai/grok-imagine-video-1.5"},"grok-build-0.1":{"id":"grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}}}}},"the-grid-ai":{"id":"the-grid-ai","env":["THEGRID_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.thegrid.ai/v1","name":"The Grid AI","doc":"https://thegrid.ai/docs","models":{"minimax-latest":{"id":"minimax-latest","name":"MiniMax Latest","description":"Lab latest market for MiniMax supply. You contract for the latest qualifying route, not a specific model.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":512000}},"text-max":{"id":"text-max","name":"Text Max","description":"Frontier models for deep reasoning, long context, and complex workflows. Any model that meets the contract spec can serve your request.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-02-26","last_updated":"2026-07-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000}},"gemini-pro-latest":{"id":"gemini-pro-latest","name":"Gemini Pro Latest","description":"Lab latest market for Gemini Pro frontier supply. You contract for the latest qualifying route, not a specific model.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":65536}},"bytedance-pro-latest":{"id":"bytedance-pro-latest","name":"ByteDance Pro Latest","description":"Lab latest market for ByteDance Pro supply. You contract for the latest qualifying route, not a specific model.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"input":256000,"output":131072}},"deepseek-pro-latest":{"id":"deepseek-pro-latest","name":"DeepSeek Pro Latest","description":"Lab latest market for DeepSeek Pro supply. You contract for the latest qualifying route, not a specific model.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":384000}},"glm-latest":{"id":"glm-latest","name":"GLM Latest","description":"Lab latest market for GLM supply. You contract for the latest qualifying route, not a specific model.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-06","last_updated":"2026-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":917504,"output":131072}},"code-standard":{"id":"code-standard","name":"Code Standard","description":"Price-optimized models for rapid autocomplete, linting, high-frequency suggestions, and batch edits. Any model that meets the contract spec can serve your request.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-04","last_updated":"2026-10-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":120000,"output":65536},"status":"beta"},"gpt-sol-latest":{"id":"gpt-sol-latest","name":"GPT Sol Latest","description":"Lab latest market for GPT frontier supply. You contract for the latest qualifying route, not a specific model.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000}},"text-prime":{"id":"text-prime","name":"Text Prime","description":"Reliable models for everyday text generation, editing, and analysis across diverse workflows. Any model that meets the contract spec can serve your request.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-26","last_updated":"2026-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":196608,"input":120000,"output":131072}},"agent-standard":{"id":"agent-standard","name":"Agent Standard","description":"Price-optimized models for fast tool calls, simple agent loops, high-throughput automation, and orchestration. Any model that meets the contract spec can serve your request.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-04","last_updated":"2026-10-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":120000,"output":65536},"status":"beta"},"kimi-latest":{"id":"kimi-latest","name":"Kimi Latest","description":"Lab latest market for Kimi supply. You contract for the latest qualifying route, not a specific model.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"input":922000,"output":943718}},"agent-max":{"id":"agent-max","name":"Agent Max","description":"Frontier models for autonomous research, deep multi-step tool chains, and complex long-horizon tasks. Any model that meets the contract spec can serve your request.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-05-04","last_updated":"2026-07-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"status":"beta"},"code-prime":{"id":"code-prime","name":"Code Prime","description":"Reliable models for everyday software tasks, code completion, review, and standard debugging. Any model that meets the contract spec can serve your request.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-04","last_updated":"2026-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":196608,"input":120000,"output":131072},"status":"beta"},"agent-prime":{"id":"agent-prime","name":"Agent Prime","description":"Reliable models for dependable agentic applications, multi-step tool use, and reasoning workflows. Any model that meets the contract spec can serve your request.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-04","last_updated":"2026-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":196608,"input":120000,"output":131072},"status":"beta"},"code-max":{"id":"code-max","name":"Code Max","description":"Frontier models for complex research, architectural decisions, debugging, and multi-file development. Any model that meets the contract spec can serve your request.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-05-04","last_updated":"2026-07-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"status":"beta"},"text-standard":{"id":"text-standard","name":"Text Standard","description":"Price-optimized models with low-latency, high-throughput and shorter maximum outputs. Any model that meets the contract spec can serve your request.","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-26","last_updated":"2026-10-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":120000,"output":65536}},"claude-opus-latest":{"id":"claude-opus-latest","name":"Claude Opus Latest","description":"Lab latest market for Claude Opus frontier supply. You contract for the latest qualifying route, not a specific model.","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-06","last_updated":"2026-07-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000}}}},"alibaba":{"id":"alibaba","env":["DASHSCOPE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://dashscope-intl.aliyuncs.com/compatible-mode/v1","name":"Alibaba","doc":"https://www.alibabacloud.com/help/en/model-studio/models","models":{"qwen-vl-plus":{"id":"qwen-vl-plus","name":"Qwen-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-08-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.21,"output":0.63}},"qwen3.5-35b-a3b":{"id":"qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2}},"qwen2-5-32b-instruct":{"id":"qwen2-5-32b-instruct","name":"Qwen2.5 32B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.7,"output":2.8}},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.5,"cache_write":3.125}},"qwen3-omni-flash":{"id":"qwen3-omni-flash","name":"Qwen3-Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":false,"limit":{"context":65536,"output":16384},"cost":{"input":0.43,"output":1.66,"input_audio":3.81,"output_audio":15.11}},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-04","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.4,"output":1.6,"cache_read":0.04,"cache_write":0.5,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.12,"cache_write":1.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.12,"cache_write":1.5}}},"qwen3-livetranslate-flash-realtime":{"id":"qwen3-livetranslate-flash-realtime","type":"realtime","name":"Qwen3-LiveTranslate Flash Realtime","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-09-22","last_updated":"2025-09-22","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":false,"limit":{"context":53248,"output":4096},"status":"deprecated","cost":{"input":10,"output":10,"input_audio":10,"output_audio":38}},"qwen3-max":{"id":"qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":1.2,"output":6}},"qwen3-14b":{"id":"qwen3-14b","name":"Qwen3 14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.35,"output":1.4,"reasoning":4.2}},"qwen-max":{"id":"qwen-max","name":"Qwen Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":1.6,"output":6.4}},"qwen-flash":{"id":"qwen-flash","name":"Qwen Flash","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.05,"output":0.4,"tiers":[{"input":0.25,"output":2,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.25,"output":2}}},"qwen3-vl-235b-a22b":{"id":"qwen3-vl-235b-a22b","name":"Qwen3-VL 235B-A22B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.7,"output":2.8,"reasoning":8.4}},"qwen3-coder-plus":{"id":"qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Hosted Qwen coder for software agents, repo edits, and long-context code","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":1,"output":5}},"qwen-vl-max":{"id":"qwen-vl-max","name":"Qwen-VL Max","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-08","last_updated":"2025-08-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.8,"output":3.2}},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.375,"output":2.25}},"qwen3.5-122b-a10b":{"id":"qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":3.2}},"qwen3-omni-flash-realtime":{"id":"qwen3-omni-flash-realtime","type":"realtime","name":"Qwen3-Omni Flash Realtime","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":false,"limit":{"context":65536,"output":16384},"status":"deprecated","cost":{"input":0.52,"output":1.99,"input_audio":4.57,"output_audio":18.13}},"qwen3-vl-30b-a3b":{"id":"qwen3-vl-30b-a3b","name":"Qwen3-VL 30B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.2,"output":0.8,"reasoning":2.4}},"qwen3.5-flash":{"id":"qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"alibaba/qwen3.5-flash"},"qwen3-asr-flash":{"id":"qwen3-asr-flash","type":"transcription","name":"Qwen3-ASR Flash","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-04","release_date":"2025-09-08","last_updated":"2025-09-08","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":53248,"output":4096}},"qwen-omni-turbo":{"id":"qwen-omni-turbo","name":"Qwen-Omni Turbo","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-01-19","last_updated":"2025-03-26","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":false,"limit":{"context":32768,"output":2048},"status":"deprecated","cost":{"input":0.07,"output":0.27,"input_audio":4.44,"output_audio":8.89}},"qwen3-vl-plus":{"id":"qwen3-vl-plus","name":"Qwen3-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":1.6,"tiers":[{"input":0.3,"output":2.4,"tier":{"type":"context","size":32000}},{"input":0.6,"output":4.8,"tier":{"type":"context","size":128000}}]}},"qwen-mt-turbo":{"id":"qwen-mt-turbo","name":"Qwen-MT Turbo","description":"Translation model for multilingual conversion, localization, and cross-language workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-01","last_updated":"2025-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":8192},"status":"deprecated","cost":{"input":0.16,"output":0.49}},"qwen3.6-max-preview":{"id":"qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":1.3,"output":7.8,"cache_read":0.13,"cache_write":1.625}},"qwen2-5-14b-instruct":{"id":"qwen2-5-14b-instruct","name":"Qwen2.5 14B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.35,"output":1.4}},"qwen2-5-vl-7b-instruct":{"id":"qwen2-5-vl-7b-instruct","name":"Qwen2.5-VL 7B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.35,"output":1.05}},"qwen3.7-flash":{"id":"qwen3.7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":131072},"cost":{"input":0.03,"output":0.13,"cache_read":0.003,"cache_write":0.0375,"tiers":[{"input":0.1,"output":0.4,"cache_read":0.01,"cache_write":0.125,"tier":{"type":"context","size":32000}},{"input":0.2,"output":0.8,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen3-235b-a22b":{"id":"qwen3-235b-a22b","name":"Qwen3 235B-A22B","description":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.7,"output":2.8,"reasoning":8.4}},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen-vl-ocr":{"id":"qwen-vl-ocr","name":"Qwen-VL OCR","description":"OCR model for extracting structured text from documents and screenshots","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-10-28","last_updated":"2025-04-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":34096,"output":4096},"cost":{"input":0.07,"output":0.16}},"qwen3.5-27b":{"id":"qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":2.4}},"qwen-mt-plus":{"id":"qwen-mt-plus","name":"Qwen-MT Plus","description":"Translation model for multilingual conversion, localization, and cross-language workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-01","last_updated":"2025-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":8192},"cost":{"input":2.46,"output":7.37}},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.3125,"tiers":[{"input":1,"output":4,"cache_read":0.1,"cache_write":1.25,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1,"output":4,"cache_read":0.1,"cache_write":1.25}}},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6}},"qwen3-coder-30b-a3b-instruct":{"id":"qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.45,"output":2.25,"tiers":[{"input":0.75,"output":3.75,"tier":{"type":"context","size":32000}},{"input":1.2,"output":6,"tier":{"type":"context","size":128000}}]}},"qwen3-next-80b-a3b-thinking":{"id":"qwen3-next-80b-a3b-thinking","name":"Qwen3-Next 80B-A3B (Thinking)","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.5,"output":6}},"qwen3.8-omni-flash":{"id":"qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"qwen-turbo":{"id":"qwen-turbo","name":"Qwen Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-11-01","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":16384},"status":"deprecated","cost":{"input":0.05,"output":0.2,"reasoning":0.5}},"qwen-plus-character-ja":{"id":"qwen-plus-character-ja","name":"Qwen Plus Character (Japanese)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-01","last_updated":"2024-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":512},"cost":{"input":0.5,"output":1.4}},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"qvq-max":{"id":"qvq-max","name":"QVQ Max","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qvq","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-03-25","last_updated":"2025-03-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":1.2,"output":4.8}},"qwen3-next-80b-a3b-instruct":{"id":"qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.5,"output":2}},"qwen3-32b":{"id":"qwen3-32b","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.7,"output":2.8,"reasoning":8.4}},"qwq-plus":{"id":"qwq-plus","name":"QwQ Plus","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-03-05","last_updated":"2025-03-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.8,"output":2.4}},"qwen3-8b":{"id":"qwen3-8b","name":"Qwen3 8B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.18,"output":0.7,"reasoning":2.1}},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.28,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"qwen2-5-vl-72b-instruct":{"id":"qwen2-5-vl-72b-instruct","name":"Qwen2.5-VL 72B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":2.8,"output":8.4}},"qwen3-coder-480b-a35b-instruct":{"id":"qwen3-coder-480b-a35b-instruct","name":"Qwen3-Coder 480B-A35B Instruct","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":1.5,"output":7.5,"tiers":[{"input":2.7,"output":13.5,"tier":{"type":"context","size":32000}},{"input":4.5,"output":22.5,"tier":{"type":"context","size":128000}}]}},"qwen3.5-plus":{"id":"qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.4,"output":2.4,"reasoning":2.4,"tiers":[{"input":0.5,"output":3,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.5,"output":3}}},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625,"tiers":[{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5}}},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.44,"output":1.32,"cache_read":0.088},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"qwen-omni-turbo-realtime":{"id":"qwen-omni-turbo-realtime","type":"realtime","name":"Qwen-Omni Turbo Realtime","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-05-08","last_updated":"2025-05-08","modalities":{"input":["text","image","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":32768,"output":2048},"status":"deprecated","cost":{"input":0.27,"output":1.07,"input_audio":4.44,"output_audio":8.89}},"qwen2-5-72b-instruct":{"id":"qwen2-5-72b-instruct","name":"Qwen2.5 72B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":1.4,"output":5.6}},"qwen2-5-omni-7b":{"id":"qwen2-5-omni-7b","name":"Qwen2.5-Omni 7B","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-12","last_updated":"2024-12","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":true,"limit":{"context":32768,"output":2048},"status":"deprecated","cost":{"input":0.1,"output":0.4,"input_audio":6.76}},"qwen2-5-7b-instruct":{"id":"qwen2-5-7b-instruct","name":"Qwen2.5 7B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.175,"output":0.7}},"qwen-plus":{"id":"qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.4,"output":1.2,"reasoning":4,"tiers":[{"input":1.2,"output":3.6,"reasoning":12,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.2,"output":3.6,"reasoning":12}}},"qwen3-coder-flash":{"id":"qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":1.5,"tiers":[{"input":0.5,"output":2.5,"tier":{"type":"context","size":32000}},{"input":0.8,"output":4,"tier":{"type":"context","size":128000}},{"input":1.6,"output":9.6,"tier":{"type":"context","size":256000}}]}},"qwen3.6-27b":{"id":"qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6}}}},"llmgateway":{"id":"llmgateway","env":["LLMGATEWAY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.llmgateway.io/v1","name":"DevPass (LLM Gateway)","doc":"https://llmgateway.io/docs","models":{"grok-4-20-reasoning":{"id":"grok-4-20-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"claude-sonnet-5-5":{"id":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.8-flash"},"qwen3.5-35b-a3b":{"id":"qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"devstral-2512":{"id":"devstral-2512","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/devstral-2512"},"qwen3-vl-30b-a3b-instruct":{"id":"qwen3-vl-30b-a3b-instruct","name":"Qwen3 VL 30B A3B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.1-flash-lite"},"qwen3-30b-a3b":{"id":"qwen3-30b-a3b","name":"Qwen3 30B A3B","description":"Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.095,"output":0.285},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"minimax-m2.1-lightning":{"id":"minimax-m2.1-lightning","name":"MiniMax M2.1 Lightning","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.12,"output":0.48}},"gpt-6-sol":{"id":"gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-6-sol"},"gpt-5.1":{"id":"gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.25,"output":3.75,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.7-max"},"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"muse-wonder-1.0":{"id":"muse-wonder-1.0","name":"Muse Wonder","description":"Muse Wonder is a frontier multimodal reasoning model from Meta for personal and professional agents, long-horizon coding, and computer use. It combines Active Memory, which carries relevant context across conversations, with Active Focus, which keeps it working through blockers toward a goal.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"seed-2-0-pro-260328":{"id":"seed-2-0-pro-260328","name":"Seed 2.0 Pro (260328)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-28","last_updated":"2026-03-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":3,"cache_read":0.1}},"seed-1-6-250615":{"id":"seed-1-6-250615","name":"Seed 1.6 (250615)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-06-25","last_updated":"2025-06-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"gpt-5-pro":{"id":"gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":64000},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen3-max":{"id":"qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.845,"output":3.38,"cache_read":0.6,"cache_write":3.75},"canonical_model_id":"alibaba/qwen3-max"},"llama-3.2-3b-instruct":{"id":"llama-3.2-3b-instruct","name":"Llama 3.2 3B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2024-09-18","last_updated":"2024-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32000},"cost":{"input":0.03,"output":0.05}},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":1050000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"openai/gpt-6-astra"},"o3-mini":{"id":"o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"gpt-5.5-pro":{"id":"gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Gemini multimodal model for text, image, audio, video, and document tasks","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-4-argon"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32766},"cost":{"input":0.04,"output":0.19,"cache_read":0.01},"canonical_model_id":"openai/gpt-oss-20b"},"grok-4-6":{"id":"grok-4-6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"qwen3-14b":{"id":"qwen3-14b","name":"Qwen3 14B","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.076,"output":0.228}},"minimax-m2.5-highspeed":{"id":"minimax-m2.5-highspeed","name":"MiniMax-M2.5-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5-highspeed"},"grok-4-7":{"id":"grok-4-7","name":"Grok 4.7","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"qwen-max":{"id":"qwen-max","name":"Qwen Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":1.6,"output":6.4},"canonical_model_id":"alibaba/qwen-max"},"step-3.7-flash":{"id":"step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"ling-3.0-flash":{"id":"ling-3.0-flash","name":"InclusionAI Ling 3.0 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.06,"output":0.18,"cache_read":0.012},"canonical_model_id":"inclusionai/ling-3.0-flash"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"glm-4.6v-flashx":{"id":"glm-4.6v-flashx","name":"GLM-4.6V FlashX","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000},"cost":{"input":0.04,"output":0.4,"cache_read":0.004}},"qwen-flash":{"id":"qwen-flash","name":"Qwen Flash","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.05,"output":0.4,"cache_read":0.01,"cache_write":0.0625},"canonical_model_id":"alibaba/qwen-flash"},"fugu-ultra-v2.0":{"id":"fugu-ultra-v2.0","name":"Fugu Ultra v2.0","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":5,"output":30,"cache_read":0.5}},"gpt-4":{"id":"gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"status":"deprecated","cost":{"input":30,"output":60},"canonical_model_id":"openai/gpt-4"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"ministral-3b-2512":{"id":"ministral-3b-2512","name":"Ministral 3B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.1,"output":0.1},"canonical_model_id":"mistral/ministral-3b-2512"},"gpt-5-mini":{"id":"gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.08,"output":0.32,"cache_read":0.017,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.168,"output":0.336,"cache_read":0.0034,"tiers":[{"input":0.8,"output":4,"cache_read":0.16,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.8,"output":4,"cache_read":0.16}},"canonical_model_id":"xiaomi/mimo-v2.5"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"seed-2-0-lite-260428":{"id":"seed-2-0-lite-260428","name":"Seed 2.0 Lite (260428)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"muse-spark-1.3":{"id":"muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"claude-haiku-5-5":{"id":"claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.522,"output":1.044,"cache_read":0.0043,"tiers":[{"input":2,"output":6,"cache_read":0.4,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.4}},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"gpt-5.4-pro":{"id":"gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.4-pro"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.248,"output":1.485},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen3.5-122b-a10b":{"id":"qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":3.2},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"ling-3.0-flash-vl":{"id":"ling-3.0-flash-vl","name":"Ling 3.0 Flash VL","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.06,"output":0.18,"cache_read":0.012},"canonical_model_id":"inclusionai/ling-3.0-flash-vl"},"glm-4.5-air":{"id":"glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":98304},"cost":{"input":0.2,"output":1.1,"cache_read":0.03,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-air"},"qwen3.8-2.4t-a95b":{"id":"qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1010000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":384000},"cost":{"input":0.076,"output":0.153,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash"},"glm-4.7-flashx":{"id":"glm-4.7-flashx","name":"GLM-4.7-FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.07,"output":0.4,"cache_read":0.01,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"qwen3-coder-next":{"id":"qwen3-coder-next","name":"Qwen3 Coder Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.108,"output":0.675,"cache_read":0.06},"canonical_model_id":"alibaba/qwen3-coder-next"},"gpt-5-nano":{"id":"gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"gemini-pro-latest":{"id":"gemini-pro-latest","name":"Gemini Pro Latest","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-27","last_updated":"2026-02-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2}},"gpt-5.2":{"id":"gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3.05,"cache_read":0.13},"canonical_model_id":"moonshotai/kimi-k2.6"},"mistral-large-4":{"id":"mistral-large-4","name":"Mistral Large 4","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":262144},"cost":{"input":1.36,"output":4.18,"cache_read":0.14},"canonical_model_id":"mistral/mistral-large-4"},"granite-4.2-8b":{"id":"granite-4.2-8b","name":"Granite 4.2 8B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"granite","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-25","last_updated":"2026-08-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.06,"output":0.25,"cache_read":0.015},"canonical_model_id":"ibm/granite-4.2-8b"},"muse-spark-1.3-contributor":{"id":"muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"mistral-large-latest":{"id":"mistral-large-latest","name":"Mistral Large (latest)","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":262144},"cost":{"input":4,"output":12},"canonical_model_id":"mistral/mistral-large-latest"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"qwen3-vl-plus":{"id":"qwen3-vl-plus","name":"Qwen3-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":1.6,"reasoning":4.8,"cache_read":0.04,"cache_write":0.25},"canonical_model_id":"alibaba/qwen3-vl-plus"},"gpt-transcribe":{"id":"gpt-transcribe","type":"transcription","name":"GPT-Transcribe","description":"High-accuracy speech-to-text model for audio files and committed Realtime turns","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-07-28","last_updated":"2026-07-28","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-transcribe"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"gpt-4o-transcribe":{"id":"gpt-4o-transcribe","name":"GPT-4o Transcribe","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-03-20","last_updated":"2025-03-20","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":16000},"cost":{"input":2.5,"output":10}},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.12,"output":0.25,"cache_read":0.0025},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":16384},"cost":{"input":0.26,"output":0.38,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":1050000},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"glm-4.5v":{"id":"glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.6,"output":1.8,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5v"},"hy-mt2-plus":{"id":"hy-mt2-plus","name":"Hy-MT2 Plus","description":"Tool-capable chat model for instruction following and agentic application workflows","family":"Hy","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0.074,"output":0.295}},"o3":{"id":"o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"gpt-4o-mini-transcribe":{"id":"gpt-4o-mini-transcribe","name":"GPT-4o Mini Transcribe","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-03-20","last_updated":"2025-03-20","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":16000},"cost":{"input":1.25,"output":5}},"gpt-6-luna":{"id":"gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"openai/gpt-6-luna"},"qwen3-235b-a22b-instruct-2507":{"id":"qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B A22B Instruct (2507)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.09,"output":0.58},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":384000},"cost":{"input":0.435,"output":0.87,"cache_read":0.003625},"canonical_model_id":"deepseek/deepseek-v4-pro"},"grok-4-1-fast-reasoning":{"id":"grok-4-1-fast-reasoning","name":"Grok 4.1 Fast Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast-reasoning"},"seed-1-6-flash-250715":{"id":"seed-1-6-flash-250715","name":"Seed 1.6 Flash (250715)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.07,"output":0.3,"cache_read":0.015}},"llama-3.1-70b-instruct":{"id":"llama-3.1-70b-instruct","name":"Llama 3.1 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":2048},"status":"beta","cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-3.1-70b-instruct"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"Qwen3.8-27B":{"id":"Qwen3.8-27B","name":"Qwen3.8 27B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":32768},"cost":{"input":0.08,"output":0.35,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.8-27b"},"minimax-m2.1":{"id":"minimax-m2.1","name":"MiniMax-M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.27,"output":1.1},"canonical_model_id":"minimax/MiniMax-M2.1"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1.38,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32766},"cost":{"input":0.032,"output":0.14,"cache_read":0.032},"canonical_model_id":"openai/gpt-oss-120b"},"qwen3-vl-235b-a22b-thinking":{"id":"qwen3-vl-235b-a22b-thinking","name":"Qwen3 VL 235B A22B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0.98,"output":3.95},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-thinking"},"seed-2-0-code-preview-260328":{"id":"seed-2-0-code-preview-260328","name":"Seed 2.0 Code Preview (260328)","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-28","last_updated":"2026-03-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":3,"cache_read":0.1}},"claude-opus-5-5":{"id":"claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"kimi-k2.7-code-highspeed":{"id":"kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"muse-glimmer-30b":{"id":"muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5},"canonical_model_id":"openai/gpt-5.6-sol"},"ministral-8b-2512":{"id":"ministral-8b-2512","name":"Ministral 8B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.15,"output":0.15},"canonical_model_id":"mistral/ministral-8b-2512"},"mistral-small-2603":{"id":"mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"mistral/mistral-small-2603"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"qwen3.7-flash":{"id":"qwen3.7-flash","name":"Qwen3.7 Flash","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":1000000},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.0375},"canonical_model_id":"alibaba/qwen3.7-flash"},"minimax-text-01":{"id":"minimax-text-01","name":"MiniMax Text 01","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-01-15","last_updated":"2025-01-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.2,"output":1.1}},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max Preview","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-19","last_updated":"2026-07-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.54,"output":4.631,"cache_read":0.198,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.737,"output":3.08,"cache_read":0.143},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gpt-6.1-sol":{"id":"gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"openai/gpt-6.1-sol"},"qwen3.5-27b":{"id":"qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":2.4},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.05,"cache_write":0.3125},"canonical_model_id":"alibaba/qwen3.6-flash"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"claude-opus-4-5-20251101":{"id":"claude-opus-4-5-20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"qwen3-coder-30b-a3b-instruct":{"id":"qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":65536},"cost":{"input":0.07,"output":0.27},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"nemotron-3-ultra-550b":{"id":"nemotron-3-ultra-550b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"cost":{"input":0.5,"output":2.2,"cache_read":0.1},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"codestral-2508":{"id":"codestral-2508","name":"Codestral","description":"Mistral coding model for code completion, generation, and developer workflows","family":"codestral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":16384},"cost":{"input":0.3,"output":0.9},"canonical_model_id":"mistral/codestral-2508"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.45,"output":2,"cache_read":0.19,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0.135,"output":0.4},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"gemma-4-26b-a4b-it":{"id":"gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.07,"output":0.34},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"gpt-4o":{"id":"gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"reka-edge-2603":{"id":"reka-edge-2603","name":"Reka Edge 2603","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"reka","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-09","last_updated":"2026-10-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":16384},"cost":{"input":0.1,"output":0.1}},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.6-flash"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.7-flash"},"glm-5.3-fast":{"id":"glm-5.3-fast","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2.8,"output":8.8,"cache_read":0.56},"canonical_model_id":"zhipuai/glm-5.3"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"gpt-4.1-nano":{"id":"gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"status":"deprecated","cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"gpt-4o-mini":{"id":"gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"grok-4-20-beta-0309-reasoning":{"id":"grok-4-20-beta-0309-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"llama-4-scout-17b-instruct":{"id":"llama-4-scout-17b-instruct","name":"Llama 4 Scout 17B Instruct","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":2048},"cost":{"input":0.18,"output":0.59},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"muse-spark-1.1":{"id":"muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"grok-build-0-1":{"id":"grok-build-0-1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}},"canonical_model_id":"xai/grok-build-0.1"},"muse-spark-1.2":{"id":"muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"glm-5-turbo":{"id":"glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"glm-4.5":{"id":"glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.574,"output":3.011,"cache_read":0.225},"canonical_model_id":"moonshotai/kimi-k2.5"},"minimax-m2":{"id":"minimax-m2","name":"MiniMax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.2,"output":1,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2"},"claude-sonnet-4-5-20250929":{"id":"claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"qwen3-next-80b-a3b-instruct":{"id":"qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":1.5},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"gpt-4-turbo":{"id":"gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"grok-4-20-beta-0309-non-reasoning":{"id":"grok-4-20-beta-0309-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"ministral-14b-2512":{"id":"ministral-14b-2512","name":"Ministral 14B","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.2,"output":0.2},"canonical_model_id":"mistral/ministral-14b-2512"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.088,"output":0.25,"cache_read":0.025},"canonical_model_id":"zhipuai/glm-5.3-flash"},"qwen3-32b":{"id":"qwen3-32b","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.36,"output":0.87,"reasoning":8.4},"canonical_model_id":"alibaba/qwen3-32b"},"qwen35-397b-a17b":{"id":"qwen35-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"glm-4-32b-0414-128k":{"id":"glm-4-32b-0414-128k","name":"GLM-4 32B (0414-128k)","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.1,"output":0.1}},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.28,"output":1.1,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M3"},"grok-4-20-non-reasoning":{"id":"grok-4-20-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"gpt-3.5-turbo":{"id":"gpt-3.5-turbo","name":"GPT-3.5-turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"status":"deprecated","cost":{"input":0.5,"output":1.5,"cache_read":0},"canonical_model_id":"openai/gpt-3.5-turbo"},"llama-4-maverick-17b-instruct":{"id":"llama-4-maverick-17b-instruct","name":"Llama 4 Maverick 17B Instruct","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":2048},"cost":{"input":0.27,"output":0.85},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"glm-4.5-x":{"id":"glm-4.5-x","name":"GLM-4.5 X","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"beta","cost":{"input":2.2,"output":8.9,"cache_read":0.45}},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.605,"output":2.123,"cache_read":0.154,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.5-flash"},"muse-spark-1.2-contributor":{"id":"muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"glm-4.7-flash":{"id":"glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.06,"output":0.4,"cache_read":0.01,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"fugu-max":{"id":"fugu-max","name":"Fugu Max","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"sakana/fugu-max"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":384000},"cost":{"input":0.135,"output":0.54,"cache_read":0.0027},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"qwen3-235b-a22b-thinking-2507":{"id":"qwen3-235b-a22b-thinking-2507","name":"Qwen3 235B A22B Thinking (2507)","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-08","last_updated":"2025-07-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":8192},"cost":{"input":0.3,"output":3}},"o1":{"id":"o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"qwen3-coder-480b-a35b-instruct":{"id":"qwen3-coder-480b-a35b-instruct","name":"Qwen3-Coder 480B-A35B Instruct","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.38,"output":1.55},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"qwen3.5-9b":{"id":"qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.1,"output":0.15},"canonical_model_id":"alibaba/qwen3.5-9b"},"grok-4-5":{"id":"grok-4-5","name":"Grok 4.5","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3},"canonical_model_id":"xai/grok-4.5"},"gpt-5.1-codex-mini":{"id":"gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"minimax-m2.5":{"id":"minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":228700,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5"},"llama-3.2-11b-instruct":{"id":"llama-3.2-11b-instruct","name":"Llama 3.2 11B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.07,"output":0.33}},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"minimax-m3.1-flash":{"id":"minimax-m3.1-flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":512000},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625,"tiers":[{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5}},"canonical_model_id":"alibaba/qwen3.6-plus"},"llama-3-70b-instruct":{"id":"llama-3-70b-instruct","name":"Llama 3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2024-04-18","last_updated":"2024-04-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8000},"cost":{"input":0.51,"output":0.74}},"qwen3-vl-235b-a22b-instruct":{"id":"qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"cost":{"input":0.2,"output":0.88,"cache_read":0.11},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"grok-4":{"id":"grok-4","name":"Grok 4","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":3,"output":15,"cache_read":0.75}},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3-flash-preview"},"glm-5":{"id":"glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131072},"cost":{"input":0.573,"output":2.58,"cache_read":0.144,"cache_write":0},"canonical_model_id":"zhipuai/glm-5"},"grok-4-1-fast-non-reasoning":{"id":"grok-4-1-fast-non-reasoning","name":"Grok 4.1 Fast Non-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"cost":{"input":0.2,"output":0.5,"cache_read":0.05}},"glm-4.5-airx":{"id":"glm-4.5-airx","name":"GLM-4.5 AirX","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.1,"output":4.5,"cache_read":0.22}},"ernie-4.5-vl-424b-a47b":{"id":"ernie-4.5-vl-424b-a47b","name":"ERNIE 4.5 VL 424B A47B","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ernie","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-06-30","last_updated":"2025-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":123000,"output":123000},"cost":{"input":0.42,"output":1.25}},"grok-4-3":{"id":"grok-4-3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"fugu-ultra":{"id":"fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"sakana/fugu-ultra"},"minimax-m2.7-highspeed":{"id":"minimax-m2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"seed-1-8-251228":{"id":"seed-1-8-251228","name":"Seed 1.8 (251228)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"gemma-4-31b-it":{"id":"gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.1,"output":0.25,"cache_read":0.01},"canonical_model_id":"google/gemma-4-31b-it"},"hy4-preview":{"id":"hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"kimi-k3-fast":{"id":"kimi-k3-fast","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1040384,"output":131072},"cost":{"input":4.5,"output":22.5,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"seed-2-0-mini-260428":{"id":"seed-2-0-mini-260428","name":"Seed 2.0 Mini (260428)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.4,"cache_read":0.02}},"mistral-small-2506":{"id":"mistral-small-2506","name":"Mistral Small 3.2","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/mistral-small-2506"},"glm-4.6":{"id":"glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.55,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6"},"gpt-5.2-pro":{"id":"gpt-5.2-pro","name":"GPT-5.2 Pro","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"seed-1-6-250915":{"id":"seed-1-6-250915","name":"Seed 1.6 (250915)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"qwen-plus":{"id":"qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.4,"output":1.2,"reasoning":4,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen-plus"},"hy3":{"id":"hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":128000},"cost":{"input":0.132,"output":0.528,"cache_read":0.033},"canonical_model_id":"tencent/hy3"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.5,"output":2,"cache_read":0.15},"canonical_model_id":"zhipuai/glm-5.3"},"mistral-medium-3-5":{"id":"mistral-medium-3-5","name":"Mistral Medium 3.5","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"mistral/mistral-medium-3-5"},"glm-5v-turbo":{"id":"glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"qwen3-coder-flash":{"id":"qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":997952,"output":65536},"cost":{"input":0.3,"output":1.5,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"alibaba/qwen3-coder-flash"},"custom":{"id":"custom","name":"Custom Model","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0,"output":0}},"gpt-live-transcribe":{"id":"gpt-live-transcribe","type":"transcription","name":"GPT-Live-Transcribe","description":"Low-latency streaming speech-to-text model with keyword and language hints","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-07-28","last_updated":"2026-07-28","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-live-transcribe"},"kimi-k2":{"id":"kimi-k2","name":"Kimi K2","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.57,"output":2.3,"cache_read":0.5}},"inkling":{"id":"inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":1048576},"status":"deprecated","cost":{"input":0.95,"output":4.05,"cache_read":0.16},"canonical_model_id":"thinkingmachines/inkling"},"glm-4.6v":{"id":"glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.05},"canonical_model_id":"zhipuai/glm-4.6v"},"qwen-plus-latest":{"id":"qwen-plus-latest","name":"Qwen Plus Latest","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-01-25","last_updated":"2025-01-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":8192},"cost":{"input":0.4,"output":1.2,"cache_read":0.08,"cache_write":0.5}},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"glm-5.2-fast":{"id":"glm-5.2-fast","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2.156,"output":6.776,"cache_read":0.429},"canonical_model_id":"zhipuai/glm-5.2"},"sonar":{"id":"sonar","name":"Sonar","description":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","family":"sonar","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":130000,"output":4096},"cost":{"input":0.25,"output":2.5,"cache_write":0.25},"canonical_model_id":"perplexity/sonar"},"auto":{"id":"auto","name":"Auto Route","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-01-01","last_updated":"2024-01-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0,"output":0}},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"qwen3-235b-a22b-fp8":{"id":"qwen3-235b-a22b-fp8","name":"Qwen3 235B A22B FP8","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":8192},"cost":{"input":0.2,"output":0.8}},"qwen3.6-27b":{"id":"qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":65536},"cost":{"input":0.15,"output":1.35},"canonical_model_id":"alibaba/qwen3.6-27b"},"mistral-large-2512":{"id":"mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/mistral-large-2512"}}},"nvidia":{"id":"nvidia","env":["NVIDIA_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://integrate.api.nvidia.com/v1","name":"Nvidia","doc":"https://docs.api.nvidia.com/nim/","models":{"meta/esm2-650m":{"id":"meta/esm2-650m","name":"esm2-650m","description":"Protein language model that turns amino-acid sequences into embeddings","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-08-29","last_updated":"2025-03-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0,"output":0}},"meta/llama-3.1-8b-instruct":{"id":"meta/llama-3.1-8b-instruct","name":"Llama 3.1 8B Instruct","description":"Compact open Llama model for lightweight chat, drafting, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta/llama-3.2-3b-instruct":{"id":"meta/llama-3.2-3b-instruct","name":"Llama 3.2 3B Instruct","description":"Small instruction-tuned Llama model for multilingual dialogue and agentic retrieval","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-3.2-3b-instruct"},"meta/llama-3.2-11b-vision-instruct":{"id":"meta/llama-3.2-11b-vision-instruct","name":"Llama 3.2 11b Vision Instruct","description":"Open multimodal Llama model for image understanding, captioning, and visual QA","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-3.2-11b-vision-instruct"},"meta/llama-4-maverick-17b-128e-instruct":{"id":"meta/llama-4-maverick-17b-128e-instruct","name":"Llama 4 Maverick 17b 128e Instruct","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"meta/llama-guard-4-12b":{"id":"meta/llama-guard-4-12b","name":"Llama Guard 4 12B","description":"Multimodal Llama 4-based safety classifier for moderating prompts and responses","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":16384},"cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-guard-4-12b"},"meta/llama-3.1-70b-instruct":{"id":"meta/llama-3.1-70b-instruct","name":"Llama 3.1 70b Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-3.1-70b-instruct"},"meta/muse-glimmer-30b":{"id":"meta/muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"meta/muse-glimmer-30b"},"meta/llama-3.2-1b-instruct":{"id":"meta/llama-3.2-1b-instruct","name":"Llama 3.2 1b Instruct","description":"Compact instruction-tuned Llama model for multilingual dialogue and on-device assistants","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-3.2-1b-instruct"},"meta/llama-3.3-70b-instruct":{"id":"meta/llama-3.3-70b-instruct","name":"Llama 3.3 70b Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta/llama-3.2-90b-vision-instruct":{"id":"meta/llama-3.2-90b-vision-instruct","name":"Llama-3.2-90B-Vision-Instruct","description":"Large open multimodal Llama model for image reasoning, document QA, and captioning","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-3.2-90b-vision-instruct"},"meta/esmfold":{"id":"meta/esmfold","name":"esmfold","description":"Protein structure prediction from a single amino-acid sequence","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-03-15","last_updated":"2025-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0,"output":0}},"sarvamai/sarvam-m":{"id":"sarvamai/sarvam-m","name":"sarvam-m","description":"Hybrid-reasoning multilingual model for Indian languages built on Mistral Small","family":"sarvam","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"sarvam/sarvam-m"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"moonshotai/kimi-k3"},"mistralai/mistral-nemotron":{"id":"mistralai/mistral-nemotron","name":"mistral-nemotron","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"nemotron","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-06-11","last_updated":"2025-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/mistral-nemotron"},"mistralai/mistral-7b-instruct-v0.3":{"id":"mistralai/mistral-7b-instruct-v0.3","name":"Mistral-7B-Instruct-v0.3","description":"Open 7B Mistral instruction model with function calling","family":"mistral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-05-22","last_updated":"2024-05-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"mistral/mistral-7b-instruct-v0.3"},"mistralai/mixtral-8x22b-instruct":{"id":"mistralai/mixtral-8x22b-instruct","name":"Mistral: Mixtral 8x22B Instruct","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-01-31","release_date":"2024-04-17","last_updated":"2024-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":13108},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"mistral/mixtral-8x22b-instruct"},"mistralai/ministral-14b-instruct-2512":{"id":"mistralai/ministral-14b-instruct-2512","name":"Ministral 3 14B Instruct 2512","description":"Open vision-language model for efficient local deployment, instruction following, and tool use","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"mistral/ministral-3-14b-instruct-2512"},"mistralai/mistral-medium-3.5-128b":{"id":"mistralai/mistral-medium-3.5-128b","name":"Mistral Medium 3.5","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"mistral/mistral-medium-2604"},"mistralai/mixtral-8x7b-instruct":{"id":"mistralai/mixtral-8x7b-instruct","name":"Mistral: Mixtral 8x7B Instruct","description":"Open sparse mixture-of-experts Mistral instruction model","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2023-12-11","last_updated":"2023-12-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"mistral/mixtral-8x7b-instruct-v0.1"},"mistralai/mistral-large-3-675b-instruct-2512":{"id":"mistralai/mistral-large-3-675b-instruct-2512","name":"Mistral Large 3 675B Instruct 2512","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"mistral/mistral-large-2512"},"mistralai/mistral-small-4-119b-2603":{"id":"mistralai/mistral-small-4-119b-2603","name":"mistral-small-4-119b-2603","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"mistral/mistral-small-2603"},"mistralai/mistral-medium-3-instruct":{"id":"mistralai/mistral-medium-3-instruct","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"input":131072,"output":32768},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"mistral/mistral-medium-3"},"mistralai/magistral-small-2506":{"id":"mistralai/magistral-small-2506","name":"Magistral Small 2506","description":"Open Mistral reasoning model for transparent step-by-step problem solving","family":"magistral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"input":32768,"output":32768},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"mistral/magistral-small-2506"},"poolside/laguna-xs-2.1":{"id":"poolside/laguna-xs-2.1","name":"Laguna XS 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-02","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-xs-2.1"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":16384},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"thinkingmachines/inkling"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-oss-20b"},"openai/whisper-large-v3":{"id":"openai/whisper-large-v3","type":"transcription","name":"Whisper Large v3","description":"Open Whisper checkpoint for robust multilingual transcription and captioning","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2023-09","release_date":"2023-11-06","last_updated":"2023-11-06","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":448,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"openai/whisper-large-v3"},"black-forest-labs/flux.1-dev":{"id":"black-forest-labs/flux.1-dev","type":"image","name":"FLUX.1-dev","description":"Open-weight FLUX.1 text-to-image model for high-quality guidance-distilled generation","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-08","release_date":"2024-08-01","last_updated":"2024-08-01","modalities":{"input":["text"],"output":["image"]},"open_weights":true,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"black-forest-labs/flux-1-dev"},"black-forest-labs/flux_2-klein-4b":{"id":"black-forest-labs/flux_2-klein-4b","type":"image","name":"FLUX.2 Klein 4B","description":"Fast 4B open-weight image generation and editing model","family":"flux","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-06","release_date":"2026-01-15","last_updated":"2026-01-15","modalities":{"input":["text","image"],"output":["image"]},"open_weights":true,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"black-forest-labs/flux-2-klein-4b"},"black-forest-labs/flux_1-kontext-dev":{"id":"black-forest-labs/flux_1-kontext-dev","type":"image","name":"FLUX.1-Kontext-dev","description":"Open-weight FLUX.1 Kontext model for instruction-based image editing","family":"flux","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-05-29","last_updated":"2025-05-29","modalities":{"input":["text","image"],"output":["image"]},"open_weights":true,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"black-forest-labs/flux-1-kontext-dev"},"black-forest-labs/flux_1-schnell":{"id":"black-forest-labs/flux_1-schnell","type":"image","name":"FLUX.1-schnell","description":"Fast few-step FLUX.1 text-to-image model under Apache 2.0","family":"flux","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"knowledge":"2024-07","release_date":"2024-08-01","last_updated":"2024-08-01","modalities":{"input":["text"],"output":["image"]},"open_weights":true,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"black-forest-labs/flux-1-schnell"},"stepfun-ai/step-3.7-flash":{"id":"stepfun-ai/step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":16384},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"stepfun/step-3.7-flash"},"stepfun-ai/step-5-preview":{"id":"stepfun-ai/step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1048576,"output":1048576},"cost":{"input":0,"output":0},"canonical_model_id":"stepfun/step-5-preview"},"stepfun-ai/step-3.5-flash":{"id":"stepfun-ai/step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":16384},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"stepfun/step-3.5-flash"},"baai/bge-m3":{"id":"baai/bge-m3","type":"embedding","name":"BGE M3","description":"Multilingual embedding model combining dense, sparse, and multi-vector retrieval","family":"bge","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-01-30","last_updated":"2024-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":0},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"baai/bge-m3"},"upstage/solar-10.7b-instruct":{"id":"upstage/solar-10.7b-instruct","name":"solar-10.7b-instruct","description":"Upstage depth-upscaled 10.7B instruction model for single-turn chat","family":"solar","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2023-12-12","last_updated":"2023-12-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"upstage/solar-10.7b-instruct-v1.0"},"microsoft/phi-4-multimodal-instruct":{"id":"microsoft/phi-4-multimodal-instruct","name":"Phi 4 Multimodal","description":"Compact Microsoft multimodal model for text, image, and speech understanding","family":"phi","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-02-26","last_updated":"2025-02-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"input":128000,"output":16384},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"microsoft/phi-4-multimodal-instruct"},"microsoft/phi-4-mini-instruct":{"id":"microsoft/phi-4-mini-instruct","name":"Phi-4-Mini","description":"Compact Microsoft instruction model tuned for efficient coding assistance, reasoning, and low-latency agent tasks","family":"phi","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-02-26","last_updated":"2025-02-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"microsoft/phi-4-mini"},"nvidia/nemotron-3.5-super-vl-120b-a12b":{"id":"nvidia/nemotron-3.5-super-vl-120b-a12b","name":"Nemotron 3.5 Super VL 120B A12B","description":"NVIDIA vision-language reasoning model for text, image, and video understanding and agentic tool use","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/nemotron-3-content-safety":{"id":"nvidia/nemotron-3-content-safety","name":"nemotron-3-content-safety","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-content-safety"},"nvidia/nemotron-3-ultra-550b-a55b":{"id":"nvidia/nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]},{"type":"budget_tokens","min":-1,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/usdvalidate":{"id":"nvidia/usdvalidate","name":"usdvalidate","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-07-24","last_updated":"2025-01-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":0,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning":{"id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning","name":"Nemotron 3 Nano Omni","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":-1,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"nvidia/cosmos-transfer2_5-2b":{"id":"nvidia/cosmos-transfer2_5-2b","type":"video","name":"cosmos-transfer2.5-2b","description":"Generates physics-aware video world states from text and spatial control videos","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":true,"limit":{"context":0,"output":4096},"cost":{"input":0,"output":0}},"nvidia/llama-3.3-nemotron-super-49b-v1":{"id":"nvidia/llama-3.3-nemotron-super-49b-v1","name":"Llama 3.3 Nemotron Super 49B v1","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-03-18","last_updated":"2025-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/llama-3.3-nemotron-super-49b-v1"},"nvidia/streampetr":{"id":"nvidia/streampetr","name":"streampetr","description":"Temporal 3D object detection model for autonomous driving","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0}},"nvidia/nemotron-mini-4b-instruct":{"id":"nvidia/nemotron-mini-4b-instruct","name":"nemotron-mini-4b-instruct","description":"Compact Nemotron model for efficient reasoning and deployable AI agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-08-21","last_updated":"2024-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-mini-4b-instruct"},"nvidia/nemotron-content-safety-reasoning-4b":{"id":"nvidia/nemotron-content-safety-reasoning-4b","name":"nemotron-content-safety-reasoning-4b","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":false,"release_date":"2026-01-22","last_updated":"2026-01-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-content-safety-reasoning-4b"},"nvidia/llama-3.1-nemotron-70b-instruct":{"id":"nvidia/llama-3.1-nemotron-70b-instruct","name":"Llama 3.1 Nemotron 70B Instruct","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-04-15","last_updated":"2025-04-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/llama-3.1-nemotron-70b-instruct"},"nvidia/nemotron-3.5-lightning-30b-a3b":{"id":"nvidia/nemotron-3.5-lightning-30b-a3b","name":"Nemotron 3.5 Lightning 30B A3B","description":"Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/cosmos-predict1-5b":{"id":"nvidia/cosmos-predict1-5b","type":"video","name":"cosmos-predict1-5b","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-03-18","last_updated":"2025-03-18","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":true,"limit":{"context":0,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/nv-embed-v1":{"id":"nvidia/nv-embed-v1","type":"embedding","name":"nv-embed-v1","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-06-07","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":2048},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/nemotron-3-nano-30b-a3b":{"id":"nvidia/nemotron-3-nano-30b-a3b","name":"nemotron-3-nano-30b-a3b","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/llama-3.1-nemotron-nano-8b-v1":{"id":"nvidia/llama-3.1-nemotron-nano-8b-v1","name":"Llama 3.1 Nemotron Nano 8B v1","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-03-18","last_updated":"2025-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/llama-3_2-nemoretriever-300m-embed-v1":{"id":"nvidia/llama-3_2-nemoretriever-300m-embed-v1","type":"embedding","name":"llama-3_2-nemoretriever-300m-embed-v1","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-07-24","last_updated":"2025-07-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":2048},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/nemotron-nano-12b-v2-vl":{"id":"nvidia/nemotron-nano-12b-v2-vl","name":"Nemotron Nano 12B v2 VL","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-10-28","last_updated":"2025-10-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-nano-12b-v2-vl"},"nvidia/riva-translate-4b-instruct-v1.1":{"id":"nvidia/riva-translate-4b-instruct-v1.1","name":"riva-translate-4b-instruct-v1_1","description":"Translation model for multilingual conversion, localization, and cross-language workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-12","last_updated":"2025-12-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/rerank-qa-mistral-4b":{"id":"nvidia/rerank-qa-mistral-4b","type":"reranking","name":"rerank-qa-mistral-4b","description":"Reranking model for improving retrieval quality in search and recommendation systems","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-03-17","last_updated":"2025-01-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/cosmos-transfer1-7b":{"id":"nvidia/cosmos-transfer1-7b","type":"video","name":"cosmos-transfer1-7b","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-13","last_updated":"2025-06-30","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":true,"limit":{"context":0,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/nvidia-nemotron-nano-9b-v2":{"id":"nvidia/nvidia-nemotron-nano-9b-v2","name":"nvidia-nemotron-nano-9b-v2","description":"Compact Nemotron model for efficient reasoning and deployable AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-09","release_date":"2025-08-18","last_updated":"2025-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-nano-9b-v2"},"nvidia/sparsedrive":{"id":"nvidia/sparsedrive","name":"sparsedrive","description":"End-to-end autonomous driving stack for perception, prediction, and planning","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-03-18","last_updated":"2025-07-20","modalities":{"input":["video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0}},"nvidia/llama-3.1-nemotron-safety-guard-8b-v3":{"id":"nvidia/llama-3.1-nemotron-safety-guard-8b-v3","name":"llama-3.1-nemotron-safety-guard-8b-v3","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-28","last_updated":"2025-10-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/llama-3.1-nemotron-safety-guard-8b-v3"},"nvidia/llama-3.1-nemotron-ultra-253b-v1":{"id":"nvidia/llama-3.1-nemotron-ultra-253b-v1","name":"Llama 3.1 Nemotron Ultra 253B","description":"Flagship Nemotron model for high-throughput reasoning and complex agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-04-07","last_updated":"2025-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/llama-3.1-nemotron-ultra-253b"},"nvidia/llama-3.3-nemotron-super-49b-v1.5":{"id":"nvidia/llama-3.3-nemotron-super-49b-v1.5","name":"Llama 3.3 Nemotron Super 49B v1.5","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"nvidia/llama-3.3-nemotron-super-49b-v1.5"},"nvidia/gliner-pii":{"id":"nvidia/gliner-pii","name":"gliner-pii","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/bevformer":{"id":"nvidia/bevformer","name":"bevformer","description":"Multi-frame bird's-eye-view 3D perception transformer for autonomous driving","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-03-18","last_updated":"2025-07-20","modalities":{"input":["video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0}},"nvidia/llama-nemotron-embed-vl-1b-v2":{"id":"nvidia/llama-nemotron-embed-vl-1b-v2","type":"embedding","name":"llama-nemotron-embed-vl-1b-v2","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"nemotron","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":2048},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/llama-nemotron-embed-vl-1b-v2"},"nvidia/nemotron-voicechat":{"id":"nvidia/nemotron-voicechat","name":"nemotron-voicechat","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-voicechat"},"nvidia/llama-3.1-nemotron-nano-vl-8b-v1":{"id":"nvidia/llama-3.1-nemotron-nano-vl-8b-v1","name":"Llama 3.1 Nemotron Nano VL 8B v1","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-04-10","last_updated":"2025-04-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/cosmos-reason2-8b":{"id":"nvidia/cosmos-reason2-8b","name":"Cosmos Reason2 8B","description":"Vision language model for physical-world understanding with structured reasoning on video and images","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/nemotron-3.5-content-safety":{"id":"nvidia/nemotron-3.5-content-safety","name":"Nemotron 3.5 Content Safety","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"nvidia/usdcode":{"id":"nvidia/usdcode","name":"usdcode","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-01-01","last_updated":"2026-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/llama-nemotron-rerank-vl-1b-v2":{"id":"nvidia/llama-nemotron-rerank-vl-1b-v2","type":"reranking","name":"llama-nemotron-rerank-vl-1b-v2","description":"Reranking model for improving retrieval quality in search and recommendation systems","family":"nemotron","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":10240,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/llama-nemotron-rerank-vl-1b-v2"},"nvidia/nv-embedcode-7b-v1":{"id":"nvidia/nv-embedcode-7b-v1","type":"embedding","name":"nv-embedcode-7b-v1","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-03-17","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":2048},"status":"deprecated","cost":{"input":0,"output":0}},"nvidia/magpie-tts-zeroshot":{"id":"nvidia/magpie-tts-zeroshot","type":"speech","name":"magpie-tts-zeroshot","description":"Expressive text-to-speech that clones a voice from a short audio sample","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-05-22","last_updated":"2025-06-12","modalities":{"input":["text","audio"],"output":["audio"]},"open_weights":true,"limit":{"context":0,"output":4096},"cost":{"input":0,"output":0}},"nvidia/studiovoice":{"id":"nvidia/studiovoice","name":"studiovoice","description":"Enhances noisy or reverberant speech recordings to studio-quality speech","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-10-03","last_updated":"2025-06-13","modalities":{"input":["audio"],"output":["audio"]},"open_weights":true,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}},"nvidia/synthetic-video-detector":{"id":"nvidia/synthetic-video-detector","name":"synthetic-video-detector","description":"Detects AI-generated (synthetic) video","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["video"],"output":["text"]},"open_weights":true,"limit":{"context":0,"output":4096},"cost":{"input":0,"output":0}},"nvidia/active-speaker-detection":{"id":"nvidia/active-speaker-detection","name":"Active Speaker Detection","description":"Detects and tracks which person is speaking across video frames","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["video"],"output":["text"]},"open_weights":true,"limit":{"context":0,"output":4096},"cost":{"input":0,"output":0}},"deepseek-ai/deepseek-v4-flash":{"id":"deepseek-ai/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek-ai/deepseek-v4-pro-0813":{"id":"deepseek-ai/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek-ai/deepseek-v4-pro":{"id":"deepseek-ai/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek-ai/deepseek-v4.1-flash":{"id":"deepseek-ai/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-ai/deepseek-v4-flash-0731":{"id":"deepseek-ai/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"bytedance/seed-oss-36b-instruct":{"id":"bytedance/seed-oss-36b-instruct","name":"ByteDance-Seed/Seed-OSS-36B-Instruct","description":"Open ByteDance Seed model for long-context reasoning, tool use, and agents","family":"seed","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-08-20","last_updated":"2025-08-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":262000},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"bytedance-seed/seed-oss-36b-instruct"},"qwen/qwen2.5-coder-32b-instruct":{"id":"qwen/qwen2.5-coder-32b-instruct","name":"Qwen2.5 Coder 32b Instruct","description":"Open coding-focused Qwen model for code generation, repair, and repository reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen2.5-coder-32b-instruct"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen-image-edit":{"id":"qwen/qwen-image-edit","type":"image","name":"Qwen Image Edit","description":"Image editing model with multilingual text editing and subject consistency","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2025-08-19","last_updated":"2025-08-19","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}},"qwen/qwen-image":{"id":"qwen/qwen-image","type":"image","name":"Qwen Image","description":"Text-to-image foundation model with multilingual text rendering","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5-397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":8192},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3-next-80b-a3b-instruct":{"id":"qwen/qwen3-next-80b-a3b-instruct","name":"Qwen3-Next-80B-A3B-Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"qwen/qwen3-coder-480b-a35b-instruct":{"id":"qwen/qwen3-coder-480b-a35b-instruct","name":"Qwen3 Coder 480B A35B Instruct","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":66536},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"abacusai/dracarys-llama-3.1-70b-instruct":{"id":"abacusai/dracarys-llama-3.1-70b-instruct","name":"dracarys-llama-3.1-70b-instruct","description":"Abacus.AI coding fine-tune of Llama 3.1 70B Instruct","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-08-14","last_updated":"2024-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"abacusai/dracarys-llama-3.1-70b-instruct"},"minimaxai/minimax-m2.7":{"id":"minimaxai/minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimaxai/minimax-m3":{"id":"minimaxai/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":16384},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"minimax/MiniMax-M3"},"google/gemma-2-2b-it":{"id":"google/gemma-2-2b-it","name":"Gemma 2 2b It","description":"Small open Gemma 2 instruction model for lightweight chat and on-device use","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2024-07-31","last_updated":"2024-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-2-2b-it"},"google/gemma-3-4b-it":{"id":"google/gemma-3-4b-it","name":"Gemma 3 4B IT","description":"Open multimodal Gemma instruction model for efficient text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-3-4b-it"},"google/gemma-3n-e2b-it":{"id":"google/gemma-3n-e2b-it","name":"Gemma 3n E2b It","description":"Efficient on-device Gemma 3n model for text, image, and audio understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-06-26","last_updated":"2025-06-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-3n-e2b-it"},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-3-12b-it"},"google/google-paligemma":{"id":"google/google-paligemma","name":"paligemma","description":"Vision-language model for answering questions about images and text","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2024-05-14","last_updated":"2024-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0}},"google/diffusiongemma-26b-a4b-it":{"id":"google/diffusiongemma-26b-a4b-it","name":"DiffusionGemma 26B A4B IT","description":"Gemma 4 based discrete text diffusion model for fast chat, reasoning, and tool use","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-09","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":250000,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/diffusiongemma-26b-a4b-it"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma-4-31B-IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemma-3n-e4b-it":{"id":"google/gemma-3n-e4b-it","name":"Gemma 3n E4b It","description":"Efficient on-device Gemma 3n model for text, image, and audio understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-06-26","last_updated":"2025-06-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-3n-e4b-it"}}},"agentrouter":{"id":"agentrouter","env":["AGENTROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://agentrouter.org/v1","name":"AgentRouter","doc":"https://agentrouter.org/docs/opencode.html","models":{"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://agentrouter.org/v1"},"canonical_model_id":"anthropic/claude-opus-4-8"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"canonical_model_id":"deepseek/deepseek-v4-flash"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"canonical_model_id":"openai/gpt-5.6-sol"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://agentrouter.org/v1"},"canonical_model_id":"anthropic/claude-opus-5"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"canonical_model_id":"zhipuai/glm-5.3"}}},"freemodel":{"id":"freemodel","env":["FREEMODEL_API_KEY"],"npm":"@ai-sdk/anthropic","api":"https://cc.freemodel.dev/v1","name":"FreeModel","doc":"https://freemodel.dev","models":{"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://api.freemodel.dev/v1"},"cost":{"input":1.75,"output":14,"cache_read":0.175,"cache_write":1.75},"canonical_model_id":"openai/gpt-5.3-codex"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://api.freemodel.dev/v1"},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":5},"canonical_model_id":"openai/gpt-5.5"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://api.freemodel.dev/v1"},"cost":{"input":2.5,"output":15,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.4"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://api.freemodel.dev/v1"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075,"cache_write":0.75},"canonical_model_id":"openai/gpt-5.4-mini"}}},"zai":{"id":"zai","env":["ZHIPU_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.z.ai/api/paas/v4","name":"Z.AI","doc":"https://docs.z.ai/guides/overview/pricing","models":{"glm-4.5-air":{"id":"glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.2,"output":1.1,"cache_read":0.03,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-air"},"glm-4.7-flashx":{"id":"glm-4.7-flashx","name":"GLM-4.7-FlashX","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.07,"output":0.4,"cache_read":0.01,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"glm-4.6v-flash":{"id":"glm-4.6v-flash","name":"GLM-4.6V-Flash","description":"Lightweight GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6v-flash"},"glm-4.5v":{"id":"glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"output":16384},"cost":{"input":0.6,"output":1.8,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5v"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7"},"glm-5-turbo":{"id":"glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24,"cache_write":0},"canonical_model_id":"zhipuai/glm-5-turbo"},"glm-4.5":{"id":"glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"glm-4.7-flash":{"id":"glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7-flash"},"glm-4.5-flash":{"id":"glm-4.5-flash","name":"GLM-4.5-Flash","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":98304},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-flash"},"glm-5.3-flashx":{"id":"glm-5.3-flashx","name":"GLM-5.3-FlashX","description":"High-speed GLM-5.3-Flash serving option for coding and agent workflows","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.37,"output":1.25,"cache_read":0.075,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5":{"id":"glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2,"cache_write":0},"canonical_model_id":"zhipuai/glm-5"},"glm-4.6":{"id":"glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"glm-5v-turbo":{"id":"glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24,"cache_write":0},"canonical_model_id":"zhipuai/glm-5v-turbo"},"glm-4.6v":{"id":"glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.05,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6v"}}},"model-oracle-ai":{"id":"model-oracle-ai","env":["MODEL_ORACLE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.modeloracle.com/api/v1","name":"Model Oracle AI","doc":"https://modeloracle.com/setup/","models":{"gpt-4.1-mini":{"id":"gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"canonical_model_id":"openai/gpt-4.1-mini"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"canonical_model_id":"openai/gpt-5.5"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"canonical_model_id":"openai/gpt-5.4"},"claude-haiku-4.5":{"id":"claude-haiku-4.5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"canonical_model_id":"anthropic/claude-haiku-4-5"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"canonical_model_id":"deepseek/deepseek-v4-pro"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"canonical_model_id":"anthropic/claude-fable-5"},"claude-opus-4.8":{"id":"claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"canonical_model_id":"anthropic/claude-opus-4-8"},"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"canonical_model_id":"openai/o4-mini"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"canonical_model_id":"zhipuai/glm-5.2"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"canonical_model_id":"anthropic/claude-sonnet-5"},"gpt-4.1":{"id":"gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"canonical_model_id":"openai/gpt-4.1"},"gpt-5":{"id":"gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"canonical_model_id":"openai/gpt-5"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"canonical_model_id":"openai/gpt-5.4-mini"},"auto":{"id":"auto","name":"Auto","description":"Model Oracle AI decision engine that selects and routes among configured coding-agent models","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-29","last_updated":"2026-07-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000}},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"canonical_model_id":"openai/gpt-5.4-nano"}}},"aimlapi":{"id":"aimlapi","env":["AIMLAPI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.aimlapi.com/v1","name":"AI/ML API","doc":"https://docs.aimlapi.com","models":{"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.71925,"output":5.84545,"cache_read":0.20631},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-glimmer-30b":{"id":"meta/muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.48139,"output":2.0631,"cache_read":0.055016},"canonical_model_id":"meta/muse-glimmer-30b"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.71925,"output":5.84545,"cache_read":0.20631},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.71925,"output":5.84545,"cache_read":0.20631},"canonical_model_id":"meta/muse-spark-1.2"},"sakana/fugu-ultra-v2":{"id":"sakana/fugu-ultra-v2","name":"Fugu Ultra v2","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-08-28","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":6.877,"output":41.262,"cache_read":0.6877},"canonical_model_id":"sakana/fugu-ultra-v2"},"sakana/fugu-max":{"id":"sakana/fugu-max","name":"Fugu Max","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.7508,"output":8.2524,"cache_read":0.34385},"canonical_model_id":"sakana/fugu-max"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":6.877,"output":41.262,"cache_read":0.6877},"canonical_model_id":"sakana/fugu-ultra"},"z-ai/glm-4.5v":{"id":"z-ai/glm-4.5v","name":"GLM 4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":0.82524,"output":2.47572,"cache_read":0.151294},"canonical_model_id":"zhipuai/glm-4.5v"},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM 5 Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":1.56,"output":5.2,"cache_read":0.312},"canonical_model_id":"zhipuai/glm-5-turbo"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.195,"output":0.65,"cache_read":0.039},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-4.7-flash":{"id":"z-ai/glm-4.7-flash","name":"GLM 4.7 Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"cost":{"input":0.171925,"output":0.6877,"cache_read":0.013754},"canonical_model_id":"zhipuai/glm-4.7-flash"},"z-ai/glm-5v-turbo":{"id":"z-ai/glm-5v-turbo","name":"GLM 5V Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.56,"output":5.2,"cache_read":0.312},"canonical_model_id":"zhipuai/glm-5v-turbo"},"z-ai/glm-4.6v":{"id":"z-ai/glm-4.6v","name":"GLM 4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.41262,"output":1.23786,"cache_read":0.075647},"canonical_model_id":"zhipuai/glm-4.6v"},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.82524,"output":3.4385,"cache_read":0.20631},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 0905","description":"September 2025 Kimi K2 instruction update with a 256K context and improved agentic coding","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":1.3754,"output":4.1262},"canonical_model_id":"moonshotai/kimi-k2-0905"},"moonshotai/kimi-k2":{"id":"moonshotai/kimi-k2","name":"Kimi K2 0711","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-12-31","release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.783978,"output":3.16342},"canonical_model_id":"moonshotai/kimi-k2"},"mistralai/devstral-2512":{"id":"mistralai/devstral-2512","name":"Devstral 2 2512","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.605176,"output":3.02588,"cache_read":0.060518},"canonical_model_id":"mistral/devstral-2512"},"mistralai/mistral-saba":{"id":"mistralai/mistral-saba","name":"Saba","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-02-17","last_updated":"2025-02-17","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":26214},"cost":{"input":0.27508,"output":0.82524,"cache_read":0.027508},"canonical_model_id":"mistral/mistral-saba"},"mistralai/ministral-3b-2512":{"id":"mistralai/ministral-3b-2512","name":"Ministral 3 3B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":104857},"cost":{"input":0.20631,"output":0.20631,"cache_read":0.013754},"canonical_model_id":"mistral/ministral-3b-2512"},"mistralai/mixtral-8x22b-instruct":{"id":"mistralai/mixtral-8x22b-instruct","name":"Mixtral 8x22B Instruct","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-01-31","release_date":"2024-04-17","last_updated":"2024-04-17","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":52428},"cost":{"input":2.7508,"output":8.2524,"cache_read":0.27508},"canonical_model_id":"mistral/mixtral-8x22b-instruct"},"mistralai/mistral-medium-3":{"id":"mistralai/mistral-medium-3","name":"Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":0.55016,"output":2.7508,"cache_read":0.055016},"canonical_model_id":"mistral/mistral-medium-3"},"mistralai/mistral-small-3.1-24b-instruct":{"id":"mistralai/mistral-small-3.1-24b-instruct","name":"Small 3.1 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"cost":{"input":0.482765,"output":0.763347},"canonical_model_id":"mistral/mistral-small-3.1-24b-instruct"},"mistralai/ministral-8b-2512":{"id":"mistralai/ministral-8b-2512","name":"Ministral 3 8B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.41262,"output":0.41262,"cache_read":0.020631},"canonical_model_id":"mistral/ministral-8b-2512"},"mistralai/mistral-small-2603":{"id":"mistralai/mistral-small-2603","name":"Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.257888,"output":1.03155,"cache_read":0.020631},"canonical_model_id":"mistral/mistral-small-2603"},"mistralai/voxtral-small-24b-2507":{"id":"mistralai/voxtral-small-24b-2507","name":"Voxtral Small 24B 2507","description":"Open audio-language model for speech transcription, audio understanding, and voice-driven tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32768},"cost":{"input":0.13754,"output":0.41262,"cache_read":0.013754},"canonical_model_id":"mistral/voxtral-small-24b-2507"},"mistralai/codestral-2508":{"id":"mistralai/codestral-2508","name":"Codestral 2508","description":"Mistral code model for low-latency completion, fill-in-the-middle, and code generation","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":8192},"cost":{"input":0.41262,"output":1.23786,"cache_read":0.041262},"canonical_model_id":"mistral/codestral-2508"},"mistralai/mistral-medium-3.1":{"id":"mistralai/mistral-medium-3.1","name":"Medium 3.1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-30","release_date":"2025-08-12","last_updated":"2025-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":0.55016,"output":2.7508,"cache_read":0.055016},"canonical_model_id":"mistral/mistral-medium-3.1"},"mistralai/ministral-14b-2512":{"id":"mistralai/ministral-14b-2512","name":"Ministral 3 14B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.48139,"output":0.48139,"cache_read":0.027508},"canonical_model_id":"mistral/ministral-14b-2512"},"mistralai/mistral-nemo":{"id":"mistralai/mistral-nemo","name":"Mistral NeMo","description":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.20631,"output":0.233818,"cache_read":0.20631},"canonical_model_id":"mistral/mistral-nemo"},"mistralai/mistral-large-2407":{"id":"mistralai/mistral-large-2407","name":"Large 2407","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-03-31","release_date":"2024-07-24","last_updated":"2024-07-24","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":104857},"cost":{"input":2.7508,"output":8.2524,"cache_read":0.27508},"canonical_model_id":"mistral/mistral-large-2407"},"mistralai/mistral-large":{"id":"mistralai/mistral-large","name":"Large","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11-30","release_date":"2024-02-26","last_updated":"2024-02-26","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":102400},"cost":{"input":2.7508,"output":8.2524,"cache_read":0.27508},"canonical_model_id":"mistral/mistral-large"},"mistralai/mistral-small-3.2-24b-instruct":{"id":"mistralai/mistral-small-3.2-24b-instruct","name":"Small 3.2 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.103155,"output":0.27508},"canonical_model_id":"mistral/mistral-small-3.2-24b-instruct"},"mistralai/mistral-medium-3-5":{"id":"mistralai/mistral-medium-3-5","name":"Medium 3.5","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":209715},"cost":{"input":2.0631,"output":10.3155},"canonical_model_id":"mistral/mistral-medium-3-5"},"mistralai/mistral-small-24b-instruct-2501":{"id":"mistralai/mistral-small-24b-instruct-2501","name":"Small 3","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-01-30","last_updated":"2025-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.06877,"output":0.110032},"canonical_model_id":"mistral/mistral-small-24b-instruct-2501"},"bytedance-seed/seed-1.6-flash":{"id":"bytedance-seed/seed-1.6-flash","name":"Seed 1.6 Flash","description":"Low-latency ByteDance Seed model for high-throughput chat, extraction, and lightweight tool use","family":"seed","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.103155,"output":0.41262},"canonical_model_id":"bytedance-seed/seed-1-6-flash"},"bytedance-seed/seed-2.0-mini":{"id":"bytedance-seed/seed-2.0-mini","name":"Seed-2.0-Mini","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.13754,"output":0.55016},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"bytedance-seed/seed-2.0-code":{"id":"bytedance-seed/seed-2.0-code","name":"Seed-2.0-Code","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.6877,"output":4.1262},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"bytedance-seed/seed-2.0-lite":{"id":"bytedance-seed/seed-2.0-lite","name":"Seed-2.0-Lite","description":"Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.34385,"output":2.7508},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"gryphe/mythomax-l2-13b":{"id":"gryphe/mythomax-l2-13b","name":"MythoMax 13B","description":"Llama 2 merge combining MythoLogic and Huginn for creative writing and roleplay","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2023-08-10","last_updated":"2023-08-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":4096},"cost":{"input":2.578875,"output":2.578875,"cache_read":2.578875},"canonical_model_id":"gryphe/mythomax-l2-13b"},"undi95/remm-slerp-l2-13b":{"id":"undi95/remm-slerp-l2-13b","name":"ReMM SLERP 13B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-06-30","release_date":"2023-07-22","last_updated":"2023-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":6144,"output":4096},"cost":{"input":0.6877,"output":1.03155},"canonical_model_id":"undi95/remm-slerp-l2-13b"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":262144},"cost":{"input":0.797732,"output":1.980576,"cache_read":0.159546},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.3754,"output":5.57037,"cache_read":0.233818},"canonical_model_id":"thinkingmachines/inkling"},"cognitivecomputations/dolphin-mistral-24b-venice-edition":{"id":"cognitivecomputations/dolphin-mistral-24b-venice-edition","name":"Venice Uncensored","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04-30","release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.27508,"output":1.23786},"canonical_model_id":"cognitivecomputations/dolphin-mistral-24b-venice-edition"},"thedrummer/unslopnemo-12b":{"id":"thedrummer/unslopnemo-12b","name":"UnslopNemo 12B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04-30","release_date":"2024-11-08","last_updated":"2024-11-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.55016,"output":0.55016},"canonical_model_id":"thedrummer/unslopnemo-12b"},"thedrummer/skyfall-36b-v2":{"id":"thedrummer/skyfall-36b-v2","name":"Skyfall 36B V2","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2025-03-10","last_updated":"2025-03-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":0.75647,"output":1.10032,"cache_read":0.34385},"canonical_model_id":"thedrummer/skyfall-36b-v2"},"thedrummer/cydonia-24b-v4.1":{"id":"thedrummer/cydonia-24b-v4.1","name":"Cydonia 24B V4.1","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04-30","release_date":"2025-09-27","last_updated":"2025-09-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.41262,"output":0.6877,"cache_read":0.20631},"canonical_model_id":"thedrummer/cydonia-24b-v4.1"},"morph/morph-v3-large":{"id":"morph/morph-v3-large","name":"V3 Large","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-07","last_updated":"2025-07-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":1.23786,"output":2.61326},"canonical_model_id":"morph/morph-v3-large"},"morph/morph-v3-fast":{"id":"morph/morph-v3-fast","name":"V3 Fast","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-07","last_updated":"2025-07-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":81920,"output":38000},"cost":{"input":1.10032,"output":1.65048},"canonical_model_id":"morph/morph-v3-fast"},"aion-labs/aion-rp-llama-3.1-8b":{"id":"aion-labs/aion-rp-llama-3.1-8b","name":"Aion-RP 1.0 (8B)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12-31","release_date":"2025-02-04","last_updated":"2025-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":32768},"cost":{"input":1.10032,"output":2.20064},"canonical_model_id":"aion-labs/aion-rp-llama-3.1-8b"},"inception/mercury-2.5":{"id":"inception/mercury-2.5","name":"Mercury 2.5","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":260000,"output":65536},"cost":{"input":0.27508,"output":1.03155,"cache_read":0.027508},"canonical_model_id":"inception/mercury-2.5"},"inception/mercury-2":{"id":"inception/mercury-2","name":"Mercury 2","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":50000},"cost":{"input":0.34385,"output":1.03155,"cache_read":0.034385},"canonical_model_id":"inception/mercury-2"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.6,"output":13,"cache_read":0.26,"cache_write":3.25},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 Mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.52,"output":2.08,"cache_read":0.13},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-4o-2024-05-13":{"id":"openai/gpt-4o-2024-05-13","name":"GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-05-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":6.5,"output":19.5},"canonical_model_id":"openai/gpt-4o-2024-05-13"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":20.631,"output":165.048},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":13,"output":65,"cache_read":1.3,"cache_write":16.25},"canonical_model_id":"openai/gpt-6-astra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3 Mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.43,"output":5.72,"cache_read":0.715},"canonical_model_id":"openai/o3-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":41.262,"output":247.572},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":117964},"cost":{"input":0.103155,"output":0.41262,"cache_read":0.051578},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-4":{"id":"openai/gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":8192},"cost":{"input":39,"output":78,"cache_read":39},"canonical_model_id":"openai/gpt-4"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.6,"output":15.6,"cache_read":0.26,"cache_write":3.25},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.325,"output":2.6,"cache_read":0.0325},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":4.8139,"output":38.5112,"cache_read":0.48139},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/o1-pro":{"id":"openai/o1-pro","name":"o1-pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2025-03-19","last_updated":"2025-03-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":206.31,"output":825.24},"canonical_model_id":"openai/o1-pro"},"openai/gpt-5-2":{"id":"openai/gpt-5-2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":2.275,"output":18.2,"cache_read":0.2275},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5.4-image-2":{"id":"openai/gpt-5.4-image-2","type":"image","name":"GPT-5.4 Image 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":true,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["image","text","pdf"],"output":["image","text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":11.0032,"output":20.631,"cache_read":2.7508},"canonical_model_id":"openai/gpt-5.4-image-2"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":41.262,"output":247.572},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.065,"output":0.52,"cache_read":0.0065},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-oss-safeguard-20b":{"id":"openai/gpt-oss-safeguard-20b","name":"gpt-oss-safeguard-20b","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.103155,"output":0.41262,"cache_read":0.051578},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.13,"output":0.65,"cache_read":0.013,"cache_write":0.1625},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":117964},"cost":{"input":0.48139,"output":1.30663,"cache_read":0.48139},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5.2,"output":26,"cache_read":0.52,"cache_write":6.5},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5-1":{"id":"openai/gpt-5-1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.625,"output":13,"cache_read":0.1625},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.6,"output":13,"cache_read":0.13,"cache_write":3.25},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.26,"output":1.56,"cache_read":0.026,"cache_write":0.325},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":3.25,"output":13,"cache_read":1.625},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-5.6-luna-pro":{"id":"openai/gpt-5.6-luna-pro","name":"GPT-5.6 Luna Pro","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.27508,"output":1.65048,"cache_read":0.027508},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 Nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.13,"output":0.52,"cache_read":0.0325},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o Mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.195,"output":0.78,"cache_read":0.0975},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4o-2024-11-20":{"id":"openai/gpt-4o-2024-11-20","name":"GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":3.25,"output":13,"cache_read":1.625},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"openai/gpt-5-4":{"id":"openai/gpt-5-4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":3.25,"output":19.5,"cache_read":0.325},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-5.6-sol-pro":{"id":"openai/gpt-5.6-sol-pro","name":"GPT-5.6 Sol Pro","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5.5016,"output":27.508,"cache_read":0.55016},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-4o-2024-08-06":{"id":"openai/gpt-4o-2024-08-06","name":"GPT-4o","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":3.25,"output":13,"cache_read":1.625},"canonical_model_id":"openai/gpt-4o-2024-08-06"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":13,"output":39},"canonical_model_id":"openai/gpt-4-turbo"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":4096},"cost":{"input":0.65,"output":1.95},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2.6,"output":10.4,"cache_read":0.65},"canonical_model_id":"openai/gpt-4.1"},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":100000},"cost":{"input":19.5,"output":78,"cache_read":9.75},"canonical_model_id":"openai/o1"},"openai/gpt-5.1-codex-max":{"id":"openai/gpt-5.1-codex-max","name":"GPT-5.1-Codex-Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":3.4385,"output":27.508,"cache_read":0.34385},"canonical_model_id":"openai/gpt-5.1-codex-max"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.625,"output":13,"cache_read":0.1625},"canonical_model_id":"openai/gpt-5"},"openai/o3-pro":{"id":"openai/o3-pro","name":"o3 Pro","description":"High-effort o3 tier for difficult technical reasoning and careful answers","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":27.508,"output":110.032},"canonical_model_id":"openai/o3-pro"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2 Pro","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":28.8834,"output":231.0672},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/gpt-5-5":{"id":"openai/gpt-5-5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":6.5,"output":39,"cache_read":0.65},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 Mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":2.0631,"output":12.3786,"cache_read":0.20631},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.6-terra-pro":{"id":"openai/gpt-5.6-terra-pro","name":"GPT-5.6 Terra Pro","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.7508,"output":16.5048,"cache_read":0.27508},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 Nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.27508,"output":1.71925,"cache_read":0.027508},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.39,"output":1.56,"cache_read":0.0078},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":2.393196,"output":4.786392,"cache_read":0.199433},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-chat-v3-0324":{"id":"deepseek/deepseek-chat-v3-0324","name":"DeepSeek V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07-31","release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":124000},"cost":{"input":0.398866,"output":1.567956,"cache_read":0.192556},"canonical_model_id":"deepseek/deepseek-chat-v3-0324"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.572,"output":1.716,"cache_read":0.0182},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.17,"output":3.51,"cache_read":0.039},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-chat":{"id":"deepseek/deepseek-chat","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.39,"output":1.56,"cache_read":0.0078},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.39,"output":1.56,"cache_read":0.0078},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.6,"output":13,"cache_read":0.26,"cache_write":3.25},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude 4.7 Opus","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":6.877,"output":34.385,"cache_read":0.6877,"cache_write":8.59625},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude 4.6 Sonnet","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":4.1262,"output":20.631,"cache_read":0.41262,"cache_write":5.15775},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude 4.5 Haiku","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.3754,"output":6.877,"cache_read":0.13754,"cache_write":1.71925},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":13,"output":65,"cache_read":1.3,"cache_write":16.25},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":6.5,"output":32.5,"cache_read":0.65,"cache_write":8.125},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.2,"output":26,"cache_read":0.26,"cache_write":6.5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude 4.5 Opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":6.877,"output":34.385,"cache_read":0.6877,"cache_write":8.59625},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude 4.5 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":4.1262,"output":20.631,"cache_read":0.41262,"cache_write":5.15775},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude 4.8 Opus","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":6.877,"output":34.385,"cache_read":0.6877,"cache_write":8.59625},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.13,"output":0.65,"cache_read":0.013,"cache_write":0.1625},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.6,"output":13,"cache_read":0.26,"cache_write":3.25},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":13,"output":65,"cache_read":0.325,"cache_write":16.25},"canonical_model_id":"anthropic/claude-fable-5-1"},"nousresearch/hermes-3-llama-3.1-405b":{"id":"nousresearch/hermes-3-llama-3.1-405b","name":"Hermes 3 405B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-16","last_updated":"2024-08-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":1.3754,"output":1.3754},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-405b"},"nousresearch/hermes-3-llama-3.1-70b":{"id":"nousresearch/hermes-3-llama-3.1-70b","name":"Hermes 3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-18","last_updated":"2024-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.96278,"output":0.96278},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-70b"},"tencent/hy-mt2-30b-a3b":{"id":"tencent/hy-mt2-30b-a3b","name":"Hy-MT2-30B-A3B","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.10178,"output":0.405743},"canonical_model_id":"tencent/hy-mt2-30b-a3b"},"tencent/hy-mt2-7b":{"id":"tencent/hy-mt2-7b","name":"Hy-MT2-7B","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.10178,"output":0.405743},"canonical_model_id":"tencent/hy-mt2-7b"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":1.147084,"output":3.439875,"cache_read":0.057767},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.27508,"output":1.10032,"cache_read":0.06877},"canonical_model_id":"tencent/hy3"},"tencent/hy-mt2-1.8b":{"id":"tencent/hy-mt2-1.8b","name":"Hy-MT2-1.8B","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.060518,"output":0.243446},"canonical_model_id":"tencent/hy-mt2-1.8b"},"ibm-granite/granite-4.0-h-micro":{"id":"ibm-granite/granite-4.0-h-micro","name":"Granite 4.0 Micro","description":"Compact open-weight hybrid Granite model for lightweight enterprise chat and tool calling","family":"granite","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-02","last_updated":"2025-10-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":0.023382,"output":0.154045},"canonical_model_id":"ibm/granite-4-h-micro"},"meta-llama/Llama-3.3-70B-Instruct-Turbo":{"id":"meta-llama/Llama-3.3-70B-Instruct-Turbo","name":"Llama 3.3 70B Instruct Turbo","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":127000},"cost":{"input":1.352,"output":1.352},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.192556,"output":0.385112,"cache_read":0.003851},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2.5-pro":{"id":"xiaomi/mimo-v2.5-pro","name":"MiMo V2.5 Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.598299,"output":1.196598,"cache_read":0.004951},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"anthracite-org/magnum-v4-72b":{"id":"anthracite-org/magnum-v4-72b","name":"Magnum v4 72B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":16384},"cost":{"input":4.1262,"output":6.877},"canonical_model_id":"anthracite-org/magnum-v4-72b"},"stepfun/step-5-preview":{"id":"stepfun/step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":64000},"cost":{"input":1.3754,"output":3.71358,"cache_read":0.06877},"canonical_model_id":"stepfun/step-5-preview"},"moonshot/kimi-k3":{"id":"moonshot/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3.9,"output":19.5,"cache_read":0.39},"canonical_model_id":"moonshotai/kimi-k3"},"zhipu/glm-5.2":{"id":"zhipu/glm-5.2","name":"GLM 5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.82,"output":5.72,"cache_read":0.338},"canonical_model_id":"zhipuai/glm-5.2"},"zhipu/glm-5.3":{"id":"zhipu/glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":131072},"cost":{"input":1.82,"output":5.72,"cache_read":0.338},"canonical_model_id":"zhipuai/glm-5.3"},"amazon/nova-pro-v1":{"id":"amazon/nova-pro-v1","name":"Nova Pro 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":5120},"cost":{"input":1.10032,"output":4.40128},"canonical_model_id":"amazon/nova-pro-v1"},"amazon/nova-lite-v1":{"id":"amazon/nova-lite-v1","name":"Nova Lite 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":5120},"cost":{"input":0.082524,"output":0.330096},"canonical_model_id":"amazon/nova-lite-v1"},"amazon/nova-micro-v1":{"id":"amazon/nova-micro-v1","name":"Nova Micro 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":5120},"cost":{"input":0.048139,"output":0.192556},"canonical_model_id":"amazon/nova-micro-v1"},"perplexity/sonar-pro":{"id":"perplexity/sonar-pro","name":"Sonar Pro","description":"Deeper Sonar search model with broader retrieval and stronger synthesis","family":"sonar-pro","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":3.9,"output":19.5},"canonical_model_id":"perplexity/sonar-pro"},"perplexity/sonar":{"id":"perplexity/sonar","name":"Sonar","description":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","family":"sonar","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":100000},"cost":{"input":1.3,"output":1.3},"canonical_model_id":"perplexity/sonar"},"relace/relace-search":{"id":"relace/relace-search","name":"Search","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":1.3754,"output":4.1262},"canonical_model_id":"relace/relace-search"},"relace/relace-apply-3":{"id":"relace/relace-apply-3","name":"Apply 3","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2025-09-26","last_updated":"2025-09-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":1.16909,"output":1.71925},"canonical_model_id":"relace/relace-apply-3"},"upstage/solar-pro-3":{"id":"upstage/solar-pro-3","name":"Solar Pro 3","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2026-01-26","last_updated":"2026-03-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.20631,"output":0.82524,"cache_read":0.020631},"canonical_model_id":"upstage/solar-pro3"},"upstage/solar-pro4":{"id":"upstage/solar-pro4","name":"Solar Pro 4","description":"Upstage's flagship model, specialized for agentic use","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.041262,"output":0.165048,"cache_read":0.008252},"canonical_model_id":"upstage/solar-pro4"},"inference-net/schematron-v2-small":{"id":"inference-net/schematron-v2-small","name":"Schematron V2 Small","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":0.06877,"output":0.316342},"canonical_model_id":"inference-net/schematron-v2-small"},"inference-net/schematron-v2-turbo":{"id":"inference-net/schematron-v2-turbo","name":"Schematron V2 Turbo","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.041262,"output":0.20631},"canonical_model_id":"inference-net/schematron-v2-turbo"},"alibaba/qwen3.8-27b":{"id":"alibaba/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.61893,"output":4.40128},"canonical_model_id":"alibaba/qwen3.8-27b"},"alibaba/qwen-max":{"id":"alibaba/qwen-max","name":"Qwen Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32000,"output":8192},"cost":{"input":2.08,"output":8.32,"cache_read":0.416},"canonical_model_id":"alibaba/qwen-max"},"alibaba/qwen3.8-max-prime":{"id":"alibaba/qwen3.8-max-prime","name":"Qwen3.8 Max Prime","description":"High-throughput edition of Qwen3.8 Max for coding, professional work, multimodal understanding, and long-running agent workflows","family":"qwen3.8-max","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":5.5016,"output":16.5048,"cache_read":0.6877},"canonical_model_id":"alibaba/qwen3.8-max-prime"},"alibaba/qwen3.8-2.4t-a95b":{"id":"alibaba/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":52429},"cost":{"input":3.4385,"output":8.59625,"cache_read":0.6877},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"alibaba/qwen3.8-max":{"id":"alibaba/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.6,"output":7.8,"cache_read":0.325},"canonical_model_id":"alibaba/qwen3.8-max"},"alibaba/qwen3.8-flash":{"id":"alibaba/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.208,"output":0.611,"cache_read":0.0208},"canonical_model_id":"alibaba/qwen3.8-flash"},"alibaba/qwen3-next-80b-a3b-instruct":{"id":"alibaba/qwen3-next-80b-a3b-instruct","name":"Qwen3 Next 80B A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":129024,"output":16384},"cost":{"input":0.20631,"output":1.65048},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"alibaba/qwen3-coder-480b-a35b-instruct":{"id":"alibaba/qwen3-coder-480b-a35b-instruct","name":"Qwen3 Coder","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":65536},"cost":{"input":1.3754,"output":6.705075,"cache_read":1.3754},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"alibaba/qwen-plus":{"id":"alibaba/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131000,"output":16384},"cost":{"input":0.52,"output":1.56,"cache_read":0.104},"canonical_model_id":"alibaba/qwen-plus"},"mancer/weaver":{"id":"mancer/weaver","name":"Weaver (alpha)","description":"General-purpose chat model for instruction following, writing, and analysis","family":"alpha","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-06-30","release_date":"2023-08-02","last_updated":"2023-08-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":2000},"cost":{"input":1.03155,"output":1.3754},"canonical_model_id":"mancer/weaver"},"sao10k/l3.1-euryale-70b":{"id":"sao10k/l3.1-euryale-70b","name":"Llama 3.1 Euryale 70B v2.2","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-28","last_updated":"2024-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":1.99488,"output":1.99488},"canonical_model_id":"sao10k/l3.1-euryale-70b"},"sao10k/l3.3-euryale-70b":{"id":"sao10k/l3.3-euryale-70b","name":"Llama 3.3 Euryale 70B","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-12-18","last_updated":"2024-12-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.89401,"output":1.03155},"canonical_model_id":"sao10k/l3.3-euryale-70b"},"sao10k/l3-lunaris-8b":{"id":"sao10k/l3-lunaris-8b","name":"Llama 3 8B Lunaris","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-13","last_updated":"2024-08-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":7372},"cost":{"input":0.06877,"output":0.06877},"canonical_model_id":"sao10k/l3-lunaris-8b"},"cohere/command-r7b-12-2024":{"id":"cohere/command-r7b-12-2024","name":"Command R7B (12-2024)","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-12-13","last_updated":"2024-12-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.051578,"output":0.20631},"canonical_model_id":"cohere/command-r7b-12-2024"},"cohere/command-r-plus-08-2024":{"id":"cohere/command-r-plus-08-2024","name":"Command R+ (08-2024)","description":"Cohere's RAG workhorse for long-context enterprise search and tool use","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":3.4385,"output":13.754},"canonical_model_id":"cohere/command-r-plus-08-2024"},"cohere/command-a":{"id":"cohere/command-a","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"cost":{"input":3.4385,"output":13.754},"canonical_model_id":"cohere/command-a"},"cohere/command-r-08-2024":{"id":"cohere/command-r-08-2024","name":"Command R (08-2024)","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.20631,"output":0.82524},"canonical_model_id":"cohere/command-r-08-2024"},"unbiased/pareto":{"id":"unbiased/pareto","name":"Pareto","description":"Blended multimodal model that runs multiple language models in parallel and returns a synthesized answer","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":3.4385,"output":10.3155,"cache_read":0.34385},"canonical_model_id":"unbiased/pareto"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.975,"output":4.875,"cache_read":0.0975},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048756,"output":65536},"cost":{"input":2.7508,"output":16.5048,"cache_read":0.27508,"cache_write":0.515775},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemini-3.1-flash-lite-preview":{"id":"google/gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.34385,"output":2.0631,"cache_read":0.034385,"cache_write":0.114617},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B","description":"Largest open Gemma 3 instruction model for multilingual text generation and visual understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":117964},"cost":{"input":0.20631,"output":0.632684,"cache_read":0.103155},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemma-3-4b-it":{"id":"google/gemma-3-4b-it","name":"Gemma 3 4B","description":"Open multimodal Gemma instruction model for efficient text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":32768},"cost":{"input":0.06877,"output":0.13754,"cache_read":0.055016},"canonical_model_id":"google/gemma-3-4b-it"},"google/gemini-flash-latest":{"id":"google/gemini-flash-latest","name":"Gemini Flash Latest","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2.0631,"output":12.3786,"cache_read":0.20631,"cache_write":0.114617},"canonical_model_id":"google/gemini-flash-latest"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":2.6,"output":15.6,"cache_read":0.65},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"cost":{"input":0.20631,"output":0.82524,"cache_read":0.20631},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.95,"output":9.75,"cache_read":0.195},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.975,"output":4.875,"cache_read":0.0975},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B","description":"Open multimodal Gemma instruction model for multilingual text generation and image understanding","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":96000},"cost":{"input":0.06877,"output":0.20631,"cache_read":0.055016},"canonical_model_id":"google/gemma-3-12b-it"},"google/gemma-2-27b-it":{"id":"google/gemma-2-27b-it","name":"Gemma 2 27B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-07-13","last_updated":"2024-07-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":2048},"cost":{"input":0.89401,"output":0.89401},"canonical_model_id":"google/gemma-2-27b-it"},"google/gemini-3-8-flash":{"id":"google/gemini-3-8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.975,"output":4.875,"cache_read":0.0975},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B Instruct","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":1.361646,"output":2.049346,"cache_read":1.361646},"canonical_model_id":"google/gemma-4-31b-it"},"x-ai/grok-4-fast-non-reasoning":{"id":"x-ai/grok-4-fast-non-reasoning","name":"Grok 4 Fast Non-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":1999000},"cost":{"input":1.625,"output":3.25,"cache_read":0.26},"canonical_model_id":"xai/grok-4-fast-non-reasoning"},"x-ai/grok-4-20-0309-non-reasoning":{"id":"x-ai/grok-4-20-0309-non-reasoning","name":"Grok 4.20 Non-Reasoning","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":1999000},"cost":{"input":1.625,"output":3.25,"cache_read":0.26},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"rekaai/reka-edge":{"id":"rekaai/reka-edge","name":"Reka Edge","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"reka","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":0.13754,"output":0.13754},"canonical_model_id":"rekaai/reka-edge"}}},"alibaba-cn":{"id":"alibaba-cn","env":["DASHSCOPE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://dashscope.aliyuncs.com/compatible-mode/v1","name":"Alibaba (China)","doc":"https://www.alibabacloud.com/help/en/model-studio/models","models":{"qwen-vl-plus":{"id":"qwen-vl-plus","name":"Qwen-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-08-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.115,"output":0.287}},"qwen-math-turbo":{"id":"qwen-math-turbo","name":"Qwen Math Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09-19","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4096,"output":3072},"status":"deprecated","cost":{"input":0.287,"output":0.861}},"qwen2-5-32b-instruct":{"id":"qwen2-5-32b-instruct","name":"Qwen2.5 32B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.287,"output":0.861}},"kimi-k2-thinking":{"id":"kimi-k2-thinking","name":"Moonshot Kimi K2 Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"status":"deprecated","cost":{"input":0.574,"output":2.294}},"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.65,"output":4.951,"cache_read":0.33,"cache_write":2.0625},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen3-omni-flash":{"id":"qwen3-omni-flash","name":"Qwen3-Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":false,"limit":{"context":65536,"output":16384},"cost":{"input":0.258,"output":0.989,"input_audio":2.265,"output_audio":8.974}},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.276,"output":1.101,"cache_read":0.0276,"cache_write":0.345,"tiers":[{"input":0.826,"output":3.301,"cache_read":0.0826,"cache_write":1.0325,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.826,"output":3.301,"cache_read":0.0826,"cache_write":1.0325}},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen3-max":{"id":"qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-23","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.359,"output":1.434,"reasoning":1.434,"tiers":[{"input":0.574,"output":2.294,"reasoning":2.294,"tier":{"type":"context","size":32000}},{"input":1.004,"output":4.014,"reasoning":4.014,"tier":{"type":"context","size":128000}}]}},"qwen2-5-math-72b-instruct":{"id":"qwen2-5-math-72b-instruct","name":"Qwen2.5-Math 72B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":3072},"status":"deprecated","cost":{"input":0.574,"output":1.721}},"qwen3-14b":{"id":"qwen3-14b","name":"Qwen3 14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":38912}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.144,"output":0.574,"reasoning":1.434}},"qwen-max":{"id":"qwen-max","name":"Qwen Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.345,"output":1.377}},"deepseek-r1-0528":{"id":"deepseek-r1-0528","name":"DeepSeek R1 0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.574,"output":2.294}},"qwen-flash":{"id":"qwen-flash","name":"Qwen Flash","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.022,"output":0.216,"tiers":[{"input":0.087,"output":0.861,"tier":{"type":"context","size":128000}},{"input":0.173,"output":1.721,"tier":{"type":"context","size":256000}}]}},"moonshot-kimi-k2-instruct":{"id":"moonshot-kimi-k2-instruct","name":"Moonshot Kimi K2 Instruct","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.574,"output":2.294}},"qwen3-vl-235b-a22b":{"id":"qwen3-vl-235b-a22b","name":"Qwen3-VL 235B-A22B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.286705,"output":1.14682,"reasoning":2.867051}},"qwen3-coder-plus":{"id":"qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Hosted Qwen coder for software agents, repo edits, and long-context code","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2026-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.574,"output":2.296,"tiers":[{"input":0.861,"output":3.444,"tier":{"type":"context","size":32000}},{"input":1.435,"output":5.74,"tier":{"type":"context","size":128000}},{"input":2.87,"output":28.7,"tier":{"type":"context","size":256000}}]}},"qwq-32b":{"id":"qwq-32b","name":"QwQ 32B","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-12","last_updated":"2024-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.287,"output":0.861}},"qwen-vl-max":{"id":"qwen-vl-max","name":"Qwen-VL Max","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-08","last_updated":"2025-08-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.23,"output":0.574}},"qwen-long":{"id":"qwen-long","name":"Qwen Long","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-01-25","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":10000000,"output":8192},"cost":{"input":0.072,"output":0.287}},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"qwen3-omni-flash-realtime":{"id":"qwen3-omni-flash-realtime","type":"realtime","name":"Qwen3-Omni Flash Realtime","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":65536,"output":16384},"status":"deprecated","cost":{"input":0.23,"output":0.918,"input_audio":3.584,"output_audio":7.168}},"qwen3-vl-30b-a3b":{"id":"qwen3-vl-30b-a3b","name":"Qwen3-VL 30B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.108,"output":0.431,"reasoning":1.076}},"qwen3.5-flash":{"id":"qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-23","last_updated":"2026-09-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.029,"output":0.287,"reasoning":0.287,"tiers":[{"input":0.115,"output":1.147,"reasoning":1.147,"tier":{"type":"context","size":128000}},{"input":0.172,"output":1.72,"reasoning":1.72,"tier":{"type":"context","size":256000}}]}},"qwen-deep-research":{"id":"qwen-deep-research","name":"Qwen Deep Research","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01","last_updated":"2024-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":7.742,"output":23.367}},"qwen3-asr-flash":{"id":"qwen3-asr-flash","type":"transcription","name":"Qwen3-ASR Flash","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2024-04","release_date":"2025-09-08","last_updated":"2025-09-08","modalities":{"input":["audio"],"output":["text"]},"open_weights":false,"limit":{"context":53248,"output":4096}},"kimi-k2.6":{"id":"kimi-k2.6","name":"Moonshot Kimi K2.6","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.929,"output":3.858}},"qwen-omni-turbo":{"id":"qwen-omni-turbo","name":"Qwen-Omni Turbo","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-01-19","last_updated":"2025-03-26","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":false,"limit":{"context":32768,"output":2048},"status":"deprecated","cost":{"input":0.058,"output":0.23,"input_audio":3.584,"output_audio":7.168}},"deepseek-v3-2-exp":{"id":"deepseek-v3-2-exp","name":"DeepSeek V3.2 Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":65536},"status":"deprecated","cost":{"input":0.287,"output":0.431}},"qwen2-5-math-7b-instruct":{"id":"qwen2-5-math-7b-instruct","name":"Qwen2.5-Math 7B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":4096,"output":3072},"status":"deprecated","cost":{"input":0.144,"output":0.287}},"qwen3-vl-plus":{"id":"qwen3-vl-plus","name":"Qwen3-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.143353,"output":1.433525,"tiers":[{"input":0.215,"output":2.15,"tier":{"type":"context","size":32000}},{"input":0.43,"output":4.301,"tier":{"type":"context","size":128000}}]}},"qwen-plus-character":{"id":"qwen-plus-character","name":"Qwen Plus Character","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-01","last_updated":"2024-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":4096},"cost":{"input":0.115,"output":0.287}},"deepseek-r1-distill-qwen-14b":{"id":"deepseek-r1-distill-qwen-14b","name":"DeepSeek R1 Distill Qwen 14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":16384},"status":"deprecated","cost":{"input":0.144,"output":0.431}},"qwen-mt-turbo":{"id":"qwen-mt-turbo","name":"Qwen-MT Turbo","description":"Translation model for multilingual conversion, localization, and cross-language workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-01","last_updated":"2025-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":8192},"status":"deprecated","cost":{"input":0.101,"output":0.28}},"deepseek-v3":{"id":"deepseek-v3","name":"DeepSeek V3","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2024-12-01","last_updated":"2024-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":8192},"status":"deprecated","cost":{"input":0.287,"output":1.147}},"qwen3.6-max-preview":{"id":"qwen3.6-max-preview","name":"Qwen3.6 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":131072}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-20","last_updated":"2026-04-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":245800,"output":65536},"status":"deprecated","cost":{"input":1.32,"output":7.9,"cache_read":0.132}},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":1.65,"output":3.301},"canonical_model_id":"deepseek/deepseek-v4-pro"},"qwen2-5-14b-instruct":{"id":"qwen2-5-14b-instruct","name":"Qwen2.5 14B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.144,"output":0.431}},"deepseek-r1":{"id":"deepseek-r1","name":"DeepSeek R1","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.574,"output":2.294}},"qwen2-5-vl-7b-instruct":{"id":"qwen2-5-vl-7b-instruct","name":"Qwen2.5-VL 7B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.287,"output":0.717}},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":131072}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-14","last_updated":"2026-04-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":128000},"cost":{"input":0.825,"output":3.301,"cache_read":0.17,"tiers":[{"input":1.1,"output":3.851,"tier":{"type":"context","size":32000}}]}},"deepseek-v3-1":{"id":"deepseek-v3-1","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":65536},"status":"deprecated","cost":{"input":0.574,"output":1.721}},"deepseek-r1-distill-qwen-32b":{"id":"deepseek-r1-distill-qwen-32b","name":"DeepSeek R1 Distill Qwen 32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":16384},"status":"deprecated","cost":{"input":0.287,"output":0.861}},"deepseek-r1-distill-llama-8b":{"id":"deepseek-r1-distill-llama-8b","name":"DeepSeek R1 Distill Llama 8B","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":16384},"status":"deprecated","cost":{"input":0,"output":0}},"qwen3.7-flash":{"id":"qwen3.7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":131072},"cost":{"input":0.02962,"output":0.1185,"cache_read":0.002962,"cache_write":0.03703,"tiers":[{"input":0.08887,"output":0.35549,"cache_read":0.008887,"cache_write":0.11109,"tier":{"type":"context","size":32000}},{"input":0.17774,"output":0.71098,"cache_read":0.017774,"cache_write":0.22218,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen3-235b-a22b":{"id":"qwen3-235b-a22b","name":"Qwen3 235B-A22B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":38912}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.287,"output":1.147,"reasoning":2.868}},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.77744,"output":5.33231,"cache_read":0.22218,"cache_write":2.22179},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen-vl-ocr":{"id":"qwen-vl-ocr","name":"Qwen-VL OCR","description":"OCR model for extracting structured text from documents and screenshots","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-10-28","last_updated":"2026-09-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":34096,"output":4096},"cost":{"input":0.043,"output":0.072}},"qwen-mt-plus":{"id":"qwen-mt-plus","name":"Qwen-MT Plus","description":"Translation model for multilingual conversion, localization, and cross-language workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-01","last_updated":"2025-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":8192},"cost":{"input":0.259,"output":0.775}},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":131072}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.165,"output":0.99,"cache_read":0.0165,"cache_write":0.20625,"tiers":[{"input":0.66,"output":3.961,"cache_read":0.066,"cache_write":0.825,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.66,"output":3.961,"cache_read":0.066,"cache_write":0.825}},"canonical_model_id":"alibaba/qwen3.6-flash"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":2.827,"output":14.133,"cache_read":0.283},"canonical_model_id":"moonshotai/kimi-k3"},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.172,"output":1.032,"reasoning":1.032,"tiers":[{"input":0.43,"output":2.58,"reasoning":2.58,"tier":{"type":"context","size":128000}}]}},"qwen3-coder-30b-a3b-instruct":{"id":"qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.216,"output":0.861,"tiers":[{"input":0.323,"output":1.291,"tier":{"type":"context","size":32000}},{"input":0.538,"output":2.151,"tier":{"type":"context","size":128000}}]}},"qwen2-5-coder-7b-instruct":{"id":"qwen2-5-coder-7b-instruct","name":"Qwen2.5-Coder 7B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-11","last_updated":"2024-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.144,"output":0.287}},"qwen3-next-80b-a3b-thinking":{"id":"qwen3-next-80b-a3b-thinking","name":"Qwen3-Next 80B-A3B (Thinking)","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.144,"output":1.434}},"qwen3.8-omni-flash":{"id":"qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.113,"output":0.382,"cache_read":0.014},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"qwen-turbo":{"id":"qwen-turbo","name":"Qwen Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":38912}],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-11-01","last_updated":"2025-07-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":16384},"status":"deprecated","cost":{"input":0.044,"output":0.087,"reasoning":0.431}},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.11875,"output":0.40073,"cache_read":0.01187,"cache_write":0.14844},"canonical_model_id":"alibaba/qwen3.8-flash"},"deepseek-r1-distill-qwen-7b":{"id":"deepseek-r1-distill-qwen-7b","name":"DeepSeek R1 Distill Qwen 7B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":16384},"status":"deprecated","cost":{"input":0.072,"output":0.144}},"kimi-k2.5":{"id":"kimi-k2.5","name":"Moonshot Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.574,"output":3.011}},"qvq-max":{"id":"qvq-max","name":"QVQ Max","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"qvq","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-03-25","last_updated":"2025-03-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":1.147,"output":4.588}},"tongyi-intent-detect-v3":{"id":"tongyi-intent-detect-v3","name":"Tongyi Intent Detect V3","description":"General-purpose chat model for instruction following, writing, and analysis","family":"yi","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-01","last_updated":"2024-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1024},"cost":{"input":0.058,"output":0.144}},"qwen3-next-80b-a3b-instruct":{"id":"qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.144,"output":0.574}},"qwen3-32b":{"id":"qwen3-32b","name":"Qwen3 32B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":38912}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.287,"output":1.147,"reasoning":2.868}},"qwq-plus":{"id":"qwq-plus","name":"QwQ Plus","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-03-05","last_updated":"2025-03-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.23,"output":0.574}},"qwen3-8b":{"id":"qwen3-8b","name":"Qwen3 8B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":38912}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.072,"output":0.287,"reasoning":0.717}},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.1,"output":3.851,"cache_read":0.275,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"qwen-doc-turbo":{"id":"qwen-doc-turbo","name":"Qwen Doc Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01","last_updated":"2024-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.087,"output":0.144}},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.29754,"output":1.19015,"cache_read":0.01488},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-r1-distill-llama-70b":{"id":"deepseek-r1-distill-llama-70b","name":"DeepSeek R1 Distill Llama 70B","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":16384},"cost":{"input":0.287,"output":0.861}},"qwen2-5-vl-72b-instruct":{"id":"qwen2-5-vl-72b-instruct","name":"Qwen2.5-VL 72B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":2.294,"output":6.881}},"qwen3-coder-480b-a35b-instruct":{"id":"qwen3-coder-480b-a35b-instruct","name":"Qwen3-Coder 480B-A35B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04","last_updated":"2025-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.861,"output":3.441,"tiers":[{"input":1.291,"output":5.161,"tier":{"type":"context","size":32000}},{"input":2.151,"output":8.602,"tier":{"type":"context","size":128000}}]}},"qwen3.5-plus":{"id":"qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-09-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.115,"output":0.688,"reasoning":0.688,"tiers":[{"input":0.287,"output":1.72,"reasoning":1.72,"tier":{"type":"context","size":128000}},{"input":0.573,"output":3.44,"reasoning":3.44,"tier":{"type":"context","size":256000}}]}},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.276,"output":1.651,"cache_read":0.0276,"cache_write":0.345,"tiers":[{"input":1.101,"output":6.602,"cache_read":0.1101,"cache_write":1.37625,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.101,"output":6.602,"cache_read":0.1101,"cache_write":1.37625}}},"qwen-omni-turbo-realtime":{"id":"qwen-omni-turbo-realtime","type":"realtime","name":"Qwen-Omni Turbo Realtime","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-05-08","last_updated":"2025-05-08","modalities":{"input":["text","image","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":32768,"output":2048},"status":"deprecated","cost":{"input":0.23,"output":0.918,"input_audio":3.584,"output_audio":7.168}},"glm-5":{"id":"glm-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":32768}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":16384},"cost":{"input":0.573,"output":2.58,"tiers":[{"input":0.86,"output":3.154,"tier":{"type":"context","size":32000}}]}},"qwen2-5-72b-instruct":{"id":"qwen2-5-72b-instruct","name":"Qwen2.5 72B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.574,"output":1.721}},"qwen2-5-omni-7b":{"id":"qwen2-5-omni-7b","name":"Qwen2.5-Omni 7B","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-12","last_updated":"2024-12","modalities":{"input":["text","image","audio","video"],"output":["text","audio"]},"open_weights":true,"limit":{"context":32768,"output":2048},"status":"deprecated","cost":{"input":0.087,"output":0.345,"input_audio":5.448}},"qwen2-5-7b-instruct":{"id":"qwen2-5-7b-instruct","name":"Qwen2.5 7B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-09","last_updated":"2024-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.072,"output":0.144}},"qwen-plus":{"id":"qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.115,"output":0.287,"reasoning":1.147,"cache_read":0.012,"cache_write":0.144,"tiers":[{"input":0.345,"output":2.868,"reasoning":3.441,"tier":{"type":"context","size":128000}},{"input":0.689,"output":6.881,"reasoning":9.175,"tier":{"type":"context","size":256000}}]}},"qwen2-5-coder-32b-instruct":{"id":"qwen2-5-coder-32b-instruct","name":"Qwen2.5-Coder 32B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-11","last_updated":"2024-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"status":"deprecated","cost":{"input":0.287,"output":0.861}},"deepseek-r1-distill-qwen-1-5b":{"id":"deepseek-r1-distill-qwen-1-5b","name":"DeepSeek R1 Distill Qwen 1.5B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"release_date":"2025-01-01","last_updated":"2025-01-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":16384},"cost":{"input":0,"output":0}},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.1,"output":3.851,"cache_read":0.275,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"qwen3-coder-flash":{"id":"qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.144,"output":0.574,"tiers":[{"input":0.216,"output":0.861,"tier":{"type":"context","size":32000}},{"input":0.359,"output":1.434,"tier":{"type":"context","size":128000}},{"input":0.717,"output":3.584,"tier":{"type":"context","size":256000}}]}},"qwen-math-plus":{"id":"qwen-math-plus","name":"Qwen Math Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-08-16","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4096,"output":3072},"status":"deprecated","cost":{"input":0.574,"output":1.721}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2}},"siliconflow/deepseek-r1-0528":{"id":"siliconflow/deepseek-r1-0528","name":"siliconflow/deepseek-r1-0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":32768},"cost":{"input":0.5,"output":2.18}},"siliconflow/deepseek-v3.1-terminus":{"id":"siliconflow/deepseek-v3.1-terminus","name":"siliconflow/deepseek-v3.1-terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-29","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":65536},"cost":{"input":0.27,"output":1}},"siliconflow/deepseek-v3.2":{"id":"siliconflow/deepseek-v3.2","name":"siliconflow/deepseek-v3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-03","last_updated":"2025-12-03","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":65536},"cost":{"input":0.27,"output":0.42}},"siliconflow/deepseek-v3-0324":{"id":"siliconflow/deepseek-v3-0324","name":"siliconflow/deepseek-v3-0324","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-12-26","last_updated":"2025-11-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":163840,"output":163840},"cost":{"input":0.25,"output":1}},"MiniMax/MiniMax-M2.7":{"id":"MiniMax/MiniMax-M2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375}},"kimi/kimi-k2.5":{"id":"kimi/kimi-k2.5","name":"kimi/kimi-k2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3,"cache_read":0.1}}}},"vispark":{"id":"vispark","env":["VISPARK_LAB_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.lab.vispark.in/v1","name":"Vispark","doc":"https://lab.vispark.in/#vision","models":{"vispark/vision-medium":{"id":"vispark/vision-medium","name":"Vision Medium","description":"Balanced multimodal model pairing a 1M-token context window with deeper reasoning for analysis, content creation, and tool use across text, image, audio, video, and PDF inputs.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-05-15","last_updated":"2026-09","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":4.21,"output":12.63},"canonical_model_id":"vispark/vision-medium"},"vispark/vision-large":{"id":"vispark/vision-large","name":"Vision Large","description":"Most capable Vision model for complex reasoning, detailed media analysis, and structured output over a 1M-token context window.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-05-15","last_updated":"2026-09","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":7.37,"output":22.11},"canonical_model_id":"vispark/vision-large"},"vispark/vision-small":{"id":"vispark/vision-small","name":"Vision Small","description":"Fast, low-cost multimodal model for understanding text, images, audio, video, and PDFs, with tool calling and a 1M-token context window.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2024-05-15","last_updated":"2026-09","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.05,"output":3.16},"canonical_model_id":"vispark/vision-small"}}},"lucidquery":{"id":"lucidquery","env":["LUCIDQUERY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.lucidquery.com/v1","name":"LucidQuery","doc":"https://lucidquery.com/docs","models":{"lucidquery-agi-01-frontier":{"id":"lucidquery-agi-01-frontier","name":"AGI-01 Frontier","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"agi","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2026-06-05","release_date":"2026-06-16","last_updated":"2026-06-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":120000},"cost":{"input":5.04675,"output":24.673}},"lucidnova-rf1-100b":{"id":"lucidnova-rf1-100b","name":"LucidNova RF1 100B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"nova","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2025-09-16","release_date":"2024-12-28","last_updated":"2025-09-10","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":120000,"output":8000},"cost":{"input":2,"output":5}},"lucidquery-agi-01-swift":{"id":"lucidquery-agi-01-swift","name":"AGI-01 Swift","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"agi","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2026-06-05","release_date":"2026-06-16","last_updated":"2026-06-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":120000},"cost":{"input":2.80375,"output":16.8225}},"lucidquery-nexus-coder":{"id":"lucidquery-nexus-coder","name":"LucidQuery Nexus Coder","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"lucid","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":false,"knowledge":"2025-08-01","release_date":"2025-09-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":250000,"output":60000},"cost":{"input":2,"output":5}}}},"stepfun-step-plan":{"id":"stepfun-step-plan","env":["STEPFUN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.stepfun.com/step_plan/v1","name":"StepFun Step Plan (China)","doc":"https://platform.stepfun.com/docs/zh/step-plan/integrations/reasoning-api","models":{"step-3.7-flash":{"id":"step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"canonical_model_id":"stepfun/step-3.7-flash"},"step-5-preview":{"id":"step-5-preview","name":"Step 5 Preview","description":"StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"input":1000000,"output":65536},"canonical_model_id":"stepfun/step-5-preview"},"step-router-v1":{"id":"step-router-v1","name":"Step Router v1","description":"StepFun routing model that dispatches requests to the appropriate Step model.","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"input":256000,"output":256000}},"step-3.5-flash":{"id":"step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"canonical_model_id":"stepfun/step-3.5-flash"},"step-3.5-flash-2603":{"id":"step-3.5-flash-2603","name":"Step 3.5 Flash 2603","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","family":"step","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":256000,"output":256000},"canonical_model_id":"stepfun/step-3.5-flash-2603"}}},"tencent-tokenhub":{"id":"tencent-tokenhub","env":["TENCENT_TOKENHUB_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://tokenhub.tencentmaas.com/v1","name":"Tencent TokenHub","doc":"https://cloud.tencent.com/document/product/1823/130050","models":{"hy3-preview":{"id":"hy3-preview","name":"Hy3 preview","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"tencent/hy3-preview"},"hy4-preview":{"id":"hy4-preview","name":"Hy4 preview","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":64000},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"hy3":{"id":"hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"tencent/hy3"}}},"llmtr":{"id":"llmtr","env":["LLMTR_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://llmtr.com/v1","name":"LLMTR","doc":"https://llmtr.com/docs","models":{"muse-glimmer-30b-tr":{"id":"muse-glimmer-30b-tr","name":"Muse Glimmer 30B (TR)","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":2,"output":5,"cache_read":0.5},"canonical_model_id":"meta/muse-glimmer-30b"},"gemma-4":{"id":"gemma-4","name":"Gemma 4","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":2,"output":5,"cache_read":0.5},"canonical_model_id":"google/gemma-4-31b-it"},"qwen3-6-35b":{"id":"qwen3-6-35b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":16384},"cost":{"input":5,"output":10},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"medgemma-4b":{"id":"medgemma-4b","name":"MedGemma 4B","description":"Multimodal medical-domain Gemma variant for text and image analysis","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-04-26","last_updated":"2026-08-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":8192},"status":"deprecated","cost":{"input":3,"output":5}},"trendyol-asure-12b":{"id":"trendyol-asure-12b","name":"Trendyol Asure 12B","description":"Turkish-language multimodal instruct model built on Gemma 3 12B for e-commerce text, chat, and image-text tasks","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-02-19","last_updated":"2026-02-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":40960},"cost":{"input":0.1,"output":0.5,"cache_read":0.025},"canonical_model_id":"trendyol/asure-12b"},"magibu-11b-v8":{"id":"magibu-11b-v8","name":"Magibu 11B v8","description":"Turkish-language chat model for instruction following and assistant flows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-06-05","last_updated":"2026-08-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0.1,"output":0.5}},"meta/muse-spark-1.2-contributor":{"id":"meta/muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.002},"canonical_model_id":"meta/muse-spark-1.2"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"sakana/fugu-ultra"},"mistral/voxtral-small-latest":{"id":"mistral/voxtral-small-latest","name":"Voxtral Small (latest)","description":"Instruct model with native audio input for speech understanding and tool use","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/voxtral-small-latest"},"poolside/laguna-xs-2.1":{"id":"poolside/laguna-xs-2.1","name":"Laguna XS 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-02","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-xs-2.1"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.58,"output":1.44,"cache_read":0.116},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.87,"output":4.68,"cache_read":0.374},"canonical_model_id":"thinkingmachines/inkling"},"publicai/apertus-8b-instruct":{"id":"publicai/apertus-8b-instruct","name":"Apertus 8B Instruct","description":"Fully open 8B multilingual LLM supporting 1800+ languages with 65K context. Trained on compliant open data. Apache 2.0, EU AI Act compliant.","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":8192},"cost":{"input":0.1,"output":0.2},"canonical_model_id":"swiss-ai/apertus-8b"},"publicai/apertus-70b-instruct":{"id":"publicai/apertus-70b-instruct","name":"Apertus 70B Instruct","description":"Fully open 70B multilingual LLM supporting 1800+ languages with 65K context. Trained on 15T tokens of compliant open data. Apache 2.0, EU AI Act compliant.","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":8192},"cost":{"input":0.82,"output":2.92},"canonical_model_id":"swiss-ai/apertus-70b"},"perplexity/sonar-deep-research":{"id":"perplexity/sonar-deep-research","name":"Sonar Deep Research","description":"Sonar search model for autonomous research and citation-backed long-form reports","family":"sonar","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-02-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32768},"cost":{"input":2,"output":8,"reasoning":3},"canonical_model_id":"perplexity/sonar-deep-research"},"mimo/mimo-v2.5":{"id":"mimo/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0.14,"output":0.28},"canonical_model_id":"xiaomi/mimo-v2.5"},"mimo/mimo-v2.5-pro":{"id":"mimo/mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0.435,"output":0.87},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"upstage/solar-pro2":{"id":"upstage/solar-pro2","name":"Solar Pro 2","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":8192},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"upstage/solar-pro2"},"upstage/solar-pro4":{"id":"upstage/solar-pro4","name":"Solar Pro 4","description":"Upstage's flagship model, specialized for agentic use","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.15,"output":0.6,"cache_read":0.03},"canonical_model_id":"upstage/solar-pro4"},"upstage/solar-pro3":{"id":"upstage/solar-pro3","name":"Solar Pro 3","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-07","release_date":"2026-01-26","last_updated":"2026-03-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":8192},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"upstage/solar-pro3"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.4,"output":1.6,"cache_read":0.08},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3-max":{"id":"qwen/qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":65536},"status":"deprecated","cost":{"input":1.2,"output":6},"canonical_model_id":"alibaba/qwen3-max"},"qwen/qwen-flash":{"id":"qwen/qwen-flash","name":"Qwen Flash","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.05,"output":0.4,"cache_read":0.01},"canonical_model_id":"alibaba/qwen-flash"},"qwen/qwen3-coder-plus":{"id":"qwen/qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Hosted Qwen coder for software agents, repo edits, and long-context code","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"status":"deprecated","cost":{"input":1,"output":5},"canonical_model_id":"alibaba/qwen3-coder-plus"},"qwen/qwen3-vl-plus":{"id":"qwen/qwen3-vl-plus","name":"Qwen3-VL Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32768},"cost":{"input":0.2,"output":1.6,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3-vl-plus"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3.5-plus":{"id":"qwen/qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.4,"output":2.4},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen/qwen-plus":{"id":"qwen/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.4,"output":1.2,"cache_read":0.08},"canonical_model_id":"alibaba/qwen-plus"},"qwen/qwen3-coder-flash":{"id":"qwen/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":1.5,"cache_read":0.06},"canonical_model_id":"alibaba/qwen3-coder-flash"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.1,"output":0.1},"canonical_model_id":"google/gemini-2.5-flash-lite"}}},"zai-coding-plan":{"id":"zai-coding-plan","env":["ZHIPU_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.z.ai/api/coding/paas/v4","name":"Z.AI Coding Plan","doc":"https://docs.z.ai/devpack/overview","models":{"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7"},"glm-5.2-highspeed":{"id":"glm-5.2-highspeed","name":"GLM-5.2 Highspeed","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5-turbo":{"id":"glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5-turbo"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"glm-5.3-highspeed":{"id":"glm-5.3-highspeed","name":"GLM-5.3 Highspeed","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"}}},"alibaba-coding-plan":{"id":"alibaba-coding-plan","env":["ALIBABA_CODING_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://coding-intl.dashscope.aliyuncs.com/v1","name":"Alibaba Coding Plan","doc":"https://www.alibabacloud.com/help/en/model-studio/coding-plan","models":{"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen3-coder-plus":{"id":"qwen3-coder-plus","name":"Qwen3 Coder Plus","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-23","last_updated":"2025-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3-coder-next":{"id":"qwen3-coder-next","name":"Qwen3 Coder Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"glm-4.7":{"id":"glm-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3.5-plus":{"id":"qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"qwen3-max-2026-01-23":{"id":"qwen3-max-2026-01-23","name":"Qwen3 Max","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-23","last_updated":"2026-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"glm-5":{"id":"glm-5","name":"GLM-5","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-11","last_updated":"2026-02-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":16384},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"input":196601,"output":24576},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0}}}},"cohere":{"id":"cohere","env":["COHERE_API_KEY"],"npm":"@ai-sdk/cohere","name":"Cohere","doc":"https://docs.cohere.com/docs/models","models":{"command-a-reasoning-08-2025":{"id":"command-a-reasoning-08-2025","name":"Command A Reasoning","description":"Cohere reasoning model for multilingual enterprise agents, tools, and complex workflows","family":"command-a","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a-reasoning-08-2025"},"tiny-aya-global":{"id":"tiny-aya-global","name":"Tiny Aya Global","description":"Compact multilingual model balanced across 70 languages and regions","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"output":8000},"canonical_model_id":"cohere/tiny-aya-global"},"command-r7b-12-2024":{"id":"command-r7b-12-2024","name":"Command R7B","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-12-13","last_updated":"2024-12-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.0375,"output":0.15},"canonical_model_id":"cohere/command-r7b-12-2024"},"command-r-plus-08-2024":{"id":"command-r-plus-08-2024","name":"Command R+","description":"Cohere's RAG workhorse for long-context enterprise search and tool use","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-r-plus-08-2024"},"command-a-vision-07-2025":{"id":"command-a-vision-07-2025","name":"Command A Vision","description":"Cohere vision model for multilingual document analysis, OCR, and image understanding","family":"command-a","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a-vision-07-2025"},"north-mini-code-1-0":{"id":"north-mini-code-1-0","name":"North Mini Code","description":"Cohere coding model for practical software engineering and agentic edits","family":"north","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-09-23","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://api.cohere.ai/compatibility/v1"},"cost":{"input":0,"output":0},"canonical_model_id":"cohere/north-mini-code-1-0"},"tiny-aya-earth":{"id":"tiny-aya-earth","name":"Tiny Aya Earth","description":"Compact multilingual model specialized for West Asian and African languages","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"output":8000},"canonical_model_id":"cohere/tiny-aya-earth"},"command-r-08-2024":{"id":"command-r-08-2024","name":"Command R","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"cohere/command-r-08-2024"},"command-a-translate-08-2025":{"id":"command-a-translate-08-2025","name":"Command A Translate","description":"Translation model for multilingual conversion, localization, and cross-language workflows","family":"command-a","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"output":8000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a-translate-08-2025"},"north-small-translate-09-2026":{"id":"north-small-translate-09-2026","name":"North Small Translate","description":"Open-weight translation model for more than 50 languages and locale variants","family":"north","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-09-09","last_updated":"2026-09-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16000},"cost":{"input":0,"output":0},"canonical_model_id":"cohere/north-small-translate-1-0"},"c4ai-aya-expanse-32b":{"id":"c4ai-aya-expanse-32b","name":"Aya Expanse 32B","description":"Open multilingual model optimized for generation across 23 languages","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-10-24","last_updated":"2024-10-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"cohere/c4ai-aya-expanse-32b"},"c4ai-aya-vision-32b":{"id":"c4ai-aya-vision-32b","name":"Aya Vision 32B","description":"Open multilingual vision model for OCR, visual reasoning, and image question answering","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-04","last_updated":"2025-05-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":16000,"output":4000},"canonical_model_id":"cohere/c4ai-aya-vision-32b"},"tiny-aya-water":{"id":"tiny-aya-water","name":"Tiny Aya Water","description":"Compact multilingual model specialized for European and Asia-Pacific languages","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"output":8000},"canonical_model_id":"cohere/tiny-aya-water"},"tiny-aya-fire":{"id":"tiny-aya-fire","name":"Tiny Aya Fire","description":"Compact multilingual model specialized for South Asian languages","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-17","last_updated":"2026-02-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8000,"output":8000},"canonical_model_id":"cohere/tiny-aya-fire"},"command-a-03-2025":{"id":"command-a-03-2025","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a-03-2025"},"command-a-plus-05-2026":{"id":"command-a-plus-05-2026","name":"Command A Plus","description":"Cohere's stronger command model for multilingual agents and enterprise workflows","family":"command-a","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04-01","release_date":"2026-05-20","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":64000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a-plus-05-2026"},"command-r7b-arabic-02-2025":{"id":"command-r7b-arabic-02-2025","name":"Command R7B Arabic","description":"Open Command R model optimized for Arabic enterprise chat, RAG, and cultural knowledge","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2025-02-27","last_updated":"2025-02-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.0375,"output":0.15},"canonical_model_id":"cohere/command-r7b-arabic-02-2025"}}},"minimax-cn":{"id":"minimax-cn","env":["MINIMAX_API_KEY"],"npm":"@ai-sdk/anthropic","api":"https://api.minimax.cn/anthropic/v1","name":"MiniMax (minimax.cn)","doc":"https://platform.minimaxi.com/docs/guides/quickstart","models":{"MiniMax-M2":{"id":"MiniMax-M2","name":"MiniMax-M2","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3137,"output":1.2549,"cache_read":0.0314,"cache_write":0.3922}},"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":524288},"cost":{"input":0.3137,"output":1.2549,"cache_read":0.0627,"tiers":[{"input":0.6275,"output":2.5098,"cache_read":0.1255,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":0.6275,"output":2.5098,"cache_read":0.1255}}},"MiniMax-M2.7":{"id":"MiniMax-M2.7","name":"MiniMax-M2.7","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3137,"output":1.2549,"cache_read":0.0627,"cache_write":0.3922}},"MiniMax-M2.5-highspeed":{"id":"MiniMax-M2.5-highspeed","name":"MiniMax-M2.5-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.6275,"output":2.5098,"cache_read":0.0314,"cache_write":0.3922}},"MiniMax-M3.1-Flash":{"id":"MiniMax-M3.1-Flash","name":"MiniMax-M3.1-Flash","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0.1494,"output":0.5976,"cache_read":0.0149},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"MiniMax-M2.1":{"id":"MiniMax-M2.1","name":"MiniMax-M2.1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3137,"output":1.2549,"cache_read":0.0314,"cache_write":0.3922}},"MiniMax-M2.7-highspeed":{"id":"MiniMax-M2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.6275,"output":2.5098,"cache_read":0.0627,"cache_write":0.3922}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"status":"deprecated","cost":{"input":0.3137,"output":1.2549,"cache_read":0.0314,"cache_write":0.3922}}}},"opper":{"id":"opper","env":["OPPER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.opper.ai/v3/compat","name":"Opper","doc":"https://opper.ai/models","models":{"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.825,"output":4.125,"cache_read":0.0825},"canonical_model_id":"google/gemini-3.8-flash"},"devstral-2512":{"id":"devstral-2512","name":"Devstral 2","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"status":"deprecated","cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/devstral-2512"},"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.5"},"grok-4.6":{"id":"grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"gpt-6-astra":{"id":"gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"gpt-5.5-pro":{"id":"gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.5-pro"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"GPT OSS 20B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.11622,"output":0.488124},"canonical_model_id":"openai/gpt-oss-20b"},"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.5811,"output":3},"canonical_model_id":"alibaba/qwen3.8-27b"},"claude-opus-4-7":{"id":"claude-opus-4-7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-4-7"},"gpt-5.6-terra":{"id":"gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"gpt-5.3-codex":{"id":"gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"minimax-m2.7":{"id":"minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.69732,"output":2.78928},"canonical_model_id":"minimax/MiniMax-M2.7"},"claude-haiku-4-5":{"id":"claude-haiku-4-5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1.1,"output":5.5,"cache_read":0.11,"cache_write":1.375},"canonical_model_id":"anthropic/claude-haiku-4-5"},"muse-spark-1.3":{"id":"muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"claude-sonnet-4-5":{"id":"claude-sonnet-4-5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125,"tiers":[{"input":6.6,"output":24.75,"cache_read":0.66,"cache_write":8.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6.6,"output":24.75,"cache_read":0.66,"cache_write":8.25}},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"gpt-5.4-pro":{"id":"gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.248,"output":1.485},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen3.8-2.4t-a95b":{"id":"qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":2.5,"output":6,"cache_read":0.63},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.25,"output":0.66},"canonical_model_id":"deepseek/deepseek-v4-flash"},"qwen3-coder-next":{"id":"qwen3-coder-next","name":"Qwen3 Coder Next","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.5811,"output":2.3244},"canonical_model_id":"alibaba/qwen3-coder-next"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3.3,"output":16.5,"cache_read":0.33,"cache_write":4.125},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":1.78812,"output":3.57624},"canonical_model_id":"deepseek/deepseek-v4-pro"},"claude-fable-5":{"id":"claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":1.1622,"output":4.88124},"canonical_model_id":"openai/gpt-oss-120b"},"gpt-5.6-sol":{"id":"gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":45,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"openai/gpt-5.6-sol"},"mistral-small-2603":{"id":"mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.5811,"output":2.44062},"canonical_model_id":"mistral/mistral-small-2603"},"claude-fable-5-1":{"id":"claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"sonar-reasoning-pro":{"id":"sonar-reasoning-pro","name":"Sonar Reasoning Pro","description":"Web-grounded Sonar for multi-step research questions that need cited reasoning","family":"sonar-reasoning","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":2,"output":8},"canonical_model_id":"perplexity/sonar-reasoning-pro"},"gpt-5.3-chat-latest":{"id":"gpt-5.3-chat-latest","name":"GPT-5.3 Chat (latest)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"status":"deprecated","cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-chat-latest"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"kimi-k3":{"id":"kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15},"canonical_model_id":"moonshotai/kimi-k3"},"gpt-5.6-luna":{"id":"gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"claude-opus-5":{"id":"claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5.5,"output":27.5,"cache_read":0.55,"cache_write":6.875},"canonical_model_id":"anthropic/claude-opus-5"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"muse-spark-1.2":{"id":"muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"glm-5.3-flash":{"id":"glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.2,"output":0.5,"cache_read":0.07},"canonical_model_id":"zhipuai/glm-5.3-flash"},"minimax-m3":{"id":"minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.12,"tiers":[{"input":1.2,"output":4.8,"cache_read":0.24,"tier":{"type":"context","size":524288}}],"context_over_200k":{"input":1.2,"output":4.8,"cache_read":0.24}},"canonical_model_id":"minimax/MiniMax-M3"},"claude-opus-4-5":{"id":"claude-opus-4-5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.62708,"output":5.811},"canonical_model_id":"zhipuai/glm-5.2"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"google/gemini-3.5-flash"},"claude-sonnet-5":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":2.2,"output":11,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"anthropic/claude-sonnet-5"},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3-flash-preview"},"gemma-4-31b-it":{"id":"gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.46488,"output":2.44062},"canonical_model_id":"google/gemma-4-31b-it"},"grok-build-0.1":{"id":"grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2},"canonical_model_id":"xai/grok-build-0.1"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.75,"output":4.6488,"cache_read":0.44},"canonical_model_id":"zhipuai/glm-5.3"},"sonar-pro":{"id":"sonar-pro","name":"Sonar Pro","description":"Deeper Sonar search model with broader retrieval and stronger synthesis","family":"sonar-pro","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8000},"cost":{"input":3,"output":15},"canonical_model_id":"perplexity/sonar-pro"},"gpt-5.4-mini":{"id":"gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"sonar":{"id":"sonar","name":"Sonar","description":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","family":"sonar","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":1,"output":1},"canonical_model_id":"perplexity/sonar"},"gpt-5.4-nano":{"id":"gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"mistral-large-2512":{"id":"mistral-large-2512","name":"Mistral Large 3","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/mistral-large-2512"}}},"wallaby":{"id":"wallaby","env":["WALLABY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.wallabytoken.com/v1","name":"Wallaby","doc":"https://wallabytoken.com/docs","models":{"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":2.7,"output":13.5,"cache_read":0.27},"canonical_model_id":"moonshotai/kimi-k3"}}},"scaleway":{"id":"scaleway","env":["SCALEWAY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.scaleway.ai/v1","name":"Scaleway","doc":"https://www.scaleway.com/en/docs/generative-apis/","models":{"bge-multilingual-gemma2":{"id":"bge-multilingual-gemma2","type":"embedding","name":"BGE Multilingual Gemma2","description":"Multilingual Gemma 2 based embedding model for retrieval and semantic search","family":"bge","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2024-07-25","last_updated":"2024-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8191,"output":0},"cost":{"input":0.11215,"output":0},"canonical_model_id":"baai/bge-multilingual-gemma2"},"qwen3.8-27b":{"id":"qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"beta","cost":{"input":0.684,"output":3.762,"reasoning":3.762,"cache_read":0.137},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen3.6-35b-a3b":{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"status":"beta","cost":{"input":0.280375,"output":1.68225},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen3-embedding-8b":{"id":"qwen3-embedding-8b","type":"embedding","name":"Qwen3 Embedding 8B","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":0},"cost":{"input":0.11215,"output":0},"canonical_model_id":"alibaba/qwen3-embedding-8b"},"qwen3-235b-a22b-instruct-2507":{"id":"qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":260000,"output":16384},"cost":{"input":0.841125,"output":2.523375},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"lightonocr-3-0.8b":{"id":"lightonocr-3-0.8b","name":"LightOnOCR-3 0.8B","description":"LightOnOCR-3-0.8B is the lightest model of LightOn's LightOnOCR-3 family, with the same layout and visual understanding features as its larger siblings.","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-10-08","last_updated":"2026-10-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":2.464,"output":0},"canonical_model_id":"lightonai/lightonocr-3-0.8b"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"GPT-OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.168225,"output":0.6729},"canonical_model_id":"openai/gpt-oss-120b"},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5 397B A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.6729,"output":4.0374},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"mistral-medium-3.5-128b":{"id":"mistral-medium-3.5-128b","name":"Mistral Medium 3.5 128B","description":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-29","last_updated":"2026-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":1.68225,"output":8.41125},"canonical_model_id":"mistral/mistral-medium-2604"},"mistral-small-3.2-24b-instruct-2506":{"id":"mistral-small-3.2-24b-instruct-2506","name":"Mistral Small 3.2 24B Instruct (2506)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.168225,"output":0.392525},"canonical_model_id":"mistral/mistral-small-2506"},"llama-3.3-70b-instruct":{"id":"llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":100000,"output":16384},"cost":{"input":1.00935,"output":1.00935},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"gemma-4-26b-a4b-it":{"id":"gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"status":"beta","cost":{"input":0.280375,"output":0.56075},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"whisper-large-v3":{"id":"whisper-large-v3","type":"transcription","name":"Whisper Large v3","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"whisper","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2023-09","release_date":"2023-09-01","last_updated":"2026-03-17","modalities":{"input":["audio"],"output":["text"]},"open_weights":true,"limit":{"context":0,"output":8192}},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":2.0187,"output":6.16825},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"status":"beta","cost":{"input":0.672,"output":2.016,"reasoning":2.016,"cache_read":0.134},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":32768},"status":"beta","cost":{"input":0.468,"output":0.936,"reasoning":0.936,"cache_read":0.0936},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"status":"beta","cost":{"input":2.01,"output":6.15},"canonical_model_id":"zhipuai/glm-5.3"}}},"daoxe":{"id":"daoxe","env":["DAOXE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://daoxe.com/v1","name":"DaoXE","doc":"https://daoxe.com/pricing","models":{"claude-opus-4-8":{"id":"claude-opus-4-8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"grok-4.5":{"id":"grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.5"},"gpt-5.5":{"id":"gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"openai/gpt-5.5"},"claude-sonnet-4-6":{"id":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"gpt-5.4":{"id":"gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2},"canonical_model_id":"google/gemini-3.1-pro-preview"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":4,"output":16,"cache_read":0.8},"canonical_model_id":"moonshotai/kimi-k2.5"},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":5},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"grok-4.3":{"id":"grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"}}},"atomic-chat":{"id":"atomic-chat","env":["ATOMIC_CHAT_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"http://127.0.0.1:1337/v1","name":"Atomic Chat","doc":"https://atomic.chat","models":{"gemma-4-E4B-it-IQ4_XS":{"id":"gemma-4-E4B-it-IQ4_XS","name":"Gemma 4 E4B Instruct (IQ4_XS)","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-E4B-it"},"Qwen3_5-9B-MLX-4bit":{"id":"Qwen3_5-9B-MLX-4bit","name":"Qwen 3.5 9B (MLX 4-bit)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.5-9b"},"Qwen3_5-9B-Q4_K_M":{"id":"Qwen3_5-9B-Q4_K_M","name":"Qwen 3.5 9B (Q4_K_M)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"alibaba/qwen3.5-9b"},"gemma-4-E4B-it-MLX-4bit":{"id":"gemma-4-E4B-it-MLX-4bit","name":"Gemma 4 E4B Instruct (MLX 4-bit)","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-E4B-it"},"Meta-Llama-3_1-8B-Instruct-GGUF":{"id":"Meta-Llama-3_1-8B-Instruct-GGUF","name":"Meta Llama 3.1 8B Instruct (GGUF)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":4096},"cost":{"input":0,"output":0},"canonical_model_id":"meta/llama-3.1-8b-instruct"}}},"jiekou":{"id":"jiekou","env":["JIEKOU_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.jiekou.ai/openai","name":"Jiekou.AI","doc":"https://docs.jiekou.ai/docs/support/quickstart?utm_source=github_models.dev","models":{"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"gemini-2.5-flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.285,"output":2.375,"cache_read":0.0285},"canonical_model_id":"google/gemini-2.5-flash"},"gpt-5.1":{"id":"gpt-5.1","name":"gpt-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.1875,"output":9.5,"cache_read":0.1187},"canonical_model_id":"openai/gpt-5.1"},"grok-4-fast-non-reasoning":{"id":"grok-4-fast-non-reasoning","name":"grok-4-fast-non-reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"status":"deprecated","cost":{"input":0.18,"output":0.45},"canonical_model_id":"xai/grok-4-fast-non-reasoning"},"gpt-5-pro":{"id":"gpt-5-pro","name":"gpt-5-pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":14.25,"output":114},"canonical_model_id":"openai/gpt-5-pro"},"o3-mini":{"id":"o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.045,"output":4.18,"cache_read":0.5225},"canonical_model_id":"openai/o3-mini"},"gemini-2.5-flash-lite-preview-06-17":{"id":"gemini-2.5-flash-lite-preview-06-17","name":"gemini-2.5-flash-lite-preview-06-17","description":"First preview of Gemini 2.5 Flash-Lite, the low-cost, low-latency 2.5 model","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"status":"deprecated","cost":{"input":0.09,"output":0.36},"canonical_model_id":"google/gemini-2.5-flash-lite-preview-06-17"},"gpt-5-chat-latest":{"id":"gpt-5-chat-latest","name":"gpt-5-chat-latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","cost":{"input":1.125,"output":9},"canonical_model_id":"openai/gpt-5-chat-latest"},"gpt-5-mini":{"id":"gpt-5-mini","name":"gpt-5-mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2375,"output":1.9,"cache_read":0.0237},"canonical_model_id":"openai/gpt-5-mini"},"gpt-5.2-codex":{"id":"gpt-5.2-codex","name":"gpt-5.2-codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"gpt-5-nano":{"id":"gpt-5-nano","name":"gpt-5-nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.0475,"output":0.38,"cache_read":0.0047},"canonical_model_id":"openai/gpt-5-nano"},"gpt-5.2":{"id":"gpt-5.2","name":"gpt-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.6625,"output":13.3,"cache_read":0.1662},"canonical_model_id":"openai/gpt-5.2"},"claude-opus-4-1-20250805":{"id":"claude-opus-4-1-20250805","name":"claude-opus-4-1-20250805","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":13.5,"output":67.5},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"gpt-5-codex":{"id":"gpt-5-codex","name":"gpt-5-codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.1875,"output":9.5,"cache_read":0.1187},"canonical_model_id":"openai/gpt-5-codex"},"o3":{"id":"o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":9.5,"output":38,"cache_read":0.475},"canonical_model_id":"openai/o3"},"grok-4-1-fast-reasoning":{"id":"grok-4-1-fast-reasoning","name":"grok-4-1-fast-reasoning","description":"xAI's fast agentic tool-calling model with a 2M context window and built-in reasoning","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.19,"output":0.475},"canonical_model_id":"xai/grok-4.1-fast-reasoning"},"grok-4-fast-reasoning":{"id":"grok-4-fast-reasoning","name":"grok-4-fast-reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"status":"deprecated","cost":{"input":0.18,"output":0.45},"canonical_model_id":"xai/grok-4-fast-reasoning"},"claude-opus-4-20250514":{"id":"claude-opus-4-20250514","name":"claude-opus-4-20250514","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","cost":{"input":13.5,"output":67.5},"canonical_model_id":"anthropic/claude-opus-4-20250514"},"claude-opus-4-5-20251101":{"id":"claude-opus-4-5-20251101","name":"claude-opus-4-5-20251101","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":65536},"cost":{"input":4.75,"output":23.75,"cache_read":0.475,"cache_write":5.9375},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"grok-code-fast-1":{"id":"grok-code-fast-1","name":"grok-code-fast-1","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":10000},"cost":{"input":0.19,"output":1.425,"cache_read":0.019},"canonical_model_id":"xai/grok-code-fast-1"},"gemini-2.5-flash-lite-preview-09-2025":{"id":"gemini-2.5-flash-lite-preview-09-2025","name":"gemini-2.5-flash-lite-preview-09-2025","description":"September 2025 preview of Gemini 2.5 Flash-Lite with better instruction following and multimodal quality","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-09-25","last_updated":"2025-09-25","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.09,"output":0.36},"canonical_model_id":"google/gemini-2.5-flash-lite-preview-09-2025"},"claude-opus-4-6":{"id":"claude-opus-4-6","name":"claude-opus-4-6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"gemini-2.5-flash-preview-05-20":{"id":"gemini-2.5-flash-preview-05-20","name":"gemini-2.5-flash-preview-05-20","description":"Preview snapshot of Gemini 2.5 Flash with controllable thinking budgets","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1425,"output":3.325,"cache_read":0.0285},"canonical_model_id":"google/gemini-2.5-flash-preview-05-20"},"grok-4-0709":{"id":"grok-4-0709","name":"grok-4-0709","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-01","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":8192},"cost":{"input":2.85,"output":14.25}},"claude-sonnet-4-5-20250929":{"id":"claude-sonnet-4-5-20250929","name":"claude-sonnet-4-5-20250929","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"o4-mini":{"id":"o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4},"canonical_model_id":"openai/o4-mini"},"claude-haiku-4-5-20251001":{"id":"claude-haiku-4-5-20251001","name":"claude-haiku-4-5-20251001","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"gpt-5.1-codex-max":{"id":"gpt-5.1-codex-max","name":"gpt-5.1-codex-max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.1875,"output":9.5,"cache_read":0.1187},"canonical_model_id":"openai/gpt-5.1-codex-max"},"gpt-5.1-codex-mini":{"id":"gpt-5.1-codex-mini","name":"gpt-5.1-codex-mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2375,"output":1.9,"cache_read":0.0237},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"gemini-3-flash-preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.475,"output":2.85,"cache_read":0.0475},"canonical_model_id":"google/gemini-3-flash-preview"},"grok-4-1-fast-non-reasoning":{"id":"grok-4-1-fast-non-reasoning","name":"grok-4-1-fast-non-reasoning","description":"xAI's fast agentic tool-calling model with a 2M context window; non-reasoning variant for low-latency responses","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.19,"output":0.475},"canonical_model_id":"xai/grok-4.1-fast"},"claude-sonnet-4-20250514":{"id":"claude-sonnet-4-20250514","name":"claude-sonnet-4-20250514","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","cost":{"input":2.7,"output":13.5},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"gemini-2.5-pro-preview-06-05":{"id":"gemini-2.5-pro-preview-06-05","name":"gemini-2.5-pro-preview-06-05","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01-31","release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["image","text","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.1875,"output":9.5,"cache_read":0.1187},"canonical_model_id":"google/gemini-2.5-pro-preview"},"gpt-5.1-codex":{"id":"gpt-5.1-codex","name":"gpt-5.1-codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.1875,"output":9.5,"cache_read":0.1187},"canonical_model_id":"openai/gpt-5.1-codex"},"gpt-5.2-pro":{"id":"gpt-5.2-pro","name":"gpt-5.2-pro","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":19.95,"output":159.6},"canonical_model_id":"openai/gpt-5.2-pro"},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"gemini-2.5-flash-lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.095,"output":0.38,"cache_read":0.0095},"canonical_model_id":"google/gemini-2.5-flash-lite"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"gemini-2.5-pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":1.1875,"output":9.5,"cache_read":0.1187},"canonical_model_id":"google/gemini-2.5-pro"},"gemini-3-pro-preview":{"id":"gemini-3-pro-preview","name":"gemini-3-pro-preview","description":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":1.8,"output":10.8},"canonical_model_id":"google/gemini-3-pro-preview"},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 0905","description":"September 2025 Kimi K2 instruction update with a 256K context and improved agentic coding","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":100352},"cost":{"input":0.6,"output":2.5},"canonical_model_id":"moonshotai/kimi-k2-0905"},"moonshotai/kimi-k2-instruct":{"id":"moonshotai/kimi-k2-instruct","name":"Kimi K2 Instruct","description":"Non-thinking Kimi K2 MoE instruction model for coding and tool-using agents","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":100352},"cost":{"input":0.57,"output":2.3},"canonical_model_id":"moonshotai/kimi-k2-instruct"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"xiaomimimo/mimo-v2-flash":{"id":"xiaomimimo/mimo-v2-flash","name":"XiaomiMiMo/MiMo-V2-Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12-01","release_date":"2025-12-16","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.1,"output":0.3,"cache_read":0.02},"canonical_model_id":"xiaomi/mimo-v2-flash"},"deepseek/deepseek-v3.1":{"id":"deepseek/deepseek-v3.1","name":"DeepSeek V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.27,"output":1},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek/deepseek-r1-0528":{"id":"deepseek/deepseek-r1-0528","name":"DeepSeek R1 0528","description":"May 2025 DeepSeek R1 update with improved mathematics, coding, and reasoning","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.7,"output":2.5,"cache_read":0.35},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek/deepseek-v3-0324":{"id":"deepseek/deepseek-v3-0324","name":"DeepSeek V3 0324","description":"March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.28,"output":1.14,"cache_read":0.14},"canonical_model_id":"deepseek/deepseek-v3-0324"},"baidu/ernie-4.5-300b-a47b-paddle":{"id":"baidu/ernie-4.5-300b-a47b-paddle","name":"ERNIE 4.5 300B A47B","description":"Open ERNIE text MoE with 300B total and 47B active parameters","family":"ernie","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-06-30","last_updated":"2025-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":123000,"output":12000},"cost":{"input":0.28,"output":1.1},"canonical_model_id":"baidu/ernie-4.5-300b-a47b"},"baidu/ernie-4.5-vl-424b-a47b":{"id":"baidu/ernie-4.5-vl-424b-a47b","name":"ERNIE 4.5 VL 424B A47B","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ernie","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-06-30","last_updated":"2025-06-30","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":123000,"output":16000},"cost":{"input":0.42,"output":1.25},"canonical_model_id":"baidu/ernie-4.5-vl-424b-a47b"},"zai-org/glm-4.5v":{"id":"zai-org/glm-4.5v","name":"GLM 4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":0.6,"output":1.8},"canonical_model_id":"zhipuai/glm-4.5v"},"zai-org/glm-4.7":{"id":"zai-org/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2},"canonical_model_id":"zhipuai/glm-4.7"},"zai-org/glm-4.5":{"id":"zai-org/glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.6,"output":2.2},"canonical_model_id":"zhipuai/glm-4.5"},"zai-org/glm-4.7-flash":{"id":"zai-org/glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.07,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flash"},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"Minimax M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.1"},"qwen/qwen3-coder-next":{"id":"qwen/qwen3-coder-next","name":"qwen/qwen3-coder-next","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.2,"output":1.5},"canonical_model_id":"alibaba/qwen3-coder-next"},"qwen/qwen3-235b-a22b-instruct-2507":{"id":"qwen/qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.15,"output":0.8},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"qwen/qwen3-30b-a3b-fp8":{"id":"qwen/qwen3-30b-a3b-fp8","name":"Qwen3 30B A3B","description":"Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":20000},"cost":{"input":0.09,"output":0.45},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"qwen/qwen3-next-80b-a3b-thinking":{"id":"qwen/qwen3-next-80b-a3b-thinking","name":"Qwen3 Next 80B A3B Thinking","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.15,"output":1.5},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"qwen/qwen3-next-80b-a3b-instruct":{"id":"qwen/qwen3-next-80b-a3b-instruct","name":"Qwen3 Next 80B A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"cost":{"input":0.15,"output":1.5},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"qwen/qwen3-32b-fp8":{"id":"qwen/qwen3-32b-fp8","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":20000},"cost":{"input":0.1,"output":0.45},"canonical_model_id":"alibaba/qwen3-32b"},"qwen/qwen3-235b-a22b-thinking-2507":{"id":"qwen/qwen3-235b-a22b-thinking-2507","name":"Qwen3 235B A22b Thinking 2507","description":"Updated large open Qwen3 MoE thinking model for reasoning, math, coding, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.3,"output":3},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"qwen/qwen3-coder-480b-a35b-instruct":{"id":"qwen/qwen3-coder-480b-a35b-instruct","name":"Qwen3 Coder 480B A35B Instruct","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.38,"output":1.55},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"qwen/qwen3-235b-a22b-fp8":{"id":"qwen/qwen3-235b-a22b-fp8","name":"Qwen3 235B A22B","description":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":20000},"cost":{"input":0.2,"output":0.8},"canonical_model_id":"alibaba/qwen3-235b-a22b"},"minimaxai/minimax-m1-80k":{"id":"minimaxai/minimax-m1-80k","name":"MiniMax M1","description":"Open hybrid-attention MoE reasoning model with a million-token context and an 80K thinking budget","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":40000},"cost":{"input":0.55,"output":2.2},"canonical_model_id":"minimax/minimax-m1-80k"}}},"sarvam":{"id":"sarvam","env":["SARVAM_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.sarvam.ai/v1","name":"Sarvam AI","doc":"https://docs.sarvam.ai/api-reference-docs/getting-started/models","models":{"sarvam-105b":{"id":"sarvam-105b","name":"Sarvam-105B","description":"Flagship Indian-language reasoning model for enterprise multilingual applications","family":"sarvam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-18","last_updated":"2026-02-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.15790269,"output":0.63161075,"cache_read":0.05263423},"canonical_model_id":"sarvam/sarvam-105b"},"sarvam-30b":{"id":"sarvam-30b","name":"Sarvam-30B","description":"Efficient Indian-language reasoning model for chat, coding, and multilingual work","family":"sarvam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-18","last_updated":"2026-02-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":65536},"status":"deprecated","canonical_model_id":"sarvam/sarvam-30b"}}},"neosmith":{"id":"neosmith","env":["NEOSMITH_API_KEY"],"npm":"@ai-sdk/openai","api":"https://router.neosmith.ai/v1","name":"NeoSmith","doc":"https://neosmith.ai/docs","models":{"neosmith.neolite":{"id":"neosmith.neolite","name":"NeoSmith NeoLite","description":"Sealed single-model budget tier. 512K context, text and images, tool use, and no escalation of any kind.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-20","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":512000,"output":64000},"cost":{"input":0.6,"output":2.4,"cache_read":0.08,"cache_write":0}},"neosmith.intelligent-basic":{"id":"neosmith.intelligent-basic","name":"NeoSmith Basic","description":"Cost-capped tier. Intelligent routing with a Claude Sonnet ceiling — Opus is never invoked.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-26","last_updated":"2026-08-03","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":1.17,"output":4.37,"cache_read":0.22,"cache_write":0}},"neosmith.intelligent-pro":{"id":"neosmith.intelligent-pro","name":"NeoSmith Pro","description":"Default production tier. Intelligent NeoSmith routing with a Claude Opus ceiling on escalation.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-26","last_updated":"2026-07-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":1.81,"output":8.39,"cache_read":0.3,"cache_write":0}},"neosmith.intelligent-maestro":{"id":"neosmith.intelligent-maestro","name":"NeoSmith Maestro","description":"Highest-accuracy coding tier. Hard, self-contained problems run NeoSmith's premium multi-model solver; everything else gets the strongest intelligence tier.","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2.4,"output":12,"cache_read":0.35,"cache_write":0}}}},"umans-ai-coding-plan":{"id":"umans-ai-coding-plan","env":["UMANS_AI_CODING_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.code.umans.ai/v1","name":"Umans AI Coding Plan","doc":"https://app.umans.ai/offers/code/docs","models":{"umans-coder":{"id":"umans-coder","name":"Umans Coder","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131071},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"umans-deepseek-v4-flash-0731":{"id":"umans-deepseek-v4-flash-0731","name":"DeepSeek V4 Flash","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393215},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"umans-glm-5.3":{"id":"umans-glm-5.3","name":"GLM 5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131071},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"umans-flash":{"id":"umans-flash","name":"Umans Flash","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"umans-glm-5.3-flash":{"id":"umans-glm-5.3-flash","name":"GLM 5.3 Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131071},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"umans-deepseek-v4.1-flash":{"id":"umans-deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393215},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"umans-kimi-k3":{"id":"umans-kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k3"},"umans-qwen3.6-35b-a3b":{"id":"umans-qwen3.6-35b-a3b","name":"Qwen3.6 35B A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"umans-deepseek-v4.1-flash-1017":{"id":"umans-deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393215},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"}}},"google":{"id":"google","env":["GOOGLE_API_KEY","GOOGLE_GENERATIVE_AI_API_KEY","GEMINI_API_KEY"],"npm":"@ai-sdk/google","name":"Google","doc":"https://ai.google.dev/gemini-api/docs/models","models":{"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.8-flash"},"gemini-omni-flash-preview":{"id":"gemini-omni-flash-preview","type":"video","name":"Gemini Omni Flash Preview","description":"Video generation and editing model for fast, conversational text- and image-to-video workflows","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":1.5,"output":17.5},"canonical_model_id":"google/gemini-omni-flash-preview"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5}},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"gemini-2.5-computer-use-preview-10-2025":{"id":"gemini-2.5-computer-use-preview-10-2025","name":"Gemini 2.5 Computer Use Preview 10-2025","description":"Specialized Gemini 2.5 model for browser-control agents that automate UI tasks","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-10-07","last_updated":"2025-10-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":64000},"cost":{"input":1.25,"output":10,"tiers":[{"input":2.5,"output":15,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15}},"canonical_model_id":"google/gemini-2.5-computer-use-preview-10-2025"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":32768},"cost":{"input":2,"output":10,"cache_read":0.1,"input_audio":2},"canonical_model_id":"google/gemini-4-argon"},"gemini-3.1-pro-preview-customtools":{"id":"gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}}},"gemini-3.1-flash-lite-preview":{"id":"gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Legacy model retained for compatibility with older integrations","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5}},"gemini-3.5-live-translate-preview":{"id":"gemini-3.5-live-translate-preview","type":"realtime","name":"Gemini 3.5 Live Translate Preview","description":"Low-latency audio-to-audio model for real-time speech translation across 70+ languages","family":"gemini-pro","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["audio"],"output":["audio","text"]},"open_weights":false,"limit":{"context":16384,"output":32768},"cost":{"input":3.5,"output":21,"input_audio":3.5,"output_audio":21},"canonical_model_id":"google/gemini-3.5-live-translate-preview"},"veo-3.1-fast-generate-preview":{"id":"veo-3.1-fast-generate-preview","type":"video","name":"Veo 3.1 fast","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-15","last_updated":"2026-01-01","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":8192},"canonical_model_id":"google/veo-3.1-fast-generate-preview"},"gemini-3.1-flash-live-preview":{"id":"gemini-3.1-flash-live-preview","type":"realtime","name":"Gemini 3.1 Flash Live Preview","description":"High-quality, low-latency Live API model for real-time dialogue and voice-first AI applications","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-26","last_updated":"2026-03-26","modalities":{"input":["text","image","video","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":0.75,"output":4.5,"input_audio":3,"output_audio":12},"canonical_model_id":"google/gemini-3.1-flash-live-preview"},"gemini-pro-latest":{"id":"gemini-pro-latest","name":"Gemini Pro Latest","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"gemini-3.1-flash-lite-image":{"id":"gemini-3.1-flash-lite-image","type":"image","name":"Nano Banana 2 Lite","description":"Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":4096},"cost":{"input":0.25,"output":30},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"gemini-3-pro-image-preview":{"id":"gemini-3-pro-image-preview","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":2,"output":120},"canonical_model_id":"google/gemini-3-pro-image-preview"},"gemini-3.1-flash-tts-preview":{"id":"gemini-3.1-flash-tts-preview","type":"speech","name":"Gemini 3.1 Flash TTS Preview","description":"Low-latency speech generation with steerable prompts and expressive audio tags","family":"gemini-flash","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-15","last_updated":"2026-04-15","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":16384},"cost":{"input":1,"output":20},"canonical_model_id":"google/gemini-3.1-flash-tts-preview"},"gemini-2.5-flash-image":{"id":"gemini-2.5-flash-image","type":"image","name":"Nano Banana","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":30,"cache_read":0.075},"canonical_model_id":"google/gemini-2.5-flash-image"},"gemini-embedding-001":{"id":"gemini-embedding-001","type":"embedding","name":"Gemini Embedding 001","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-05","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2048,"output":1},"cost":{"input":0.15,"output":0}},"veo-3.1-generate-preview":{"id":"veo-3.1-generate-preview","type":"video","name":"Veo 3.1","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2025-10-15","last_updated":"2026-01","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":8192},"status":"beta","canonical_model_id":"google/veo-3.1-generate-preview"},"gemini-flash-latest":{"id":"gemini-flash-latest","name":"Gemini Flash Latest","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75}},"deep-research-max-preview-04-2026":{"id":"deep-research-max-preview-04-2026","name":"Deep Research Max Preview (Apr-21-2026)","description":"Maximum-comprehensiveness agentic researcher for multi-step investigation, synthesis, and cited reports","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/deep-research-max-preview-04-2026"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gemini-3.1-flash-image-preview":{"id":"gemini-3.1-flash-image-preview","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":65536},"cost":{"input":0.5,"output":60}},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}}},"gemma-4-26b-a4b-it":{"id":"gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768}},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.7-flash"},"gemini-2.5-flash-preview-tts":{"id":"gemini-2.5-flash-preview-tts","type":"speech","name":"Gemini 2.5 Flash Preview TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gemini-flash","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-05-01","last_updated":"2025-05-01","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":16384},"cost":{"input":0.5,"output":10}},"lyria-3-clip-preview":{"id":"lyria-3-clip-preview","name":"Lyria 3 Clip Preview","description":"Music generation model for short 30-second clips, loops, and previews from text or image prompts","family":"lyria","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-25","last_updated":"2026-03-25","modalities":{"input":["text","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/lyria-3-clip-preview"},"gemini-flash-lite-latest":{"id":"gemini-flash-lite-latest","name":"Gemini Flash-Lite Latest","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03}},"veo-3.1-lite-generate-preview":{"id":"veo-3.1-lite-generate-preview","type":"video","name":"Veo 3.1 lite","description":"Video model for prompt-guided generation, editing, and motion workflows","family":"veo","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image"],"output":["video"]},"open_weights":false,"limit":{"context":480,"output":8192},"canonical_model_id":"google/veo-3.1-lite-generate-preview"},"gemini-2.5-pro-preview-tts":{"id":"gemini-2.5-pro-preview-tts","type":"speech","name":"Gemini 2.5 Pro Preview TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gemini-flash","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-05-01","last_updated":"2025-05-01","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":16384},"cost":{"input":1,"output":20}},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5}},"gemini-3.1-flash-image":{"id":"gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","video","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":60},"canonical_model_id":"google/gemini-3.1-flash-image"},"lyria-3-pro-preview":{"id":"lyria-3-pro-preview","name":"Lyria 3 Pro Preview","description":"Music generation model for full-length songs from text or images with vocals and structure","family":"lyria","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-25","last_updated":"2026-03-25","modalities":{"input":["text","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/lyria-3-pro-preview"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1}},"gemma-4-31b-it":{"id":"gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768}},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3}},"deep-research-preview-04-2026":{"id":"deep-research-preview-04-2026","name":"Deep Research Preview (Apr-21-2026)","description":"Agentic model for autonomous multi-step research, synthesis, and cited reports","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/deep-research-preview-04-2026"},"gemini-3-pro-image":{"id":"gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":120},"canonical_model_id":"google/gemini-3-pro-image"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"gemini-embedding-2":{"id":"gemini-embedding-2","type":"embedding","name":"Gemini Embedding 2","description":"Multimodal embedding model mapping text, images, video, audio, and PDFs into a unified embedding space","family":"gemini","attachment":true,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-11","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":1},"cost":{"input":0.2,"output":0,"input_audio":6.5},"canonical_model_id":"google/gemini-embedding-2"}}},"iteracompute":{"id":"iteracompute","env":["ITERACOMPUTE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.iteracompute.com/v1","name":"IteraCompute","doc":"https://iteracompute.com/docs.html","models":{"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.08,"output":0.28,"cache_read":0.023},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.82,"output":2.88,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5.3"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.69,"output":3.29,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":999999},"cost":{"input":2.35,"output":11.76,"cache_read":0.23},"canonical_model_id":"moonshotai/kimi-k3"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.79,"output":2.38,"cache_read":0.026},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.17,"output":0.7,"cache_read":0.005},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731 (legacy alias)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.17,"output":0.7,"cache_read":0.005},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":524288},"cost":{"input":0.3,"output":1.19,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M3"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":8192},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.8-2.4t-a95b":{"id":"qwen/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.95,"output":5.7,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"}}},"alibaba-token-plan-cn":{"id":"alibaba-token-plan-cn","env":["ALIBABA_TOKEN_PLAN_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://token-plan.cn-beijing.maas.aliyuncs.com/compatible-mode/v1","name":"Alibaba Token Plan (China)","doc":"https://www.alibabacloud.com/help/zh/model-studio/token-plan-overview","models":{"qwen3.7-max":{"id":"qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen3.7-plus":{"id":"qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen-image-2.0-pro":{"id":"qwen-image-2.0-pro","type":"image","name":"Qwen Image 2.0 Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"deepseek-v4-flash":{"id":"deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"kimi-k2.6":{"id":"kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"deepseek-v3.2":{"id":"deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-03","last_updated":"2025-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"status":"deprecated","cost":{"input":0,"output":0}},"deepseek-v4.1-flash-1017":{"id":"deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"qwen-image-2.0":{"id":"qwen-image-2.0","type":"image","name":"Qwen Image 2.0","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"deepseek-v4-pro-0813":{"id":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek-v4-pro":{"id":"deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-pro"},"glm-5.1":{"id":"glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":128000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"qwen3.8-max-preview":{"id":"qwen3.8-max-preview","name":"Qwen3.8 Max Preview","description":"Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-07-19","last_updated":"2026-07-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.8-max-preview"},"qwen3.8-max":{"id":"qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.8-max"},"kimi-k2.7-code":{"id":"kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"qwen3.6-flash":{"id":"qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":131072}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-flash"},"happyhorse-1.1-r2v":{"id":"happyhorse-1.1-r2v","type":"video","name":"HappyHorse 1.1 Reference-to-Video","description":"Video model for reference-guided video generation","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-17","last_updated":"2026-07-17","modalities":{"input":["image","text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}},"happyhorse-1.1-i2v":{"id":"happyhorse-1.1-i2v","type":"video","name":"HappyHorse 1.1 Image-to-Video","description":"Video model for image-to-video generation","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-17","last_updated":"2026-07-17","modalities":{"input":["image","text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}},"wan2.7-image-pro":{"id":"wan2.7-image-pro","type":"image","name":"Wan2.7 Image Pro","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"qwen3.8-flash":{"id":"qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.8-flash"},"kimi-k2.5":{"id":"kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.5"},"glm-5.2":{"id":"glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"deepseek-v4.1-flash":{"id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"qwen3.6-plus":{"id":"qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","max":131072}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"alibaba/qwen3.6-plus"},"deepseek-v4-flash-0731":{"id":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"glm-5":{"id":"glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":16384},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5"},"glm-5.3":{"id":"glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.3"},"wan2.7-image":{"id":"wan2.7-image","type":"image","name":"Wan2.7 Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text"],"output":["image"]},"open_weights":false,"limit":{"context":8192,"output":0},"cost":{"input":0,"output":0}},"happyhorse-1.1-t2v":{"id":"happyhorse-1.1-t2v","type":"video","name":"HappyHorse 1.1 Text-to-Video","description":"Video model for prompt-driven text-to-video generation","attachment":false,"reasoning":false,"tool_call":false,"temperature":true,"release_date":"2026-07-17","last_updated":"2026-07-17","modalities":{"input":["text"],"output":["video"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0}},"MiniMax-M2.5":{"id":"MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"input":196601,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"minimax/MiniMax-M2.5"}}},"kosmik":{"id":"kosmik","env":["KOSMIK_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.koscompute.com/v1","name":"Kosmik Compute","doc":"https://api.koscompute.com/docs/","models":{"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.35,"output":2.2,"cache_read":0.09},"canonical_model_id":"alibaba/qwen3.8-27b"}}},"google-vertex":{"id":"google-vertex","env":["GOOGLE_VERTEX_PROJECT","GOOGLE_VERTEX_LOCATION","GOOGLE_APPLICATION_CREDENTIALS"],"npm":"@ai-sdk/google-vertex","name":"Vertex","doc":"https://cloud.google.com/vertex-ai/generative-ai/docs/models","models":{"gemini-3.8-flash":{"id":"gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.8-flash"},"gemini-3.1-flash-lite":{"id":"gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1}},"claude-opus-4@20250514":{"id":"claude-opus-4@20250514","name":"Claude Opus 4","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-20250514"},"claude-opus-5-5@default":{"id":"claude-opus-5-5@default","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"claude-fable-5@default":{"id":"claude-fable-5@default","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"gemini-4-argon":{"id":"gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1,"input_audio":2},"canonical_model_id":"google/gemini-4-argon"},"claude-fable-5-1@default":{"id":"claude-fable-5-1@default","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"gemini-3.1-pro-preview-customtools":{"id":"gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"beta","cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"claude-sonnet-5@default":{"id":"claude-sonnet-5@default","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"claude-opus-5@default":{"id":"claude-opus-5@default","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"claude-haiku-4-5@20251001":{"id":"claude-haiku-4-5@20251001","name":"Claude Haiku 4.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"gemini-3.1-flash-lite-image":{"id":"gemini-3.1-flash-lite-image","type":"image","name":"Nano Banana 2 Lite","description":"Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","video","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":4096},"cost":{"input":0.25,"output":30,"cache_read":0.025},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"gemini-2.5-flash-image":{"id":"gemini-2.5-flash-image","type":"image","name":"Nano Banana","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":32768},"cost":{"input":0.3,"output":30},"canonical_model_id":"google/gemini-2.5-flash-image"},"gemini-embedding-001":{"id":"gemini-embedding-001","type":"embedding","name":"Gemini Embedding 001","description":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","family":"gemini","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-05","release_date":"2025-05-20","last_updated":"2025-05-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2048,"output":1},"cost":{"input":0.15,"output":0},"canonical_model_id":"google/gemini-embedding-001"},"codestral-2":{"id":"codestral-2","name":"Codestral 2","description":"Mistral code model for low-latency completion, fill-in-the-middle, and code generation","family":"codestral","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"provider":{"npm":"@ai-sdk/mistral","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/publishers/mistralai/models"},"cost":{"input":0.3,"output":0.9},"canonical_model_id":"mistral/codestral-2508"},"mistral-medium-3":{"id":"mistral-medium-3","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"provider":{"npm":"@ai-sdk/mistral","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/publishers/mistralai/models"},"cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/mistral-medium-2505"},"claude-haiku-5-5@default":{"id":"claude-haiku-5-5@default","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"claude-opus-4-6@default":{"id":"claude-opus-4-6@default","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"gemini-3.5-flash-lite":{"id":"gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"gemini-2.5-pro-tts":{"id":"gemini-2.5-pro-tts","type":"speech","name":"Gemini 2.5 Pro TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gemini-pro","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-09-30","last_updated":"2025-12-10","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":32768,"input":8192,"output":16384},"cost":{"input":1,"output":20},"canonical_model_id":"google/gemini-2.5-pro-tts"},"gemini-nano-banana-2.1":{"id":"gemini-nano-banana-2.1","type":"image","name":"Nano Banana 2.1","description":"Image model for prompt-driven generation, multi-turn editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","medium","high"]}],"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image","video","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":1.5,"output":30,"cache_read":0.15},"canonical_model_id":"google/gemini-nano-banana-2.1"},"gemini-3.1-pro-preview":{"id":"gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"beta","cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"gemini-3.6-flash":{"id":"gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.6-flash"},"gemini-3.7-flash":{"id":"gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.7-flash"},"claude-opus-4-8@default":{"id":"claude-opus-4-8@default","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"claude-opus-4-5@20251101":{"id":"claude-opus-4-5@20251101","name":"Claude Opus 4.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"claude-sonnet-4@20250514":{"id":"claude-sonnet-4@20250514","name":"Claude Sonnet 4","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-20250514"},"gemini-3.5-flash":{"id":"gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"input_audio":1.5},"canonical_model_id":"google/gemini-3.5-flash"},"claude-opus-4-1@20250805":{"id":"claude-opus-4-1@20250805","name":"Claude Opus 4.1","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1-20250805"},"gemini-2.5-flash-tts":{"id":"gemini-2.5-flash-tts","type":"speech","name":"Gemini 2.5 Flash TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gemini-flash","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"knowledge":"2025-01","release_date":"2025-09-30","last_updated":"2025-12-10","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":32768,"input":8192,"output":16384},"cost":{"input":0.5,"output":10},"canonical_model_id":"google/gemini-2.5-flash-tts"},"mistral-small-2503":{"id":"mistral-small-2503","name":"Mistral Small 3.1 24B","description":"Efficient multimodal model for instruction following, coding, reasoning, and function calling","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"provider":{"npm":"@ai-sdk/mistral","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/publishers/mistralai/models"},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/mistral-small-3-1-24b-instruct-2503"},"gemini-3.1-flash-image":{"id":"gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","high"]}],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","video","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":60,"cache_read":0.05},"canonical_model_id":"google/gemini-3.1-flash-image"},"gemini-3-flash-preview":{"id":"gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"beta","cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"claude-sonnet-4-6@default":{"id":"claude-sonnet-4-6@default","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"claude-sonnet-4-5@20250929":{"id":"claude-sonnet-4-5@20250929","name":"Claude Sonnet 4.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":512,"max":24576}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"gemini-3-pro-image":{"id":"gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":120,"cache_read":0.2},"canonical_model_id":"google/gemini-3-pro-image"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":1.25,"output":10,"cache_read":0.125,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"claude-sonnet-5-5@default":{"id":"claude-sonnet-5-5@default","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"claude-opus-4-7@default":{"id":"claude-opus-4-7@default","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/google-vertex/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"meta/llama-4-maverick-17b-128e-instruct-maas":{"id":"meta/llama-4-maverick-17b-128e-instruct-maas","name":"Llama 4 Maverick 17B 128E Instruct","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":8192},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.35,"output":1.15}},"meta/llama-4-scout-17b-16e-instruct-maas":{"id":"meta/llama-4-scout-17b-16e-instruct-maas","name":"Llama 4 Scout 17B Instruct","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1310720,"output":8192},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.25,"output":0.7},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"meta/llama-3.3-70b-instruct-maas":{"id":"meta/llama-3.3-70b-instruct-maas","name":"Llama 3.3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.72,"output":0.72}},"moonshotai/kimi-k2-thinking-maas":{"id":"moonshotai/kimi-k2-thinking-maas","name":"Kimi K2 Thinking","description":"Kimi reasoning model for long-horizon research, planning, and tool use","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.6,"output":2.5,"cache_read":0.06}},"openai/gpt-oss-20b-maas":{"id":"openai/gpt-oss-20b-maas","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.07,"output":0.25,"cache_read":0.007}},"openai/gpt-oss-120b-maas":{"id":"openai/gpt-oss-120b-maas","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.09,"output":0.36},"canonical_model_id":"openai/gpt-oss-120b"},"zai-org/glm-4.7-maas":{"id":"zai-org/glm-4.7-maas","name":"GLM-4.7","description":"GLM model for coding, tool use, and complex reasoning","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-06","last_updated":"2026-01-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.6,"output":2.2,"cache_read":0.06}},"zai-org/glm-5-maas":{"id":"zai-org/glm-5-maas","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":1,"output":3.2,"cache_read":0.1},"canonical_model_id":"zhipuai/glm-5"},"zai-org/glm-5.2-maas":{"id":"zai-org/glm-5.2-maas","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":64000},"status":"beta","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.2"},"xai/grok-4.6":{"id":"xai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":500000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"xai/grok-4.7":{"id":"xai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":500000},"status":"beta","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"xai/grok-4.1-fast-reasoning":{"id":"xai/grok-4.1-fast-reasoning","name":"Grok 4.1 Fast (Reasoning)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":30000},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast-reasoning"},"xai/grok-4.20-reasoning":{"id":"xai/grok-4.20-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"xai/grok-4.20-non-reasoning":{"id":"xai/grok-4.20-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"xai/grok-4.1-fast-non-reasoning":{"id":"xai/grok-4.1-fast-non-reasoning","name":"Grok 4.1 Fast","description":"Fast Grok model for responsive chat, tool-assisted work, and low-latency responses","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":30000},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.2,"output":0.5,"cache_read":0.05},"canonical_model_id":"xai/grok-4.1-fast"},"xai/grok-4.3":{"id":"xai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":30000},"status":"beta","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"deepseek-ai/deepseek-v3.2-maas":{"id":"deepseek-ai/deepseek-v3.2-maas","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2025-12-17","last_updated":"2026-04-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":65536},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.56,"output":1.68,"cache_read":0.056}},"deepseek-ai/deepseek-v3.1-maas":{"id":"deepseek-ai/deepseek-v3.1-maas","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.6,"output":1.7,"cache_read":0.06}},"qwen/qwen3-235b-a22b-instruct-2507-maas":{"id":"qwen/qwen3-235b-a22b-instruct-2507-maas","name":"Qwen3 235B A22B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-13","last_updated":"2025-08-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"status":"deprecated","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.22,"output":0.88}},"google/gemma-4-26b-a4b-it-maas":{"id":"google/gemma-4-26b-a4b-it-maas","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":128000},"status":"alpha","provider":{"npm":"@ai-sdk/openai-compatible","api":"https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi"},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"google/gemma-4-26b-a4b-it"}}},"llmgateway-providers":{"id":"llmgateway-providers","env":["LLMGATEWAY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.llmgateway.io/v1","name":"LLM Gateway","doc":"https://llmgateway.io/docs","models":{"meta/muse-wonder-1.0":{"id":"meta/muse-wonder-1.0","name":"Muse Wonder 1.0 (Meta)","description":"Muse Wonder is a frontier multimodal reasoning model from Meta for personal and professional agents, long-horizon coding, and computer use. It combines Active Memory, which carries relevant context across conversations, with Active Focus, which keeps it working through blockers toward a goal.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3 (Meta)","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1 (Meta)","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2 (Meta)","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"consensusprotocol/gpt-oss-20b":{"id":"consensusprotocol/gpt-oss-20b","name":"GPT OSS 20B (Consensus Protocol)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32768},"cost":{"input":0.04,"output":0.19,"cache_read":0.01},"canonical_model_id":"openai/gpt-oss-20b"},"consensusprotocol/deepseek-v4-flash":{"id":"consensusprotocol/deepseek-v4-flash","name":"DeepSeek V4 Flash (Consensus Protocol)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.1,"output":0.2,"cache_read":0.005},"canonical_model_id":"deepseek/deepseek-v4-flash"},"consensusprotocol/mimo-v2.6-flash":{"id":"consensusprotocol/mimo-v2.6-flash","name":"MiMo V2.6 Flash (Consensus Protocol)","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.12,"output":0.25,"cache_read":0.0025},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"consensusprotocol/Qwen3.8-27B":{"id":"consensusprotocol/Qwen3.8-27B","name":"Qwen3.8 27B (Consensus Protocol)","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":32768},"cost":{"input":0.08,"output":0.35,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.8-27b"},"consensusprotocol/glm-5.3-flash":{"id":"consensusprotocol/glm-5.3-flash","name":"GLM-5.3 Flash (Consensus Protocol)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.25,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-5.3-flash"},"consensusprotocol/minimax-m3":{"id":"consensusprotocol/minimax-m3","name":"MiniMax M3 (Consensus Protocol)","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.28,"output":1.1,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M3"},"consensusprotocol/deepseek-v4.1-flash":{"id":"consensusprotocol/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Consensus Protocol)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.15,"output":0.55,"cache_read":0.005},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"consensusprotocol/gemma-4-31b-it":{"id":"consensusprotocol/gemma-4-31b-it","name":"Gemma 4 31B IT (Consensus Protocol)","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.1,"output":0.25,"cache_read":0.01},"canonical_model_id":"google/gemma-4-31b-it"},"sakana/fugu-ultra-v2.0":{"id":"sakana/fugu-ultra-v2.0","name":"Fugu Ultra v2.0 (Sakana AI)","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-08-28","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"sakana/fugu-ultra-v2"},"sakana/fugu-max":{"id":"sakana/fugu-max","name":"Fugu Max (Sakana AI)","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"sakana/fugu-max"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra (Sakana AI)","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":5,"output":30,"cache_read":0.5},"canonical_model_id":"sakana/fugu-ultra"},"azure-anthropic/claude-sonnet-5-5":{"id":"azure-anthropic/claude-sonnet-5-5","name":"Claude Sonnet 5.5 (Azure Anthropic)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"azure-anthropic/claude-opus-4-8":{"id":"azure-anthropic/claude-opus-4-8","name":"Claude Opus 4.8 (Azure Anthropic)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"azure-anthropic/claude-opus-4-7":{"id":"azure-anthropic/claude-opus-4-7","name":"Claude Opus 4.7 (Azure Anthropic)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"azure-anthropic/claude-fable-5":{"id":"azure-anthropic/claude-fable-5","name":"Claude Fable 5 (Azure Anthropic)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"azure-anthropic/claude-opus-5-5":{"id":"azure-anthropic/claude-opus-5-5","name":"Claude Opus 5.5 (Azure Anthropic)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"azure-anthropic/claude-fable-5-1":{"id":"azure-anthropic/claude-fable-5-1","name":"Claude Fable 5.1 (Azure Anthropic)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"azure-anthropic/claude-opus-5":{"id":"azure-anthropic/claude-opus-5","name":"Claude Opus 5 (Azure Anthropic)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"azure-anthropic/claude-opus-4-6":{"id":"azure-anthropic/claude-opus-4-6","name":"Claude Opus 4.6 (Azure Anthropic)","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"azure-anthropic/claude-sonnet-5":{"id":"azure-anthropic/claude-sonnet-5","name":"Claude Sonnet 5 (Azure Anthropic)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"azure/gpt-6-sol":{"id":"azure/gpt-6-sol","name":"GPT-6 Sol (Azure)","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-6-sol"},"azure/gpt-5.1":{"id":"azure/gpt-5.1","name":"GPT-5.1 (Azure)","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"azure/gpt-4.1-mini":{"id":"azure/gpt-4.1-mini","name":"GPT-4.1 Mini (Azure)","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"azure/gpt-6-astra":{"id":"azure/gpt-6-astra","name":"GPT-6 Astra (Azure)","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"openai/gpt-6-astra"},"azure/gpt-5.6-terra":{"id":"azure/gpt-5.6-terra","name":"GPT-5.6 Terra (Azure)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"azure/gpt-5-mini":{"id":"azure/gpt-5-mini","name":"GPT-5 Mini (Azure)","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"azure/gpt-5.3-codex":{"id":"azure/gpt-5.3-codex","name":"GPT-5.3 Codex (Azure)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"azure/gpt-5.2-codex":{"id":"azure/gpt-5.2-codex","name":"GPT-5.2 Codex (Azure)","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"azure/gpt-5.4-pro":{"id":"azure/gpt-5.4-pro","name":"GPT-5.4 Pro (Azure)","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.4-pro"},"azure/gpt-5-nano":{"id":"azure/gpt-5-nano","name":"GPT-5 Nano (Azure)","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"azure/gpt-5.2":{"id":"azure/gpt-5.2","name":"GPT-5.2 (Azure)","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"azure/gpt-5.5":{"id":"azure/gpt-5.5","name":"GPT-5.5 (Azure)","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"azure/gpt-5.4":{"id":"azure/gpt-5.4","name":"GPT-5.4 (Azure)","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"azure/gpt-6-luna":{"id":"azure/gpt-6-luna","name":"GPT-6 Luna (Azure)","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"openai/gpt-6-luna"},"azure/gpt-oss-120b":{"id":"azure/gpt-oss-120b","name":"GPT OSS 120B (Azure)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"azure/gpt-5.6-sol":{"id":"azure/gpt-5.6-sol","name":"GPT-5.6 Sol (Azure)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5},"canonical_model_id":"openai/gpt-5.6-sol"},"azure/gpt-6.1-sol":{"id":"azure/gpt-6.1-sol","name":"GPT-6.1 Sol (Azure)","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"openai/gpt-6.1-sol"},"azure/gpt-5.6-luna":{"id":"azure/gpt-5.6-luna","name":"GPT-5.6 Luna (Azure)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"azure/gpt-4o":{"id":"azure/gpt-4o","name":"GPT-4o (Azure)","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"azure/gpt-4.1":{"id":"azure/gpt-4.1","name":"GPT-4.1 (Azure)","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"azure/gpt-5.1-codex-mini":{"id":"azure/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini (Azure)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"azure/gpt-5":{"id":"azure/gpt-5","name":"GPT-5 (Azure)","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"azure/gpt-5.1-codex":{"id":"azure/gpt-5.1-codex","name":"GPT-5.1 Codex (Azure)","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":1.25,"output":10},"canonical_model_id":"openai/gpt-5.1-codex"},"azure/gpt-5.4-mini":{"id":"azure/gpt-5.4-mini","name":"GPT-5.4 Mini (Azure)","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"azure/gpt-5.4-nano":{"id":"azure/gpt-5.4-nano","name":"GPT-5.4 Nano (Azure)","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"scx-ai/minimax-m2.7":{"id":"scx-ai/minimax-m2.7","name":"MiniMax M2.7 (SCX.ai (Turbo))","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":196608},"cost":{"input":0.48,"output":1.79,"cache_read":0.05},"canonical_model_id":"minimax/MiniMax-M2.7"},"scx-ai/gpt-oss-120b":{"id":"scx-ai/gpt-oss-120b","name":"GPT OSS 120B (SCX.ai (Turbo))","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.17,"output":0.55},"canonical_model_id":"openai/gpt-oss-120b"},"scx-ai/qwen3-32b":{"id":"scx-ai/qwen3-32b","name":"Qwen3 32B (SCX.ai (Turbo))","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":8192},"cost":{"input":0.36,"output":0.87},"canonical_model_id":"alibaba/qwen3-32b"},"google-ai-studio/gemini-3.8-flash":{"id":"google-ai-studio/gemini-3.8-flash","name":"Gemini 3.8 Flash (Google AI Studio)","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.8-flash"},"google-ai-studio/gemini-3.1-flash-lite":{"id":"google-ai-studio/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite (Google AI Studio)","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.08333,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google-ai-studio/gemini-4-argon":{"id":"google-ai-studio/gemini-4-argon","name":"Gemini 4 Argon (Google AI Studio)","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-4-argon"},"google-ai-studio/gemini-pro-latest":{"id":"google-ai-studio/gemini-pro-latest","name":"Gemini Pro Latest (Google AI Studio)","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-27","last_updated":"2026-02-27","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2}},"google-ai-studio/gemini-3.5-flash-lite":{"id":"google-ai-studio/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite (Google AI Studio)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google-ai-studio/gemini-3.1-pro-preview":{"id":"google-ai-studio/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro (Preview) (Google AI Studio)","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google-ai-studio/gemini-3.6-flash":{"id":"google-ai-studio/gemini-3.6-flash","name":"Gemini 3.6 Flash (Google AI Studio)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.6-flash"},"google-ai-studio/gemini-3-flash-preview":{"id":"google-ai-studio/gemini-3-flash-preview","name":"Gemini 3 Flash (Preview) (Google AI Studio)","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"luminal/deepseek-v4.1-flash":{"id":"luminal/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (luminal)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.135,"output":0.54,"cache_read":0.0027},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"mistral/devstral-2512":{"id":"mistral/devstral-2512","name":"Devstral 2 (Mistral AI)","description":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","family":"devstral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.4,"output":2},"canonical_model_id":"mistral/devstral-2512"},"mistral/ministral-3b-2512":{"id":"mistral/ministral-3b-2512","name":"Ministral 3B (Mistral AI)","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.1,"output":0.1},"canonical_model_id":"mistral/ministral-3b-2512"},"mistral/mistral-large-4":{"id":"mistral/mistral-large-4","name":"Mistral Large 4 (Mistral AI)","description":"Mistral's 1T-parameter multimodal MoE for reasoning, coding, and agentic work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":262144},"cost":{"input":1.36,"output":4.18,"cache_read":0.14},"canonical_model_id":"mistral/mistral-large-4"},"mistral/mistral-large-latest":{"id":"mistral/mistral-large-latest","name":"Mistral Large Latest (Mistral AI)","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-11","release_date":"2024-11-01","last_updated":"2025-12-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":262144},"cost":{"input":4,"output":12},"canonical_model_id":"mistral/mistral-large-latest"},"mistral/ministral-8b-2512":{"id":"mistral/ministral-8b-2512","name":"Ministral 8B (Mistral AI)","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.15,"output":0.15},"canonical_model_id":"mistral/ministral-8b-2512"},"mistral/mistral-small-2603":{"id":"mistral/mistral-small-2603","name":"Mistral Small 4 (Mistral AI)","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"mistral/mistral-small-2603"},"mistral/codestral-2508":{"id":"mistral/codestral-2508","name":"Codestral (Mistral AI)","description":"Mistral code model for low-latency completion, fill-in-the-middle, and code generation","family":"codestral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.3,"output":0.9},"canonical_model_id":"mistral/codestral-2508"},"mistral/ministral-14b-2512":{"id":"mistral/ministral-14b-2512","name":"Ministral 14B (Mistral AI)","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.2,"output":0.2},"canonical_model_id":"mistral/ministral-14b-2512"},"mistral/mistral-small-2506":{"id":"mistral/mistral-small-2506","name":"Mistral Small 3.2 (Mistral AI)","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-03","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"mistral/mistral-small-2506"},"mistral/glm-5.3":{"id":"mistral/glm-5.3","name":"GLM-5.3 (Mistral AI)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.14},"canonical_model_id":"zhipuai/glm-5.3"},"mistral/mistral-medium-3-5":{"id":"mistral/mistral-medium-3-5","name":"Mistral Medium 3.5 (Mistral AI)","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"mistral/mistral-medium-3-5"},"mistral/mistral-large-2512":{"id":"mistral/mistral-large-2512","name":"Mistral Large 3 (Mistral AI)","description":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"mistral/mistral-large-2512"},"groq/gpt-oss-20b":{"id":"groq/gpt-oss-20b","name":"GPT OSS 20B (Groq)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32766},"cost":{"input":0.1,"output":0.5},"canonical_model_id":"openai/gpt-oss-20b"},"groq/gpt-oss-120b":{"id":"groq/gpt-oss-120b","name":"GPT OSS 120B (Groq)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32766},"cost":{"input":0.15,"output":0.75},"canonical_model_id":"openai/gpt-oss-120b"},"runware/deepseek-v4-flash":{"id":"runware/deepseek-v4-flash","name":"DeepSeek V4 Flash (Runware)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.076,"output":0.153,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash"},"runware/kimi-k2.6":{"id":"runware/kimi-k2.6","name":"Kimi K2.6 (Runware)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.6,"output":3.05,"cache_read":0.13},"canonical_model_id":"moonshotai/kimi-k2.6"},"runware/deepseek-v4-pro":{"id":"runware/deepseek-v4-pro","name":"DeepSeek V4 Pro (Runware)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.961,"output":1.922,"cache_read":0.079},"canonical_model_id":"deepseek/deepseek-v4-pro"},"runware/gpt-oss-120b":{"id":"runware/gpt-oss-120b","name":"GPT OSS 120B (Runware)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.032,"output":0.14,"cache_read":0.032},"canonical_model_id":"openai/gpt-oss-120b"},"runware/kimi-k3":{"id":"runware/kimi-k3","name":"Kimi K3 (Runware)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"runware/glm-5.3-flash":{"id":"runware/glm-5.3-flash","name":"GLM-5.3 Flash (Runware)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"runware/glm-5.2":{"id":"runware/glm-5.2","name":"GLM-5.2 (Runware)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":128000},"cost":{"input":0.8,"output":2.55,"cache_read":0.16},"canonical_model_id":"zhipuai/glm-5.2"},"runware/deepseek-v4.1-flash":{"id":"runware/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Runware)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.01},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"runware/gemma-4-31b-it":{"id":"runware/gemma-4-31b-it","name":"Gemma 4 31B IT (Runware)","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.102,"output":0.297,"cache_read":0.012},"canonical_model_id":"google/gemma-4-31b-it"},"runware/glm-5.3":{"id":"runware/glm-5.3","name":"GLM-5.3 (Runware)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5.3"},"aws-mantle/gpt-6-astra":{"id":"aws-mantle/gpt-6-astra","name":"GPT-6 Astra (AWS Mantle)","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"openai/gpt-6-astra"},"aws-mantle/gpt-5.6-terra":{"id":"aws-mantle/gpt-5.6-terra","name":"GPT-5.6 Terra (AWS Mantle)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":921600,"output":128000},"cost":{"input":2.2,"output":13.2,"cache_read":0.22,"cache_write":2.75},"canonical_model_id":"openai/gpt-5.6-terra"},"aws-mantle/gpt-5.6-sol":{"id":"aws-mantle/gpt-5.6-sol","name":"GPT-5.6 Sol (AWS Mantle)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":921600,"output":128000},"cost":{"input":4.4,"output":22,"cache_read":0.44,"cache_write":5.5},"canonical_model_id":"openai/gpt-5.6-sol"},"aws-mantle/gpt-5.6-luna":{"id":"aws-mantle/gpt-5.6-luna","name":"GPT-5.6 Luna (AWS Mantle)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":921600,"output":128000},"cost":{"input":0.22,"output":1.32,"cache_read":0.022,"cache_write":0.275},"canonical_model_id":"openai/gpt-5.6-luna"},"cerebras/qwen3-235b-a22b-instruct-2507":{"id":"cerebras/qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B A22B Instruct 2507 (Cerebras)","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":8192},"cost":{"input":0.6,"output":1.2},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"cerebras/gpt-oss-120b":{"id":"cerebras/gpt-oss-120b","name":"GPT OSS 120B (Cerebras)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.35,"output":0.75},"canonical_model_id":"openai/gpt-oss-120b"},"cerebras/glm-4.7":{"id":"cerebras/glm-4.7","name":"GLM-4.7 (Cerebras)","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":2.25,"output":2.75},"canonical_model_id":"zhipuai/glm-4.7"},"cerebras/llama-3.3-70b-instruct":{"id":"cerebras/llama-3.3-70b-instruct","name":"Llama 3.3 70B Instruct (Cerebras)","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"cost":{"input":0.85,"output":1.2},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"cerebras/gemma-4-31b-it":{"id":"cerebras/gemma-4-31b-it","name":"Gemma 4 31B IT (Cerebras)","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.99,"output":1.49},"canonical_model_id":"google/gemma-4-31b-it"},"azure-ai-foundry/grok-4-1-fast-reasoning":{"id":"azure-ai-foundry/grok-4-1-fast-reasoning","name":"Grok 4.1 Fast Reasoning (Azure AI Foundry)","description":"xAI's fast agentic tool-calling model with a 2M context window and built-in reasoning","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5},"canonical_model_id":"xai/grok-4.1-fast-reasoning"},"azure-ai-foundry/grok-4-1-fast-non-reasoning":{"id":"azure-ai-foundry/grok-4-1-fast-non-reasoning","name":"Grok 4.1 Fast Non-Reasoning (Azure AI Foundry)","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":0.2,"output":0.5}},"azure-ai-foundry/grok-4-3":{"id":"azure-ai-foundry/grok-4-3","name":"Grok 4.3 (Azure AI Foundry)","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":20000,"output":8192},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"gonka24/minimax-m2.7":{"id":"gonka24/minimax-m2.7","name":"MiniMax M2.7 (Gonka24)","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.08,"output":0.32,"cache_read":0.017},"canonical_model_id":"minimax/MiniMax-M2.7"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol (OpenAI)","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1 (OpenAI)","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 Mini (OpenAI)","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 Pro (OpenAI)","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra (OpenAI)","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"openai/gpt-6-astra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3 Mini (OpenAI)","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro (OpenAI)","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-4":{"id":"openai/gpt-4","name":"GPT-4 (OpenAI)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"status":"deprecated","cost":{"input":30,"output":60},"canonical_model_id":"openai/gpt-4"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra (OpenAI)","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini (OpenAI)","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex (OpenAI)","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro (OpenAI)","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano (OpenAI)","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2 (OpenAI)","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5 (OpenAI)","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-transcribe":{"id":"openai/gpt-transcribe","type":"transcription","name":"GPT Transcribe (OpenAI)","description":"High-accuracy speech-to-text model for audio files and committed Realtime turns","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-07-28","last_updated":"2026-07-28","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-transcribe"},"openai/gpt-4o-transcribe":{"id":"openai/gpt-4o-transcribe","name":"GPT-4o Transcribe (OpenAI)","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-20","last_updated":"2025-03-20","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":2000},"cost":{"input":2.5,"output":10,"input_audio":6}},"openai/o3":{"id":"openai/o3","name":"o3 (OpenAI)","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4 (OpenAI)","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-4o-mini-transcribe":{"id":"openai/gpt-4o-mini-transcribe","name":"GPT-4o Mini Transcribe (OpenAI)","description":"Speech transcription model for accurate audio-to-text and captioning workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-20","last_updated":"2025-03-20","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":16000,"output":2000},"cost":{"input":1.25,"output":5,"input_audio":3}},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna (OpenAI)","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol (OpenAI)","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol (OpenAI)","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna (OpenAI)","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o (OpenAI)","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 Nano (OpenAI)","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"status":"deprecated","cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o Mini (OpenAI)","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo (OpenAI)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4 Mini (OpenAI)","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5 Turbo (OpenAI)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"status":"deprecated","cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1 (OpenAI)","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/o1":{"id":"openai/o1","name":"o1 (OpenAI)","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"status":"deprecated","cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5 (OpenAI)","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2 Pro (OpenAI)","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/gpt-live-transcribe":{"id":"openai/gpt-live-transcribe","type":"transcription","name":"GPT Live Transcribe (OpenAI)","description":"Low-latency streaming speech-to-text model with keyword and language hints","family":"gpt","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-07-28","last_updated":"2026-07-28","modalities":{"input":["text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":0,"output":0},"cost":{"input":0,"output":0},"canonical_model_id":"openai/gpt-live-transcribe"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 Mini (OpenAI)","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 Nano (OpenAI)","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (DeepSeek)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":393216},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro (DeepSeek)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":393216},"cost":{"input":0.435,"output":0.87,"cache_read":0.003625},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (DeepSeek)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":393216},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"anthropic/claude-sonnet-5-5":{"id":"anthropic/claude-sonnet-5-5","name":"Claude Sonnet 5.5 (Anthropic)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4-8":{"id":"anthropic/claude-opus-4-8","name":"Claude Opus 4.8 (Anthropic)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-opus-4-7":{"id":"anthropic/claude-opus-4-7","name":"Claude Opus 4.7 (Anthropic)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-haiku-4-5":{"id":"anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5 (Anthropic)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-haiku-5-5":{"id":"anthropic/claude-haiku-5-5","name":"Claude Haiku 5.5 (Anthropic)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-sonnet-4-5":{"id":"anthropic/claude-sonnet-4-5","name":"Claude Sonnet 4.5 (Anthropic)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-sonnet-4-6":{"id":"anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6 (Anthropic)","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5 (Anthropic)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-5-5":{"id":"anthropic/claude-opus-5-5","name":"Claude Opus 5.5 (Anthropic)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-fable-5-1":{"id":"anthropic/claude-fable-5-1","name":"Claude Fable 5.1 (Anthropic)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"anthropic/claude-opus-4-5-20251101":{"id":"anthropic/claude-opus-4-5-20251101","name":"Claude Opus 4.5 (Anthropic)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5 (Anthropic)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-4-6":{"id":"anthropic/claude-opus-4-6","name":"Claude Opus 4.6 (Anthropic)","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-sonnet-4-5-20250929":{"id":"anthropic/claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5 (2025-09-29) (Anthropic)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"anthropic/claude-haiku-4-5-20251001":{"id":"anthropic/claude-haiku-4-5-20251001","name":"Claude Haiku 4.5 (2025-10-01) (Anthropic)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5 (Anthropic)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"ranoai/deepseek-v4-flash":{"id":"ranoai/deepseek-v4-flash","name":"DeepSeek V4 Flash (RanoAI)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"vertex-openai/grok-4-20-reasoning":{"id":"vertex-openai/grok-4-20-reasoning","name":"Grok 4.20 Reasoning (Vertex AI (OpenAI-compatible))","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"vertex-openai/grok-4-6":{"id":"vertex-openai/grok-4-6","name":"Grok 4.6 (Vertex AI (OpenAI-compatible))","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"vertex-openai/grok-4-7":{"id":"vertex-openai/grok-4-7","name":"Grok 4.7 (Vertex AI (OpenAI-compatible))","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":524288},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"vertex-openai/grok-4-20-non-reasoning":{"id":"vertex-openai/grok-4-20-non-reasoning","name":"Grok 4.20 Non-Reasoning (Vertex AI (OpenAI-compatible))","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"tencent/minimax-m2.7":{"id":"tencent/minimax-m2.7","name":"MiniMax M2.7 (Tencent Cloud)","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"tencent/deepseek-v4-flash":{"id":"tencent/deepseek-v4-flash","name":"DeepSeek V4 Flash (Tencent Cloud)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"tencent/kimi-k2.6":{"id":"tencent/kimi-k2.6","name":"Kimi K2.6 (Tencent Cloud)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.858,"output":3.566,"cache_read":0.145},"canonical_model_id":"moonshotai/kimi-k2.6"},"tencent/hy-mt2-plus":{"id":"tencent/hy-mt2-plus","name":"Hy-MT2 Plus (Tencent Cloud)","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":4096},"cost":{"input":0.074,"output":0.295}},"tencent/deepseek-v4-pro":{"id":"tencent/deepseek-v4-pro","name":"DeepSeek V4 Pro (Tencent Cloud)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.435,"output":0.87,"cache_read":0.00363},"canonical_model_id":"deepseek/deepseek-v4-pro"},"tencent/glm-5.1":{"id":"tencent/glm-5.1","name":"GLM-5.1 (Tencent Cloud)","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"tencent/kimi-k2.7-code-highspeed":{"id":"tencent/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed (Tencent Cloud)","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"tencent/kimi-k2.7-code":{"id":"tencent/kimi-k2.7-code","name":"Kimi K2.7 Code (Tencent Cloud)","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"tencent/kimi-k3":{"id":"tencent/kimi-k3","name":"Kimi K3 (Tencent Cloud)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"tencent/glm-5-turbo":{"id":"tencent/glm-5-turbo","name":"GLM-5 Turbo (Tencent Cloud)","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"tencent/minimax-m3":{"id":"tencent/minimax-m3","name":"MiniMax M3 (Tencent Cloud)","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"tencent/glm-5.2":{"id":"tencent/glm-5.2","name":"GLM-5.2 (Tencent Cloud)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"tencent/glm-5":{"id":"tencent/glm-5","name":"GLM-5 (Tencent Cloud)","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 Preview (Tencent Cloud)","description":"A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0.834,"output":2.501,"cache_read":0.042},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3 (Tencent Cloud)","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.132,"output":0.528,"cache_read":0.033},"canonical_model_id":"tencent/hy3"},"tencent/glm-5v-turbo":{"id":"tencent/glm-5v-turbo","name":"GLM-5V Turbo (Tencent Cloud)","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"novita/qwen3-vl-30b-a3b-instruct":{"id":"novita/qwen3-vl-30b-a3b-instruct","name":"Qwen3 VL 30B A3B Instruct (NovitaAI)","description":"Open vision-language MoE model for image and video understanding, documents, and visual agents","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.2,"output":0.7},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"novita/qwen3.7-max":{"id":"novita/qwen3.7-max","name":"Qwen3.7 Max (NovitaAI)","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.25,"output":3.75,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.7-max"},"novita/qwen3-max":{"id":"novita/qwen3-max","name":"Qwen3 Max (NovitaAI)","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.845,"output":3.38},"canonical_model_id":"alibaba/qwen3-max"},"novita/llama-3.2-3b-instruct":{"id":"novita/llama-3.2-3b-instruct","name":"Llama 3.2 3B Instruct (NovitaAI)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-09-18","last_updated":"2024-09-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":32000},"cost":{"input":0.03,"output":0.05}},"novita/qwen3.8-27b":{"id":"novita/qwen3.8-27b","name":"Qwen3.8 27B (NovitaAI)","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.42,"output":3,"cache_read":0.085},"canonical_model_id":"alibaba/qwen3.8-27b"},"novita/step-3.7-flash":{"id":"novita/step-3.7-flash","name":"Step 3.7 Flash (NovitaAI)","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":256000},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"novita/ling-3.0-flash":{"id":"novita/ling-3.0-flash","name":"InclusionAI Ling 3.0 Flash (NovitaAI)","description":"Hybrid-linear MoE reasoning model with 124B total and 5.1B active parameters for efficient agents","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012},"canonical_model_id":"inclusionai/ling-3.0-flash"},"novita/minimax-m2.7":{"id":"novita/minimax-m2.7","name":"MiniMax M2.7 (NovitaAI)","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"novita/mimo-v2.5":{"id":"novita/mimo-v2.5","name":"MiMo V2.5 (NovitaAI)","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.168,"output":0.336,"cache_read":0.0034,"tiers":[{"input":0.8,"output":4,"cache_read":0.16,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.8,"output":4,"cache_read":0.16}},"canonical_model_id":"xiaomi/mimo-v2.5"},"novita/mimo-v2.5-pro":{"id":"novita/mimo-v2.5-pro","name":"MiMo V2.5 Pro (NovitaAI)","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.522,"output":1.044,"cache_read":0.0043,"tiers":[{"input":2,"output":6,"cache_read":0.4,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.4}},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"novita/qwen3.6-35b-a3b":{"id":"novita/qwen3.6-35b-a3b","name":"Qwen3.6 35B A3B (NovitaAI)","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":64000},"cost":{"input":0.248,"output":1.485},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"novita/qwen3.5-122b-a10b":{"id":"novita/qwen3.5-122b-a10b","name":"Qwen3.5 122B A10B (NovitaAI)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":3.2},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"novita/qwen3.8-2.4t-a95b":{"id":"novita/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B (NovitaAI)","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"novita/deepseek-v4-flash":{"id":"novita/deepseek-v4-flash","name":"DeepSeek V4 Flash (NovitaAI)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":393216},"cost":{"input":0.14,"output":0.28,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v4-flash"},"novita/kimi-k2.6":{"id":"novita/kimi-k2.6","name":"Kimi K2.6 (NovitaAI)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.8,"output":3.4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"novita/deepseek-v3.2":{"id":"novita/deepseek-v3.2","name":"DeepSeek V3.2 (NovitaAI)","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":65536},"cost":{"input":0.269,"output":0.4,"cache_read":0.1345},"canonical_model_id":"deepseek/deepseek-v3.2"},"novita/deepseek-v4.1-flash-1017":{"id":"novita/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (NovitaAI)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"novita/glm-4.5v":{"id":"novita/glm-4.5v","name":"GLM-4.5V (NovitaAI)","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16000},"cost":{"input":0.6,"output":1.8,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5v"},"novita/qwen3-235b-a22b-instruct-2507":{"id":"novita/qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B A22B Instruct 2507 (NovitaAI)","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.09,"output":0.58},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"novita/minimax-m2.1":{"id":"novita/minimax-m2.1","name":"MiniMax M2.1 (NovitaAI)","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.1"},"novita/glm-5.1":{"id":"novita/glm-5.1","name":"GLM-5.1 (NovitaAI)","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1.38,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"novita/qwen3-vl-235b-a22b-thinking":{"id":"novita/qwen3-vl-235b-a22b-thinking","name":"Qwen3 VL 235B A22B Thinking (NovitaAI)","description":"Qwen vision-language thinking model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.98,"output":3.95},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-thinking"},"novita/qwen3.8-max":{"id":"novita/qwen3.8-max","name":"Qwen3.8 Max (NovitaAI)","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max"},"novita/kimi-k2.7-code":{"id":"novita/kimi-k2.7-code","name":"Kimi K2.7 Code (NovitaAI)","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"novita/kimi-k3":{"id":"novita/kimi-k3","name":"Kimi K3 (NovitaAI)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"novita/qwen3-coder-30b-a3b-instruct":{"id":"novita/qwen3-coder-30b-a3b-instruct","name":"Qwen3 Coder 30B A3B Instruct (NovitaAI)","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":160000,"output":32768},"cost":{"input":0.07,"output":0.27},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"novita/glm-4.7":{"id":"novita/glm-4.7","name":"GLM-4.7 (NovitaAI)","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128000},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"novita/llama-3.3-70b-instruct":{"id":"novita/llama-3.3-70b-instruct","name":"Llama 3.3 70B Instruct (NovitaAI)","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":120000},"cost":{"input":0.135,"output":0.4},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"novita/gemma-4-26b-a4b-it":{"id":"novita/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT (NovitaAI)","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.13,"output":0.4},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"novita/qwen3.8-flash":{"id":"novita/qwen3.8-flash","name":"Qwen3.8 Flash (NovitaAI)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-flash"},"novita/llama-4-scout-17b-instruct":{"id":"novita/llama-4-scout-17b-instruct","name":"Llama 4 Scout 17B Instruct (NovitaAI)","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.18,"output":0.59},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"novita/qwen3-next-80b-a3b-instruct":{"id":"novita/qwen3-next-80b-a3b-instruct","name":"Qwen3 Next 80B A3B Instruct (NovitaAI)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":1.5},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"novita/glm-5.3-flash":{"id":"novita/glm-5.3-flash","name":"GLM-5.3 Flash (NovitaAI)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"novita/qwen35-397b-a17b":{"id":"novita/qwen35-397b-a17b","name":"Qwen3.5 397B A17B (NovitaAI)","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":64000},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"novita/llama-4-maverick-17b-instruct":{"id":"novita/llama-4-maverick-17b-instruct","name":"Llama 4 Maverick 17B Instruct (NovitaAI)","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":8192},"cost":{"input":0.27,"output":0.85},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"novita/glm-5.2":{"id":"novita/glm-5.2","name":"GLM-5.2 (NovitaAI)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"novita/deepseek-v4.1-flash":{"id":"novita/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (NovitaAI)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"novita/qwen3-235b-a22b-thinking-2507":{"id":"novita/qwen3-235b-a22b-thinking-2507","name":"Qwen3 235B A22B Thinking 2507 (NovitaAI)","description":"Updated large open Qwen3 MoE thinking model for reasoning, math, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":3},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"novita/qwen3-coder-480b-a35b-instruct":{"id":"novita/qwen3-coder-480b-a35b-instruct","name":"Qwen3 Coder 480B A35B Instruct (NovitaAI)","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.38,"output":1.55},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"novita/minimax-m2.5":{"id":"novita/minimax-m2.5","name":"MiniMax M2.5 (NovitaAI)","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"novita/llama-3-70b-instruct":{"id":"novita/llama-3-70b-instruct","name":"Llama 3 70B Instruct (NovitaAI)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-04-18","last_updated":"2024-04-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8000},"cost":{"input":0.51,"output":0.74}},"novita/qwen3-vl-235b-a22b-instruct":{"id":"novita/qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct (NovitaAI)","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":1.5},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"novita/glm-5":{"id":"novita/glm-5","name":"GLM-5 (NovitaAI)","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"novita/ernie-4.5-vl-424b-a47b":{"id":"novita/ernie-4.5-vl-424b-a47b","name":"ERNIE 4.5 VL 424B A47B (NovitaAI)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ernie","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-06-30","last_updated":"2025-06-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":123000,"output":16000},"cost":{"input":0.42,"output":1.25}},"novita/gemma-4-31b-it":{"id":"novita/gemma-4-31b-it","name":"Gemma 4 31B IT (NovitaAI)","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"novita/glm-4.6":{"id":"novita/glm-4.6","name":"GLM-4.6 (NovitaAI)","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.55,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"novita/hy3":{"id":"novita/hy3","name":"Hy3 (NovitaAI)","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":262144},"cost":{"input":0.14,"output":0.58,"cache_read":0.035},"canonical_model_id":"tencent/hy3"},"novita/glm-5.3":{"id":"novita/glm-5.3","name":"GLM-5.3 (NovitaAI)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"novita/kimi-k2":{"id":"novita/kimi-k2","name":"Kimi K2 (NovitaAI)","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"cost":{"input":0.57,"output":2.3}},"novita/glm-4.6v":{"id":"novita/glm-4.6v","name":"GLM-4.6V (NovitaAI)","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16000},"cost":{"input":0.3,"output":0.9,"cache_read":0.055},"canonical_model_id":"zhipuai/glm-4.6v"},"novita/qwen3-235b-a22b-fp8":{"id":"novita/qwen3-235b-a22b-fp8","name":"Qwen3 235B A22B FP8 (NovitaAI)","description":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":20000},"cost":{"input":0.2,"output":0.8},"canonical_model_id":"alibaba/qwen3-235b-a22b"},"meta-contributor/muse-spark-1.3-contributor":{"id":"meta-contributor/muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor (Meta Contributor)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta-contributor/muse-spark-1.2-contributor":{"id":"meta-contributor/muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor (Meta Contributor)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"quartz/gemini-3.1-pro-preview":{"id":"quartz/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro (Preview) (Quartz)","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"baidu/deepseek-v4-flash":{"id":"baidu/deepseek-v4-flash","name":"DeepSeek V4 Flash (Baidu)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash"},"baidu/kimi-k2.6":{"id":"baidu/kimi-k2.6","name":"Kimi K2.6 (Baidu)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"baidu/deepseek-v4-pro":{"id":"baidu/deepseek-v4-pro","name":"DeepSeek V4 Pro (Baidu)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.32,"output":3.96,"cache_read":0.042},"canonical_model_id":"deepseek/deepseek-v4-pro"},"baidu/glm-5.1":{"id":"baidu/glm-5.1","name":"GLM-5.1 (Baidu)","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"baidu/glm-5.2":{"id":"baidu/glm-5.2","name":"GLM-5.2 (Baidu)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"baidu/deepseek-v4.1-flash":{"id":"baidu/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Baidu)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"baidu/glm-5.3":{"id":"baidu/glm-5.3","name":"GLM-5.3 (Baidu)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo V2.6 Pro (Xiaomi)","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"xiaomi/mimo-v2.6-flash":{"id":"xiaomi/mimo-v2.6-flash","name":"MiMo V2.6 Flash (Xiaomi)","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"vertex-anthropic/claude-opus-4-7":{"id":"vertex-anthropic/claude-opus-4-7","name":"Claude Opus 4.7 (Vertex AI (Anthropic))","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"vertex-anthropic/claude-haiku-4-5":{"id":"vertex-anthropic/claude-haiku-4-5","name":"Claude Haiku 4.5 (Vertex AI (Anthropic))","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"vertex-anthropic/claude-sonnet-4-5":{"id":"vertex-anthropic/claude-sonnet-4-5","name":"Claude Sonnet 4.5 (Vertex AI (Anthropic))","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"vertex-anthropic/claude-sonnet-4-6":{"id":"vertex-anthropic/claude-sonnet-4-6","name":"Claude Sonnet 4.6 (Vertex AI (Anthropic))","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"vertex-anthropic/claude-opus-4-5-20251101":{"id":"vertex-anthropic/claude-opus-4-5-20251101","name":"Claude Opus 4.5 (Vertex AI (Anthropic))","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"vertex-anthropic/claude-opus-4-6":{"id":"vertex-anthropic/claude-opus-4-6","name":"Claude Opus 4.6 (Vertex AI (Anthropic))","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"vertex-anthropic/claude-sonnet-5":{"id":"vertex-anthropic/claude-sonnet-5","name":"Claude Sonnet 5 (Vertex AI (Anthropic))","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"embercloud/glm-4.5-air":{"id":"embercloud/glm-4.5-air","name":"GLM-4.5 Air (EmberCloud)","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":96000},"cost":{"input":0.13,"output":0.85,"cache_read":0.025},"canonical_model_id":"zhipuai/glm-4.5-air"},"embercloud/qwen3-coder-next":{"id":"embercloud/qwen3-coder-next","name":"Qwen3 Coder Next (EmberCloud)","description":"Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.108,"output":0.675,"cache_read":0.06},"canonical_model_id":"alibaba/qwen3-coder-next"},"embercloud/glm-5.1":{"id":"embercloud/glm-5.1","name":"GLM-5.1 (EmberCloud)","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131000},"cost":{"input":0.931,"output":2.93,"cache_read":0.173},"canonical_model_id":"zhipuai/glm-5.1"},"embercloud/glm-4.7":{"id":"embercloud/glm-4.7","name":"GLM-4.7 (EmberCloud)","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131000},"cost":{"input":0.38,"output":1.98,"cache_read":0.19},"canonical_model_id":"zhipuai/glm-4.7"},"embercloud/glm-4.5":{"id":"embercloud/glm-4.5","name":"GLM-4.5 (EmberCloud)","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":96000},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5"},"embercloud/kimi-k2.5":{"id":"embercloud/kimi-k2.5","name":"Kimi K2.5 (EmberCloud)","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.405,"output":1.98,"cache_read":0.225},"canonical_model_id":"moonshotai/kimi-k2.5"},"embercloud/glm-5.2":{"id":"embercloud/glm-5.2","name":"GLM-5.2 (EmberCloud)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131000},"cost":{"input":1.26,"output":3.96,"cache_read":0.234},"canonical_model_id":"zhipuai/glm-5.2"},"embercloud/glm-4.7-flash":{"id":"embercloud/glm-4.7-flash","name":"GLM-4.7 Flash (EmberCloud)","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131000},"cost":{"input":0.06,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flash"},"embercloud/glm-5":{"id":"embercloud/glm-5","name":"GLM-5 (EmberCloud)","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":203000,"output":131000},"cost":{"input":0.72,"output":2.3,"cache_read":0.144},"canonical_model_id":"zhipuai/glm-5"},"moonshot/kimi-k2.6":{"id":"moonshot/kimi-k2.6","name":"Kimi K2.6 (Moonshot AI)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshot/kimi-k2.7-code-highspeed":{"id":"moonshot/kimi-k2.7-code-highspeed","name":"Kimi K2.7 Code Highspeed (Moonshot AI)","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":1.9,"output":8,"cache_read":0.38},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"moonshot/kimi-k2.7-code":{"id":"moonshot/kimi-k2.7-code","name":"Kimi K2.7 Code (Moonshot AI)","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshot/kimi-k3":{"id":"moonshot/kimi-k3","name":"Kimi K3 (Moonshot AI)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"moonshot/kimi-k2.5":{"id":"moonshot/kimi-k2.5","name":"Kimi K2.5 (Moonshot AI)","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.6,"output":3,"cache_read":0.1},"canonical_model_id":"moonshotai/kimi-k2.5"},"deepinfra/qwen3-vl-30b-a3b-instruct":{"id":"deepinfra/qwen3-vl-30b-a3b-instruct","name":"Qwen3 VL 30B A3B Instruct (DeepInfra)","description":"Open vision-language MoE model for image and video understanding, documents, and visual agents","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"deepinfra/ling-3.0-flash":{"id":"deepinfra/ling-3.0-flash","name":"InclusionAI Ling 3.0 Flash (DeepInfra)","description":"Hybrid-linear MoE reasoning model with 124B total and 5.1B active parameters for efficient agents","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012},"canonical_model_id":"inclusionai/ling-3.0-flash"},"deepinfra/ling-3.0-flash-vl":{"id":"deepinfra/ling-3.0-flash-vl","name":"Ling 3.0 Flash VL (DeepInfra)","description":"Native multimodal hybrid MoE model for image and video understanding, reasoning, and visual agents","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012},"canonical_model_id":"inclusionai/ling-3.0-flash-vl"},"deepinfra/deepseek-v4-flash":{"id":"deepinfra/deepseek-v4-flash","name":"DeepSeek V4 Flash (DeepInfra)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.08,"output":0.18,"cache_read":0.016},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepinfra/granite-4.2-8b":{"id":"deepinfra/granite-4.2-8b","name":"Granite 4.2 8B (DeepInfra)","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"granite","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-25","last_updated":"2026-08-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.06,"output":0.25,"cache_read":0.015},"canonical_model_id":"ibm/granite-4.2-8b"},"deepinfra/mimo-v2.6-pro":{"id":"deepinfra/mimo-v2.6-pro","name":"MiMo V2.6 Pro (DeepInfra)","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"deepinfra/mimo-v2.6-flash":{"id":"deepinfra/mimo-v2.6-flash","name":"MiMo V2.6 Flash (DeepInfra)","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":65536},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"deepinfra/deepseek-v3.2":{"id":"deepinfra/deepseek-v3.2","name":"DeepSeek V3.2 (DeepInfra)","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":160000,"output":65536},"cost":{"input":0.26,"output":0.38,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepinfra/deepseek-v4.1-flash-1017":{"id":"deepinfra/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (DeepInfra)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.2,"output":0.6,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepinfra/deepseek-v4-pro":{"id":"deepinfra/deepseek-v4-pro","name":"DeepSeek V4 Pro (DeepInfra)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":1.3,"output":2.6,"cache_read":0.1},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepinfra/muse-glimmer-30b":{"id":"deepinfra/muse-glimmer-30b","name":"Muse Glimmer 30B (DeepInfra)","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.3,"output":1.2,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"deepinfra/nemotron-3-ultra-550b":{"id":"deepinfra/nemotron-3-ultra-550b","name":"Nemotron 3 Ultra 550B (DeepInfra)","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":262144},"cost":{"input":0.5,"output":2.2,"cache_read":0.1}},"deepinfra/gemma-4-26b-a4b-it":{"id":"deepinfra/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT (DeepInfra)","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.07,"output":0.34},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"deepinfra/deepseek-v4.1-flash":{"id":"deepinfra/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (DeepInfra)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.2,"output":0.6,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepinfra/qwen3.5-9b":{"id":"deepinfra/qwen3.5-9b","name":"Qwen3.5 9B (DeepInfra)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.1,"output":0.15},"canonical_model_id":"alibaba/qwen3.5-9b"},"deepinfra/qwen3-vl-235b-a22b-instruct":{"id":"deepinfra/qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct (DeepInfra)","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":0.88,"cache_read":0.11},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"deepinfra/gemma-4-31b-it":{"id":"deepinfra/gemma-4-31b-it","name":"Gemma 4 31B IT (DeepInfra)","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"deepinfra/inkling":{"id":"deepinfra/inkling","name":"Inkling (DeepInfra)","description":"Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":262144},"cost":{"input":0.95,"output":4.05,"cache_read":0.16},"canonical_model_id":"thinkingmachines/inkling"},"entrim/glm-5.3-flash":{"id":"entrim/glm-5.3-flash","name":"GLM-5.3 Flash (entrim)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.1,"output":0.35,"cache_read":0.0175},"canonical_model_id":"zhipuai/glm-5.3-flash"},"entrim/deepseek-v4.1-flash":{"id":"entrim/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (entrim)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.2,"output":0.6,"cache_read":0.005},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"entrim/glm-5.3":{"id":"entrim/glm-5.3","name":"GLM-5.3 (entrim)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.5,"output":2,"cache_read":0.15},"canonical_model_id":"zhipuai/glm-5.3"},"fireworks/kimi-k3":{"id":"fireworks/kimi-k3","name":"Kimi K3 (Fireworks AI)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1040384,"output":1040384},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"fireworks/glm-5.3-flash":{"id":"fireworks/glm-5.3-flash","name":"GLM-5.3 Flash (Fireworks AI)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"fireworks/deepseek-v4.1-flash":{"id":"fireworks/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Fireworks AI)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.22,"output":0.66,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"fireworks/kimi-k3-fast":{"id":"fireworks/kimi-k3-fast","name":"Kimi K3 Fast (Fireworks AI)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1040384,"output":1040384},"cost":{"input":4.5,"output":22.5,"cache_read":0.45},"canonical_model_id":"moonshotai/kimi-k3"},"fireworks/glm-5.3":{"id":"fireworks/glm-5.3","name":"GLM-5.3 (Fireworks AI)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"perplexity/sonar":{"id":"perplexity/sonar","name":"Sonar (Perplexity)","description":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","family":"sonar","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-09-01","release_date":"2024-01-01","last_updated":"2025-09-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":130000,"output":4096},"cost":{"input":0.25,"output":2.5,"cache_write":0.25},"canonical_model_id":"perplexity/sonar"},"inference.net/deepseek-v4.1-flash-1017":{"id":"inference.net/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (Inference.net)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.3,"output":1,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"inference.net/glm-5.3-flash":{"id":"inference.net/glm-5.3-flash","name":"GLM-5.3 Flash (Inference.net)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":128000},"cost":{"input":0.09,"output":0.28,"cache_read":0.02},"canonical_model_id":"zhipuai/glm-5.3-flash"},"inference.net/llama-3.2-11b-instruct":{"id":"inference.net/llama-3.2-11b-instruct","name":"Llama 3.2 11B Instruct (Inference.net)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.07,"output":0.33}},"inference.net/glm-5.3":{"id":"inference.net/glm-5.3","name":"GLM-5.3 (Inference.net)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.9,"output":3,"cache_read":0.15},"canonical_model_id":"zhipuai/glm-5.3"},"together-ai/qwen3.8-2.4t-a95b":{"id":"together-ai/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B (Together AI)","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1010000,"output":909000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"together-ai/deepseek-v4-flash":{"id":"together-ai/deepseek-v4-flash","name":"DeepSeek V4 Flash (Together AI)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":163840},"cost":{"input":0.14,"output":0.28,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4-flash"},"together-ai/deepseek-v4.1-flash-1017":{"id":"together-ai/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (Together AI)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"together-ai/deepseek-v4-pro":{"id":"together-ai/deepseek-v4-pro","name":"DeepSeek V4 Pro (Together AI)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":163840},"cost":{"input":1.32,"output":3.96,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v4-pro"},"together-ai/gpt-oss-120b":{"id":"together-ai/gpt-oss-120b","name":"GPT OSS 120B (Together AI)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"openai/gpt-oss-120b"},"together-ai/muse-glimmer-30b":{"id":"together-ai/muse-glimmer-30b","name":"Muse Glimmer 30B (Together AI)","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.35,"output":1.5,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"together-ai/kimi-k3":{"id":"together-ai/kimi-k3","name":"Kimi K3 (Together AI)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":1000000},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"together-ai/glm-4.7":{"id":"together-ai/glm-4.7","name":"GLM-4.7 (Together AI)","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":128000},"cost":{"input":0.45,"output":2},"canonical_model_id":"zhipuai/glm-4.7"},"together-ai/glm-5.3-flash":{"id":"together-ai/glm-5.3-flash","name":"GLM-5.3 Flash (Together AI)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943717},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"together-ai/minimax-m3":{"id":"together-ai/minimax-m3","name":"MiniMax M3 (Together AI)","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"together-ai/deepseek-v4.1-flash":{"id":"together-ai/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Together AI)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"together-ai/glm-5.3":{"id":"together-ai/glm-5.3","name":"GLM-5.3 (Together AI)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943717},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"aws-bedrock/claude-sonnet-5-5":{"id":"aws-bedrock/claude-sonnet-5-5","name":"Claude Sonnet 5.5 (AWS Bedrock)","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"aws-bedrock/claude-opus-4-8":{"id":"aws-bedrock/claude-opus-4-8","name":"Claude Opus 4.8 (AWS Bedrock)","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"aws-bedrock/grok-4-6":{"id":"aws-bedrock/grok-4-6","name":"Grok 4.6 (AWS Bedrock)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"aws-bedrock/grok-4-7":{"id":"aws-bedrock/grok-4-7","name":"Grok 4.7 (AWS Bedrock)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":524288},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"aws-bedrock/claude-opus-4-7":{"id":"aws-bedrock/claude-opus-4-7","name":"Claude Opus 4.7 (AWS Bedrock)","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"aws-bedrock/claude-haiku-4-5":{"id":"aws-bedrock/claude-haiku-4-5","name":"Claude Haiku 4.5 (AWS Bedrock)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"aws-bedrock/claude-haiku-5-5":{"id":"aws-bedrock/claude-haiku-5-5","name":"Claude Haiku 5.5 (AWS Bedrock)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125},"canonical_model_id":"anthropic/claude-haiku-5-5"},"aws-bedrock/claude-sonnet-4-5":{"id":"aws-bedrock/claude-sonnet-4-5","name":"Claude Sonnet 4.5 (AWS Bedrock)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"aws-bedrock/claude-sonnet-4-6":{"id":"aws-bedrock/claude-sonnet-4-6","name":"Claude Sonnet 4.6 (AWS Bedrock)","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"aws-bedrock/llama-3.1-70b-instruct":{"id":"aws-bedrock/llama-3.1-70b-instruct","name":"Llama 3.1 70B Instruct (AWS Bedrock)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":2048},"status":"deprecated","cost":{"input":0.72,"output":0.72},"canonical_model_id":"meta/llama-3.1-70b-instruct"},"aws-bedrock/claude-fable-5":{"id":"aws-bedrock/claude-fable-5","name":"Claude Fable 5 (AWS Bedrock)","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"aws-bedrock/claude-opus-5-5":{"id":"aws-bedrock/claude-opus-5-5","name":"Claude Opus 5.5 (AWS Bedrock)","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"aws-bedrock/claude-fable-5-1":{"id":"aws-bedrock/claude-fable-5-1","name":"Claude Fable 5.1 (AWS Bedrock)","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"aws-bedrock/kimi-k3":{"id":"aws-bedrock/kimi-k3","name":"Kimi K3 (AWS Bedrock)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"moonshotai/kimi-k3"},"aws-bedrock/claude-opus-4-5-20251101":{"id":"aws-bedrock/claude-opus-4-5-20251101","name":"Claude Opus 4.5 (AWS Bedrock)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1024,"max":31999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5-20251101"},"aws-bedrock/claude-opus-5":{"id":"aws-bedrock/claude-opus-5","name":"Claude Opus 5 (AWS Bedrock)","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"aws-bedrock/claude-opus-4-6":{"id":"aws-bedrock/claude-opus-4-6","name":"Claude Opus 4.6 (AWS Bedrock)","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"aws-bedrock/llama-4-scout-17b-instruct":{"id":"aws-bedrock/llama-4-scout-17b-instruct","name":"Llama 4 Scout 17B Instruct (AWS Bedrock)","description":"Open Llama with long-context vision for efficient multimodal agents","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":2048},"status":"deprecated","cost":{"input":0.17,"output":0.66},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"aws-bedrock/claude-sonnet-4-5-20250929":{"id":"aws-bedrock/claude-sonnet-4-5-20250929","name":"Claude Sonnet 4.5 (2025-09-29) (AWS Bedrock)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]},{"type":"budget_tokens","min":1024,"max":63999}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5-20250929"},"aws-bedrock/claude-haiku-4-5-20251001":{"id":"aws-bedrock/claude-haiku-4-5-20251001","name":"Claude Haiku 4.5 (2025-10-01) (AWS Bedrock)","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5-20251001"},"aws-bedrock/llama-4-maverick-17b-instruct":{"id":"aws-bedrock/llama-4-maverick-17b-instruct","name":"Llama 4 Maverick 17B Instruct (AWS Bedrock)","description":"Open multimodal Llama for strong reasoning with efficient everyday serving","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":2048},"status":"deprecated","cost":{"input":0.24,"output":0.97},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"aws-bedrock/claude-sonnet-5":{"id":"aws-bedrock/claude-sonnet-5","name":"Claude Sonnet 5 (AWS Bedrock)","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"aws-bedrock/grok-4-3":{"id":"aws-bedrock/grok-4-3","name":"Grok 4.3 (AWS Bedrock)","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"minimax/minimax-m2.1-lightning":{"id":"minimax/minimax-m2.1-lightning","name":"MiniMax M2.1 Lightning (MiniMax)","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":196608,"output":131072},"cost":{"input":0.12,"output":0.48}},"minimax/minimax-m2.5-highspeed":{"id":"minimax/minimax-m2.5-highspeed","name":"MiniMax M2.5 Highspeed (MiniMax)","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.6,"output":2.4,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5-highspeed"},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax M2.7 (MiniMax)","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax M2.1 (MiniMax)","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.27,"output":1.1},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-text-01":{"id":"minimax/minimax-text-01","name":"MiniMax Text 01 (MiniMax)","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-01-15","last_updated":"2025-01-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.2,"output":1.1}},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax M2 (MiniMax)","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.2,"output":1,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax M3 (MiniMax)","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":512000,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax M2.5 (MiniMax)","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"minimax/minimax-m3.1-flash":{"id":"minimax/minimax-m3.1-flash","name":"MiniMax M3.1 Flash (MiniMax)","description":"Frontier multimodal coding model with 1M context window and tunable thinking depth","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":524288},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax M2.7 Highspeed (MiniMax)","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131100},"cost":{"input":0.6,"output":2.4,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"sciforium/deepseek-v4.1-flash":{"id":"sciforium/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (sciforium)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.246,"output":0.944,"cache_read":0.012},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"canopywave/deepseek-v4-flash":{"id":"canopywave/deepseek-v4-flash","name":"DeepSeek V4 Flash (CanopyWave)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.14,"output":0.28,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4-flash"},"canopywave/kimi-k2.6":{"id":"canopywave/kimi-k2.6","name":"Kimi K2.6 (CanopyWave)","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"canopywave/deepseek-v4-pro":{"id":"canopywave/deepseek-v4-pro","name":"DeepSeek V4 Pro (CanopyWave)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":1.74,"output":3.48,"cache_read":0.01},"canonical_model_id":"deepseek/deepseek-v4-pro"},"canopywave/kimi-k3":{"id":"canopywave/kimi-k3","name":"Kimi K3 (CanopyWave)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"canopywave/glm-5.2":{"id":"canopywave/glm-5.2","name":"GLM-5.2 (CanopyWave)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":32768},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"xai/grok-4-6":{"id":"xai/grok-4-6","name":"Grok 4.6 (xAI)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.6"},"xai/grok-4-7":{"id":"xai/grok-4-7","name":"Grok 4.7 (xAI)","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5},"canonical_model_id":"xai/grok-4.7"},"xai/grok-4-20-beta-0309-reasoning":{"id":"xai/grok-4-20-beta-0309-reasoning","name":"Grok 4.20 Beta Reasoning (0309) (xAI)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"xai/grok-build-0-1":{"id":"xai/grok-build-0-1","name":"Grok Build 0.1 (xAI)","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}},"canonical_model_id":"xai/grok-build-0.1"},"xai/grok-4-20-beta-0309-non-reasoning":{"id":"xai/grok-4-20-beta-0309-non-reasoning","name":"Grok 4.20 Beta Non-Reasoning (0309) (xAI)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"xai/grok-4-5":{"id":"xai/grok-4-5","name":"Grok 4.5 (xAI)","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3},"canonical_model_id":"xai/grok-4.5"},"xai/grok-4":{"id":"xai/grok-4","name":"Grok 4 (xAI)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":3,"output":15,"cache_read":0.75}},"xai/grok-4-3":{"id":"xai/grok-4-3","name":"Grok 4.3 (xAI)","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":1000000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"scx-ai-gp/qwen3.8-max":{"id":"scx-ai-gp/qwen3.8-max","name":"Qwen3.8 Max (SCX.ai)","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.54,"output":4.631,"cache_read":0.198,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"scx-ai-gp/kimi-k2.7-code":{"id":"scx-ai-gp/kimi-k2.7-code","name":"Kimi K2.7 Code (SCX.ai)","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.737,"output":3.08,"cache_read":0.143},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"scx-ai-gp/kimi-k3":{"id":"scx-ai-gp/kimi-k3","name":"Kimi K3 (SCX.ai)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3.135,"output":15.675,"cache_read":0.319},"canonical_model_id":"moonshotai/kimi-k3"},"scx-ai-gp/glm-5.3-fast":{"id":"scx-ai-gp/glm-5.3-fast","name":"GLM 5.3 Fast (SCX.ai)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":2.8,"output":8.8,"cache_read":0.56},"canonical_model_id":"zhipuai/glm-5.3"},"scx-ai-gp/glm-5.3-flash":{"id":"scx-ai-gp/glm-5.3-flash","name":"GLM-5.3 Flash (SCX.ai)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.088,"output":0.25,"cache_read":0.025},"canonical_model_id":"zhipuai/glm-5.3-flash"},"scx-ai-gp/glm-5.2":{"id":"scx-ai-gp/glm-5.2","name":"GLM-5.2 (SCX.ai)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0.605,"output":2.123,"cache_read":0.154},"canonical_model_id":"zhipuai/glm-5.2"},"scx-ai-gp/deepseek-v4.1-flash":{"id":"scx-ai-gp/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (SCX.ai)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.15,"output":0.6,"cache_read":0.01},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"scx-ai-gp/glm-5.3":{"id":"scx-ai-gp/glm-5.3","name":"GLM-5.3 (SCX.ai)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.924,"output":2.904,"cache_read":0.1716},"canonical_model_id":"zhipuai/glm-5.3"},"scx-ai-gp/glm-5.2-fast":{"id":"scx-ai-gp/glm-5.2-fast","name":"GLM-5.2 Turbo (SCX.ai)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":2.156,"output":6.776,"cache_read":0.429},"canonical_model_id":"zhipuai/glm-5.2"},"runpod/kimi-k3":{"id":"runpod/kimi-k3","name":"Kimi K3 (Runpod)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"alibaba/qwen3.5-35b-a3b":{"id":"alibaba/qwen3.5-35b-a3b","name":"Qwen3.5 35B A3B (Alibaba Cloud)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"alibaba/qwen3.7-max":{"id":"alibaba/qwen3.7-max","name":"Qwen3.7 Max (Alibaba Cloud)","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":65536},"cost":{"input":2.5,"output":7.5,"cache_read":0.5,"cache_write":3.125},"canonical_model_id":"alibaba/qwen3.7-max"},"alibaba/qwen3.7-plus":{"id":"alibaba/qwen3.7-plus","name":"Qwen3.7 Plus (Alibaba Cloud)","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":65536},"cost":{"input":0.4,"output":1.6,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen3.7-plus"},"alibaba/qwen-max":{"id":"alibaba/qwen-max","name":"Qwen Max (Alibaba Cloud)","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":1.6,"output":6.4},"canonical_model_id":"alibaba/qwen-max"},"alibaba/qwen-flash":{"id":"alibaba/qwen-flash","name":"Qwen Flash (Alibaba Cloud)","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":0.05,"output":0.4,"cache_read":0.01,"cache_write":0.0625},"canonical_model_id":"alibaba/qwen-flash"},"alibaba/qwen3.6-35b-a3b":{"id":"alibaba/qwen3.6-35b-a3b","name":"Qwen3.6 35B A3B (Alibaba Cloud)","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.375,"output":2.25},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"alibaba/qwen3.5-122b-a10b":{"id":"alibaba/qwen3.5-122b-a10b","name":"Qwen3.5 122B A10B (Alibaba Cloud)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.4,"output":3.2},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"alibaba/deepseek-v4-flash":{"id":"alibaba/deepseek-v4-flash","name":"DeepSeek V4 Flash (Alibaba Cloud)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.2,"output":0.4,"cache_read":0.04},"canonical_model_id":"deepseek/deepseek-v4-flash"},"alibaba/qwen3-vl-plus":{"id":"alibaba/qwen3-vl-plus","name":"Qwen3 VL Plus (Alibaba Cloud)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.2,"output":1.6,"cache_read":0.04,"cache_write":0.25},"canonical_model_id":"alibaba/qwen3-vl-plus"},"alibaba/deepseek-v4.1-flash-1017":{"id":"alibaba/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017 (Alibaba Cloud)","description":"October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"alibaba/deepseek-v4-pro":{"id":"alibaba/deepseek-v4-pro","name":"DeepSeek V4 Pro (Alibaba Cloud)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":2.4,"output":4.8,"cache_read":0.2},"canonical_model_id":"deepseek/deepseek-v4-pro"},"alibaba/qwen3.7-flash":{"id":"alibaba/qwen3.7-flash","name":"Qwen3.7 Flash (Alibaba Cloud)","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"input":991808,"output":65536},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.0375},"canonical_model_id":"alibaba/qwen3.7-flash"},"alibaba/qwen3.8-max":{"id":"alibaba/qwen3.8-max","name":"Qwen3.8 Max (Alibaba Cloud)","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"alibaba/qwen3.5-27b":{"id":"alibaba/qwen3.5-27b","name":"Qwen3.5 27B (Alibaba Cloud)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":2.4},"canonical_model_id":"alibaba/qwen3.5-27b"},"alibaba/qwen3.6-flash":{"id":"alibaba/qwen3.6-flash","name":"Qwen3.6 Flash (Alibaba Cloud)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.05,"cache_write":0.3125},"canonical_model_id":"alibaba/qwen3.6-flash"},"alibaba/kimi-k3":{"id":"alibaba/kimi-k3","name":"Kimi K3 (Alibaba Cloud)","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":1048576},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"alibaba/qwen3.8-flash":{"id":"alibaba/qwen3.8-flash","name":"Qwen3.8 Flash (Alibaba Cloud)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016,"cache_write":0.2},"canonical_model_id":"alibaba/qwen3.8-flash"},"alibaba/kimi-k2.5":{"id":"alibaba/kimi-k2.5","name":"Kimi K2.5 (Alibaba Cloud)","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.574,"output":3.011},"canonical_model_id":"moonshotai/kimi-k2.5"},"alibaba/qwen35-397b-a17b":{"id":"alibaba/qwen35-397b-a17b","name":"Qwen3.5 397B A17B (Alibaba Cloud)","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"alibaba/glm-5.2":{"id":"alibaba/glm-5.2","name":"GLM-5.2 (Alibaba Cloud)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.28},"canonical_model_id":"zhipuai/glm-5.2"},"alibaba/deepseek-v4.1-flash":{"id":"alibaba/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash (Alibaba Cloud)","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":393216},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"alibaba/qwen3.6-plus":{"id":"alibaba/qwen3.6-plus","name":"Qwen3.6 Plus (Alibaba Cloud)","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":983616,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"cache_write":0.625,"tiers":[{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":2,"output":6,"cache_read":0.2,"cache_write":2.5}},"canonical_model_id":"alibaba/qwen3.6-plus"},"alibaba/glm-5":{"id":"alibaba/glm-5","name":"GLM-5 (Alibaba Cloud)","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":16384},"cost":{"input":0.573,"output":2.58},"canonical_model_id":"zhipuai/glm-5"},"alibaba/qwen-plus":{"id":"alibaba/qwen-plus","name":"Qwen Plus (Alibaba Cloud)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32000},"cost":{"input":0.4,"output":1.2,"cache_read":0.08,"cache_write":0.5},"canonical_model_id":"alibaba/qwen-plus"},"alibaba/glm-5.3":{"id":"alibaba/glm-5.3","name":"GLM-5.3 (Alibaba Cloud)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.28},"canonical_model_id":"zhipuai/glm-5.3"},"alibaba/qwen3-coder-flash":{"id":"alibaba/qwen3-coder-flash","name":"Qwen3 Coder Flash (Alibaba Cloud)","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":997952,"output":65536},"cost":{"input":0.3,"output":1.5,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"alibaba/qwen3-coder-flash"},"alibaba/qwen-plus-latest":{"id":"alibaba/qwen-plus-latest","name":"Qwen Plus Latest (Alibaba Cloud)","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2024-09-09","last_updated":"2024-09-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"cost":{"input":0.4,"output":1.2,"cache_read":0.08,"cache_write":0.5}},"bytedance/seed-2-0-pro-260328":{"id":"bytedance/seed-2-0-pro-260328","name":"Seed 2.0 Pro (260328) (ByteDance)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-28","last_updated":"2026-03-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.5,"output":3,"cache_read":0.1}},"bytedance/seed-1-6-250615":{"id":"bytedance/seed-1-6-250615","name":"Seed 1.6 (250615) (ByteDance)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-06-25","last_updated":"2025-06-25","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"bytedance/seed-2-0-lite-260428":{"id":"bytedance/seed-2-0-lite-260428","name":"Seed 2.0 Lite (260428) (ByteDance)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"bytedance/deepseek-v4-flash":{"id":"bytedance/deepseek-v4-flash","name":"DeepSeek V4 Flash (ByteDance)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.44,"output":1.32,"cache_read":0.014},"canonical_model_id":"deepseek/deepseek-v4-flash"},"bytedance/deepseek-v3.2":{"id":"bytedance/deepseek-v3.2","name":"DeepSeek V3.2 (ByteDance)","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.28,"output":0.42,"cache_read":0.056},"canonical_model_id":"deepseek/deepseek-v3.2"},"bytedance/deepseek-v4-pro":{"id":"bytedance/deepseek-v4-pro","name":"DeepSeek V4 Pro (ByteDance)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":1.32,"output":3.96,"cache_read":0.044},"canonical_model_id":"deepseek/deepseek-v4-pro"},"bytedance/seed-1-6-flash-250715":{"id":"bytedance/seed-1-6-flash-250715","name":"Seed 1.6 Flash (250715) (ByteDance)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-26","last_updated":"2025-07-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.07,"output":0.3,"cache_read":0.015}},"bytedance/gpt-oss-120b":{"id":"bytedance/gpt-oss-120b","name":"GPT OSS 120B (ByteDance)","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"cost":{"input":0.1,"output":0.5,"cache_read":0.02},"canonical_model_id":"openai/gpt-oss-120b"},"bytedance/seed-2-0-code-preview-260328":{"id":"bytedance/seed-2-0-code-preview-260328","name":"Seed 2.0 Code Preview (260328) (ByteDance)","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-28","last_updated":"2026-03-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.5,"output":3,"cache_read":0.1}},"bytedance/glm-4.7":{"id":"bytedance/glm-4.7","name":"GLM-4.7 (ByteDance)","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"bytedance/glm-5.2":{"id":"bytedance/glm-5.2","name":"GLM-5.2 (ByteDance)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"bytedance/seed-1-8-251228":{"id":"bytedance/seed-1-8-251228","name":"Seed 1.8 (251228) (ByteDance)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"bytedance/seed-2-0-mini-260428":{"id":"bytedance/seed-2-0-mini-260428","name":"Seed 2.0 Mini (260428) (ByteDance)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.1,"output":0.4,"cache_read":0.02}},"bytedance/seed-1-6-250915":{"id":"bytedance/seed-1-6-250915","name":"Seed 1.6 (250915) (ByteDance)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-15","last_updated":"2025-09-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":256000},"cost":{"input":0.25,"output":2,"cache_read":0.05}},"zai/glm-4.6v-flashx":{"id":"zai/glm-4.6v-flashx","name":"GLM-4.6V FlashX (Z AI)","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000},"cost":{"input":0.04,"output":0.4,"cache_read":0.004}},"zai/glm-4.5-air":{"id":"zai/glm-4.5-air","name":"GLM-4.5 Air (Z AI)","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":98304},"cost":{"input":0.2,"output":1.1,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-4.5-air"},"zai/glm-4.7-flashx":{"id":"zai/glm-4.7-flashx","name":"GLM-4.7 FlashX (Z AI)","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.07,"output":0.4,"cache_read":0.01},"canonical_model_id":"zhipuai/glm-4.7-flashx"},"zai/glm-4.5v":{"id":"zai/glm-4.5v","name":"GLM-4.5V (Z AI)","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16000},"cost":{"input":0.6,"output":1.8,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5v"},"zai/glm-5.1":{"id":"zai/glm-5.1","name":"GLM-5.1 (Z AI)","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai/glm-4.7":{"id":"zai/glm-4.7","name":"GLM-4.7 (Z AI)","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.7"},"zai/glm-4.5":{"id":"zai/glm-4.5","name":"GLM-4.5 (Z AI)","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5"},"zai/glm-5.3-flash":{"id":"zai/glm-5.3-flash","name":"GLM-5.3 Flash (Z AI)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"zai/glm-4-32b-0414-128k":{"id":"zai/glm-4-32b-0414-128k","name":"GLM-4 32B (0414-128k) (Z AI)","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":0.1,"output":0.1}},"zai/glm-4.5-x":{"id":"zai/glm-4.5-x","name":"GLM-4.5 X (Z AI)","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":2.2,"output":8.9,"cache_read":0.45}},"zai/glm-5.2":{"id":"zai/glm-5.2","name":"GLM-5.2 (Z AI)","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai/glm-5":{"id":"zai/glm-5","name":"GLM-5 (Z AI)","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202800,"output":131100},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"zai/glm-4.5-airx":{"id":"zai/glm-4.5-airx","name":"GLM-4.5 AirX (Z AI)","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000},"cost":{"input":1.1,"output":4.5,"cache_read":0.22}},"zai/glm-4.6":{"id":"zai/glm-4.6","name":"GLM-4.6 (Z AI)","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"zai/glm-5.3":{"id":"zai/glm-5.3","name":"GLM-5.3 (Z AI)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.3"},"zai/glm-4.6v":{"id":"zai/glm-4.6v","name":"GLM-4.6V (Z AI)","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16000},"cost":{"input":0.3,"output":0.9,"cache_read":0.05},"canonical_model_id":"zhipuai/glm-4.6v"},"greenference/qwen3-30b-a3b":{"id":"greenference/qwen3-30b-a3b","name":"Qwen3 30B A3B (greenference)","description":"Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.095,"output":0.285},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"greenference/qwen3-14b":{"id":"greenference/qwen3-14b","name":"Qwen3 14B (greenference)","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":8192},"cost":{"input":0.076,"output":0.228}},"greenference/glm-5.3-flash":{"id":"greenference/glm-5.3-flash","name":"GLM-5.3 Flash (greenference)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.14,"output":0.49,"cache_read":0.028},"canonical_model_id":"zhipuai/glm-5.3-flash"},"greenference/qwen3.6-27b":{"id":"greenference/qwen3.6-27b","name":"Qwen3.6 27B (greenference)","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":8192},"cost":{"input":0.15,"output":1.35},"canonical_model_id":"alibaba/qwen3.6-27b"},"google-vertex/gemini-3.8-flash":{"id":"google-vertex/gemini-3.8-flash","name":"Gemini 3.8 Flash (Google Vertex AI)","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.8-flash"},"google-vertex/gemini-3.1-flash-lite":{"id":"google-vertex/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite (Google Vertex AI)","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"cache_write":0.08333,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google-vertex/gemini-4-argon":{"id":"google-vertex/gemini-4-argon","name":"Gemini 4 Argon (Google Vertex AI)","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":1048576},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-4-argon"},"google-vertex/gemini-3.5-flash-lite":{"id":"google-vertex/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite (Google Vertex AI)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google-vertex/gemini-3.1-pro-preview":{"id":"google-vertex/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro (Preview) (Google Vertex AI)","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google-vertex/gemini-3.6-flash":{"id":"google-vertex/gemini-3.6-flash","name":"Gemini 3.6 Flash (Google Vertex AI)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.6-flash"},"google-vertex/gemini-3.7-flash":{"id":"google-vertex/gemini-3.7-flash","name":"Gemini 3.7 Flash (Google Vertex AI)","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"cache_write":0.08333},"canonical_model_id":"google/gemini-3.7-flash"},"google-vertex/gemini-3.5-flash":{"id":"google-vertex/gemini-3.5-flash","name":"Gemini 3.5 Flash (Google Vertex AI)","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":0.08333,"input_audio":3},"canonical_model_id":"google/gemini-3.5-flash"},"google-vertex/gemini-3-flash-preview":{"id":"google-vertex/gemini-3-flash-preview","name":"Gemini 3 Flash (Preview) (Google Vertex AI)","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"reka-ai/deepseek-v4-pro":{"id":"reka-ai/deepseek-v4-pro","name":"DeepSeek V4 Pro (reka-ai)","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":1.2,"output":3.3,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-v4-pro"},"reka-ai/reka-edge-2603":{"id":"reka-ai/reka-edge-2603","name":"Reka Edge 2603 (reka-ai)","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"reka","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-09","last_updated":"2026-10-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":16384,"output":16384},"cost":{"input":0.1,"output":0.1}},"reka-ai/glm-5.3-flash":{"id":"reka-ai/glm-5.3-flash","name":"GLM-5.3 Flash (reka-ai)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.15,"output":0.5,"cache_read":0.02},"canonical_model_id":"zhipuai/glm-5.3-flash"},"reka-ai/glm-5.3":{"id":"reka-ai/glm-5.3","name":"GLM-5.3 (reka-ai)","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":1.17,"output":3.96,"cache_read":0.234},"canonical_model_id":"zhipuai/glm-5.3"}}},"kimi-code-plan-cn":{"id":"kimi-code-plan-cn","env":["KIMI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.kimi.com/coding/v1","name":"Kimi For Coding (kimi.com)","doc":"https://www.kimi.com/code/docs/en/kimi-code/models.html","models":{"kimi-for-coding":{"id":"kimi-for-coding","name":"Kimi K2.8 Preview","description":"Kimi coding model with more efficient thinking and up to 1M context, available through Kimi Code","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.8-preview"},"kimi-for-coding-highspeed":{"id":"kimi-for-coding-highspeed","name":"Kimi For Coding HighSpeed","description":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.7-code-highspeed"},"k3":{"id":"k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k3"},"k3-256k":{"id":"k3-256k","name":"Kimi K3-256K","description":"256K-context version of Kimi K3, reducing token consumption for shorter coding sessions","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k3"}}},"qiniu-ai":{"id":"qiniu-ai","env":["QINIU_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.qnaigc.com/v1","name":"Qiniu","doc":"https://developer.qiniu.com/aitokenapi","models":{"claude-3.5-haiku":{"id":"claude-3.5-haiku","name":"Claude 3.5 Haiku","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-07-31","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"status":"deprecated","canonical_model_id":"anthropic/claude-3-5-haiku-20241022"},"doubao-seed-2.0-pro":{"id":"doubao-seed-2.0-pro","name":"Doubao Seed 2.0 Pro","description":"Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"status":"deprecated","cost":{"input":0.46376812,"output":2.31884058,"cache_read":0.09275362,"tiers":[{"input":0.69565217,"output":3.47826087,"cache_read":0.13913043,"tier":{"type":"context","size":32000}},{"input":1.39130435,"output":6.95652174,"cache_read":0.27826087,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-pro"},"qwen3-30b-a3b":{"id":"qwen3-30b-a3b","name":"Qwen3 30B A3B","description":"Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40000,"output":4096},"status":"deprecated","cost":{"input":0.10869565,"output":0.43478261},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"gemini-2.5-flash":{"id":"gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":64000},"status":"deprecated","canonical_model_id":"google/gemini-2.5-flash"},"doubao-seed-2.0-code":{"id":"doubao-seed-2.0-code","name":"Doubao Seed 2.0 Code","description":"ByteDance Seed coding model for multimodal software engineering and long-running agents","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"status":"deprecated","cost":{"input":0.46376812,"output":2.31884058,"cache_read":0.09275362,"tiers":[{"input":0.69565217,"output":3.47826087,"cache_read":0.13913043,"tier":{"type":"context","size":32000}},{"input":1.39130435,"output":6.95652174,"cache_read":0.27826087,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"qwen3-max":{"id":"qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.86956522,"output":3.47826087,"tiers":[{"input":1.44927536,"output":5.79710145,"tier":{"type":"context","size":32000}},{"input":2.17391304,"output":8.69565217,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3-max"},"gpt-oss-20b":{"id":"gpt-oss-20b","name":"gpt-oss-20b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","canonical_model_id":"openai/gpt-oss-20b"},"deepseek-v3.1":{"id":"deepseek-v3.1","name":"DeepSeek-V3.1","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"status":"deprecated","cost":{"input":0.57971014,"output":1.73913043},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek-r1-0528":{"id":"deepseek-r1-0528","name":"DeepSeek-R1-0528","description":"May 2025 DeepSeek R1 update with improved mathematics, coding, and reasoning","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"status":"deprecated","cost":{"input":0.57971014,"output":2.31884058},"canonical_model_id":"deepseek/deepseek-r1-0528"},"qwen3-30b-a3b-thinking-2507":{"id":"qwen3-30b-a3b-thinking-2507","name":"Qwen3 30b A3b Thinking 2507","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-04","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":126000,"output":32000},"status":"deprecated","cost":{"input":0.10869565,"output":1.08695652}},"claude-4.0-opus":{"id":"claude-4.0-opus","name":"Claude 4.0 Opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","canonical_model_id":"anthropic/claude-opus-4-0"},"qwen-vl-max-2025-01-25":{"id":"qwen-vl-max-2025-01-25","name":"Qwen VL-MAX-2025-01-25","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.43478261,"output":1.30434783}},"claude-4.5-opus":{"id":"claude-4.5-opus","name":"Claude 4.5 Opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":200000},"status":"deprecated","canonical_model_id":"anthropic/claude-opus-4-5"},"glm-4.5-air":{"id":"glm-4.5-air","name":"GLM 4.5 Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":4096},"status":"deprecated","canonical_model_id":"zhipuai/glm-4.5-air"},"claude-4.1-opus":{"id":"claude-4.1-opus","name":"Claude 4.1 Opus","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"status":"deprecated","canonical_model_id":"anthropic/claude-opus-4-1"},"qwen3-30b-a3b-instruct-2507":{"id":"qwen3-30b-a3b-instruct-2507","name":"Qwen3 30b A3b Instruct 2507","description":"Updated compact Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"status":"deprecated","cost":{"input":0.10869565,"output":0.43478261},"canonical_model_id":"alibaba/qwen3-30b-a3b-instruct-2507"},"claude-4.0-sonnet":{"id":"claude-4.0-sonnet","name":"Claude 4.0 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-22","last_updated":"2025-05-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","canonical_model_id":"anthropic/claude-sonnet-4-0"},"claude-4.5-haiku":{"id":"claude-4.5-haiku","name":"Claude 4.5 Haiku","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","canonical_model_id":"anthropic/claude-haiku-4-5"},"qwen2.5-vl-7b-instruct":{"id":"qwen2.5-vl-7b-instruct","name":"Qwen 2.5 VL 7B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"status":"deprecated"},"gemini-2.5-flash-image":{"id":"gemini-2.5-flash-image","type":"image","name":"Gemini 2.5 Flash Image","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["image"]},"open_weights":false,"limit":{"context":32768,"output":8192},"status":"deprecated","canonical_model_id":"google/gemini-2.5-flash-image"},"deepseek-v3":{"id":"deepseek-v3","name":"DeepSeek-V3","description":"Open DeepSeek MoE chat model for coding, math, and general reasoning","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2024-12-26","last_updated":"2024-12-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16000},"status":"deprecated","cost":{"input":0.28985507,"output":1.15942029},"canonical_model_id":"deepseek/deepseek-v3"},"gemini-2.0-flash":{"id":"gemini-2.0-flash","name":"Gemini 2.0 Flash","description":"Earlier Gemini Flash workhorse for responsive multimodal apps and tool use","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":8192},"status":"deprecated","canonical_model_id":"google/gemini-2.0-flash"},"doubao-1.5-thinking-pro":{"id":"doubao-1.5-thinking-pro","name":"Doubao 1.5 Thinking Pro","description":"Flagship model for demanding analysis, coding, and production agent workflows","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000},"status":"deprecated","cost":{"input":0.57971014,"output":2.31884058}},"doubao-seed-1.6-flash":{"id":"doubao-seed-1.6-flash","name":"Doubao-Seed 1.6 Flash","description":"Low-latency ByteDance Seed model for high-throughput chat, extraction, and lightweight tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"status":"deprecated","cost":{"input":0.02173913,"output":0.2173913,"tiers":[{"input":0.04347826,"output":0.43478261,"tier":{"type":"context","size":32000}},{"input":0.08695652,"output":0.86956522,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-6-flash"},"qwen3-235b-a22b-instruct-2507":{"id":"qwen3-235b-a22b-instruct-2507","name":"Qwen3 235b A22B Instruct 2507","description":"Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":64000},"status":"deprecated","cost":{"input":0.28985507,"output":1.15942029},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"mimo-v2-flash":{"id":"mimo-v2-flash","name":"Mimo-V2-Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12-01","release_date":"2025-12-16","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"status":"deprecated","cost":{"input":0.1,"output":0.3,"cache_read":0.01},"canonical_model_id":"xiaomi/mimo-v2-flash"},"qwen2.5-vl-72b-instruct":{"id":"qwen2.5-vl-72b-instruct","name":"Qwen 2.5 VL 72B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-01-26","last_updated":"2025-01-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"status":"deprecated","cost":{"input":2.31884058,"output":6.95652174},"canonical_model_id":"alibaba/qwen2-5-vl-72b-instruct"},"doubao-1.5-pro-32k":{"id":"doubao-1.5-pro-32k","name":"Doubao 1.5 Pro 32k","description":"Flagship model for demanding analysis, coding, and production agent workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":12000},"status":"deprecated","cost":{"input":0.11594203,"output":0.28985507}},"deepseek-r1":{"id":"deepseek-r1","name":"DeepSeek-R1","description":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-01-20","last_updated":"2025-05-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"status":"deprecated","cost":{"input":0.57971014,"output":2.31884058},"canonical_model_id":"deepseek/deepseek-r1"},"gemini-3.0-pro-image-preview":{"id":"gemini-3.0-pro-image-preview","type":"image","name":"Gemini 3.0 Pro Image Preview","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":8192},"status":"deprecated","canonical_model_id":"google/gemini-3-pro-image-preview"},"gpt-oss-120b":{"id":"gpt-oss-120b","name":"gpt-oss-120b","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4096},"status":"deprecated","canonical_model_id":"openai/gpt-oss-120b"},"doubao-seed-2.0-lite":{"id":"doubao-seed-2.0-lite","name":"Doubao Seed 2.0 Lite","description":"Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.08695652,"output":0.52173913,"cache_read":0.0173913,"tiers":[{"input":0.13043478,"output":0.7826087,"cache_read":0.02608696,"tier":{"type":"context","size":32000}},{"input":0.26086957,"output":1.56521739,"cache_read":0.05217391,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"doubao-seed-1.6":{"id":"doubao-seed-1.6","name":"Doubao-Seed 1.6","description":"ByteDance Seed model for long-context reasoning, instruction following, and tool-assisted tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"status":"deprecated","cost":{"input":0.11594203,"output":1.15942029,"tiers":[{"input":0.17391304,"output":2.31884058,"tier":{"type":"context","size":32000}},{"input":0.34782609,"output":3.47826087,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-6"},"gemini-2.0-flash-lite":{"id":"gemini-2.0-flash-lite","name":"Gemini 2.0 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06","release_date":"2024-12-11","last_updated":"2024-12-11","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":8192},"status":"deprecated","canonical_model_id":"google/gemini-2.0-flash-lite"},"claude-3.5-sonnet":{"id":"claude-3.5-sonnet","name":"Claude 3.5 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-04-30","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8192},"status":"deprecated","canonical_model_id":"anthropic/claude-3-5-sonnet-20241022"},"qwen3-235b-a22b":{"id":"qwen3-235b-a22b","name":"Qwen 3 235B A22B","description":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"status":"deprecated","cost":{"input":0.28985507,"output":1.15942029},"canonical_model_id":"alibaba/qwen3-235b-a22b"},"qwen3.5-397b-a17b":{"id":"qwen3.5-397b-a17b","name":"Qwen3.5 397B A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0.17391304,"output":1.04347826,"tiers":[{"input":0.43478261,"output":2.60869565,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen3-next-80b-a3b-thinking":{"id":"qwen3-next-80b-a3b-thinking","name":"Qwen3 Next 80B A3B Thinking","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.14492754,"output":1.44927536},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"doubao-1.5-vision-pro":{"id":"doubao-1.5-vision-pro","name":"Doubao 1.5 Vision Pro","description":"Flagship model for demanding analysis, coding, and production agent workflows","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16000},"status":"deprecated","cost":{"input":0.43478261,"output":1.30434783}},"qwen-turbo":{"id":"qwen-turbo","name":"Qwen-Turbo","description":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-11-01","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":4096},"status":"deprecated","cost":{"input":0.04347826,"output":0.08695652},"canonical_model_id":"alibaba/qwen-turbo"},"qwen3-max-preview":{"id":"qwen3-max-preview","name":"Qwen3 Max Preview","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-05","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":64000},"status":"deprecated","cost":{"input":0.86956522,"output":3.47826087,"tiers":[{"input":1.44927536,"output":5.79710145,"tier":{"type":"context","size":32000}},{"input":2.17391304,"output":8.69565217,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3-max-preview"},"glm-4.5":{"id":"glm-4.5","name":"GLM 4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"status":"deprecated","cost":{"input":0.57971014,"output":2.31884058},"canonical_model_id":"zhipuai/glm-4.5"},"qwen3-next-80b-a3b-instruct":{"id":"qwen3-next-80b-a3b-instruct","name":"Qwen3 Next 80B A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.14492754,"output":0.57971014},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"qwen3-vl-30b-a3b-thinking":{"id":"qwen3-vl-30b-a3b-thinking","name":"Qwen3-Vl 30b A3b Thinking","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-09","last_updated":"2026-02-09","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000},"status":"deprecated","cost":{"input":0.10869565,"output":1.08695652}},"claude-3.7-sonnet":{"id":"claude-3.7-sonnet","name":"Claude 3.7 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2025-02-19","last_updated":"2025-02-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000},"status":"deprecated","canonical_model_id":"anthropic/claude-3-7-sonnet-20250219"},"doubao-seed-1.6-thinking":{"id":"doubao-seed-1.6-thinking","name":"Doubao-Seed 1.6 Thinking","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-15","last_updated":"2025-08-15","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"status":"deprecated","cost":{"input":0.11594203,"output":1.15942029,"tiers":[{"input":0.17391304,"output":2.31884058,"tier":{"type":"context","size":32000}},{"input":0.34782609,"output":3.47826087,"tier":{"type":"context","size":128000}}]}},"qwen3-32b":{"id":"qwen3-32b","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40000,"output":4096},"status":"deprecated","cost":{"input":0.28985507,"output":1.15942029},"canonical_model_id":"alibaba/qwen3-32b"},"kling-v2-6":{"id":"kling-v2-6","type":"video","name":"Kling-V2 6","description":"Video model for prompt-guided generation, editing, and motion workflows","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-01-13","last_updated":"2026-01-13","modalities":{"input":["text","image","video"],"output":["video"]},"open_weights":false,"limit":{"context":99999999,"output":99999999}},"doubao-seed-2.0-mini":{"id":"doubao-seed-2.0-mini","name":"Doubao Seed 2.0 Mini","description":"Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":32000},"cost":{"input":0.02898551,"output":0.28985507,"cache_read":0.0057971,"tiers":[{"input":0.05797101,"output":0.57971014,"cache_read":0.0115942,"tier":{"type":"context","size":32000}},{"input":0.11594203,"output":1.15942029,"cache_read":0.02318841,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"qwen3-235b-a22b-thinking-2507":{"id":"qwen3-235b-a22b-thinking-2507","name":"Qwen3 235B A22B Thinking 2507","description":"Updated large open Qwen3 MoE thinking model for reasoning, math, coding, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":4096},"status":"deprecated","cost":{"input":0.28985507,"output":2.89855072},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"qwen-max-2025-01-25":{"id":"qwen-max-2025-01-25","name":"Qwen2.5-Max-2025-01-25","description":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-04-03","last_updated":"2025-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"status":"deprecated","cost":{"input":0.34782609,"output":1.39130435},"canonical_model_id":"alibaba/qwen-max"},"qwen3-coder-480b-a35b-instruct":{"id":"qwen3-coder-480b-a35b-instruct","name":"Qwen3 Coder 480B A35B Instruct","description":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262000,"output":4096},"status":"deprecated","cost":{"input":0.86956522,"output":3.47826087,"tiers":[{"input":1.30434783,"output":5.2173913,"tier":{"type":"context","size":32000}},{"input":2.17391304,"output":8.69565217,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"gemini-3.0-pro-preview":{"id":"gemini-3.0-pro-preview","name":"Gemini 3.0 Pro Preview","description":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-18","last_updated":"2025-11-18","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","canonical_model_id":"google/gemini-3-pro-preview"},"gemini-3.0-flash-preview":{"id":"gemini-3.0-flash-preview","name":"Gemini 3.0 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","canonical_model_id":"google/gemini-3-flash-preview"},"gemini-2.5-flash-lite":{"id":"gemini-2.5-flash-lite","name":"Gemini 2.5 Flash Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":64000},"status":"deprecated","canonical_model_id":"google/gemini-2.5-flash-lite"},"gemini-2.5-pro":{"id":"gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","canonical_model_id":"google/gemini-2.5-pro"},"kimi-k2":{"id":"kimi-k2","name":"Kimi K2","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-12-31","release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":128000},"status":"deprecated","cost":{"input":0.57971014,"output":2.31884058},"canonical_model_id":"moonshotai/kimi-k2"},"claude-4.5-sonnet":{"id":"claude-4.5-sonnet","name":"Claude 4.5 Sonnet","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"status":"deprecated","canonical_model_id":"anthropic/claude-sonnet-4-5"},"MiniMax-M1":{"id":"MiniMax-M1","name":"MiniMax M1","description":"Open hybrid-attention MoE reasoning model with a million-token context and an 80K thinking budget","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":80000},"status":"deprecated","cost":{"input":0.57971014,"output":2.31884058},"canonical_model_id":"minimax/minimax-m1-80k"},"deepseek-v3-0324":{"id":"deepseek-v3-0324","name":"DeepSeek-V3-0324","description":"March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16000},"status":"deprecated","canonical_model_id":"deepseek/deepseek-v3-0324"},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"Z-Ai/GLM 4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":200000},"status":"deprecated","cost":{"input":0.45913043,"output":1.81565217},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"Z-Ai/GLM 5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"cost":{"input":0.57971014,"output":2.60869565,"tiers":[{"input":0.86956522,"output":3.1884058,"tier":{"type":"context","size":32000}}]},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-4.6":{"id":"z-ai/glm-4.6","name":"Z-AI/GLM 4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":200000},"status":"deprecated","cost":{"input":1.04347826,"output":1.82608696},"canonical_model_id":"zhipuai/glm-4.6"},"z-ai/autoglm-phone-9b":{"id":"z-ai/autoglm-phone-9b","name":"Z-Ai/Autoglm Phone 9b","description":"GLM vision model for visual reasoning, documents, and multimodal agents","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":12800,"output":4096},"status":"deprecated"},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":100000},"status":"deprecated","cost":{"input":0.57971014,"output":2.31884058},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 0905","description":"September 2025 Kimi K2 instruction update with a 256K context and improved agentic coding","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":100000},"status":"deprecated","canonical_model_id":"moonshotai/kimi-k2-0905"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Moonshotai/Kimi-K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"status":"deprecated","cost":{"input":0.57971014,"output":3.04347826},"canonical_model_id":"moonshotai/kimi-k2.5"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"OpenAI/GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5":{"id":"openai/gpt-5","name":"OpenAI/GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"status":"deprecated","canonical_model_id":"openai/gpt-5"},"deepseek/deepseek-v3.2-exp":{"id":"deepseek/deepseek-v3.2-exp","name":"DeepSeek/DeepSeek-V3.2-Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000},"status":"deprecated","cost":{"input":0.28985507,"output":0.43478261}},"deepseek/deepseek-v3.1-terminus":{"id":"deepseek/deepseek-v3.1-terminus","name":"DeepSeek/DeepSeek-V3.1-Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-22","last_updated":"2025-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000},"status":"deprecated","cost":{"input":0.57971014,"output":1.73913043}},"deepseek/deepseek-v3.1-terminus-thinking":{"id":"deepseek/deepseek-v3.1-terminus-thinking","name":"DeepSeek/DeepSeek-V3.1-Terminus-Thinking","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-22","last_updated":"2025-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000},"status":"deprecated","cost":{"input":0.57971014,"output":1.73913043}},"deepseek/deepseek-v3.2-exp-thinking":{"id":"deepseek/deepseek-v3.2-exp-thinking","name":"DeepSeek/DeepSeek-V3.2-Exp-Thinking","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000},"status":"deprecated","cost":{"input":0.28985507,"output":0.43478261}},"deepseek/deepseek-v3.2-251201":{"id":"deepseek/deepseek-v3.2-251201","name":"Deepseek/DeepSeek-V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":32000},"status":"deprecated","cost":{"input":0.28985507,"output":0.43478261,"cache_read":0.02898551},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-math-v2":{"id":"deepseek/deepseek-math-v2","name":"Deepseek/Deepseek-Math-V2","description":"DeepSeek chat model for instruction following, coding, and analysis","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-12-04","last_updated":"2025-12-04","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":160000,"output":160000},"status":"deprecated"},"xiaomi/mimo-v2-flash":{"id":"xiaomi/mimo-v2-flash","name":"Xiaomi/Mimo-V2-Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-12-01","release_date":"2025-12-16","last_updated":"2026-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":256000},"status":"deprecated","cost":{"input":0.1,"output":0.3,"cache_read":0.01},"canonical_model_id":"xiaomi/mimo-v2-flash"},"stepfun/step-3.5-flash":{"id":"stepfun/step-3.5-flash","name":"Stepfun/Step-3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","family":"step","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":64000,"input":256000,"output":4096},"status":"deprecated","canonical_model_id":"stepfun/step-3.5-flash"},"stepfun-ai/gelab-zero-4b-preview":{"id":"stepfun-ai/gelab-zero-4b-preview","name":"Stepfun-Ai/Gelab Zero 4b Preview","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":4096},"status":"deprecated"},"minimax/minimax-m2.5-highspeed":{"id":"minimax/minimax-m2.5-highspeed","name":"Minimax/Minimax-M2.5 Highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128000},"status":"deprecated","cost":{"input":0.60869565,"output":2.43478261},"canonical_model_id":"minimax/MiniMax-M2.5-highspeed"},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"Minimax/Minimax-M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128000},"status":"deprecated","cost":{"input":0.30434783,"output":1.2173913},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"Minimax/Minimax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":128000},"status":"deprecated","cost":{"input":0.30434783,"output":1.2173913},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"Minimax/Minimax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":128000},"cost":{"input":0.30434783,"output":1.2173913},"canonical_model_id":"minimax/MiniMax-M2.5"},"meituan/longcat-flash-lite":{"id":"meituan/longcat-flash-lite","name":"Meituan/Longcat-Flash-Lite","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-06","last_updated":"2026-02-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":320000},"status":"deprecated","cost":{"input":0,"output":0}},"meituan/longcat-flash-chat":{"id":"meituan/longcat-flash-chat","name":"Meituan/Longcat-Flash-Chat","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-11-05","last_updated":"2025-11-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":131072},"status":"deprecated"},"x-ai/grok-4-fast-non-reasoning":{"id":"x-ai/grok-4-fast-non-reasoning","name":"X-Ai/Grok-4-Fast-Non-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"status":"deprecated","canonical_model_id":"xai/grok-4-fast-non-reasoning"},"x-ai/grok-4.1-fast-reasoning":{"id":"x-ai/grok-4.1-fast-reasoning","name":"X-Ai/Grok 4.1 Fast Reasoning","description":"xAI's fast agentic tool-calling model with a 2M context window and built-in reasoning","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":20000000,"output":2000000},"status":"deprecated","canonical_model_id":"xai/grok-4.1-fast-reasoning"},"x-ai/grok-4.1-fast":{"id":"x-ai/grok-4.1-fast","name":"x-AI/Grok-4.1-Fast","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"status":"deprecated"},"x-ai/grok-4-fast-reasoning":{"id":"x-ai/grok-4-fast-reasoning","name":"X-Ai/Grok-4-Fast-Reasoning","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-18","last_updated":"2025-12-18","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"status":"deprecated","canonical_model_id":"xai/grok-4-fast-reasoning"},"x-ai/grok-4-fast":{"id":"x-ai/grok-4-fast","name":"x-AI/Grok-4-Fast","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-20","last_updated":"2025-09-20","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"status":"deprecated"},"x-ai/grok-code-fast-1":{"id":"x-ai/grok-code-fast-1","name":"x-AI/Grok-Code-Fast 1","description":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","family":"grok","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-09-02","last_updated":"2025-09-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":10000},"status":"deprecated","canonical_model_id":"xai/grok-code-fast-1"},"x-ai/grok-4.1-fast-non-reasoning":{"id":"x-ai/grok-4.1-fast-non-reasoning","name":"X-Ai/Grok 4.1 Fast Non Reasoning","description":"xAI's fast agentic tool-calling model with a 2M context window; non-reasoning variant for low-latency responses","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-11-19","last_updated":"2025-11-19","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":2000000},"status":"deprecated","canonical_model_id":"xai/grok-4.1-fast"}}},"xiaomi-token-plan-ams":{"id":"xiaomi-token-plan-ams","env":["XIAOMI_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://token-plan-ams.xiaomimimo.com/v1","name":"Xiaomi Token Plan (Europe)","doc":"https://platform.xiaomimimo.com/#/docs","models":{"mimo-v2.6-pro":{"id":"mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"mimo-v2.6-flash":{"id":"mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo Flash model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"mimo-v2.5-tts-voiceclone":{"id":"mimo-v2.5-tts-voiceclone","type":"speech","name":"MiMo-V2.5-TTS-VoiceClone","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.5-tts-voiceclone"},"mimo-v2-pro":{"id":"mimo-v2-pro","name":"MiMo-V2-Pro","description":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2-pro"},"mimo-v2.5-tts":{"id":"mimo-v2.5-tts","type":"speech","name":"MiMo-V2.5-TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.5-tts"},"mimo-v2.5-tts-voicedesign":{"id":"mimo-v2.5-tts-voicedesign","type":"speech","name":"MiMo-V2.5-TTS-VoiceDesign","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2.5-tts-voicedesign"},"mimo-v2-tts":{"id":"mimo-v2-tts","type":"speech","name":"MiMo-V2-TTS","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"mimo","attachment":false,"reasoning":false,"tool_call":false,"temperature":false,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["audio"]},"open_weights":false,"limit":{"context":8192,"output":8192},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"xiaomi/mimo-v2-tts"},"mimo-v2.5-pro":{"id":"mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"mimo-v2.5":{"id":"mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"xiaomi/mimo-v2.5"}}},"submodel":{"id":"submodel","env":["SUBMODEL_INSTAGEN_ACCESS_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://llm.submodel.ai/v1","name":"submodel","doc":"https://submodel.gitbook.io","models":{"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.1,"output":0.5},"canonical_model_id":"openai/gpt-oss-120b"},"zai-org/GLM-4.5-FP8":{"id":"zai-org/GLM-4.5-FP8","name":"GLM 4.5 FP8","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.2,"output":0.8},"canonical_model_id":"zhipuai/glm-4.5"},"zai-org/GLM-4.5-Air":{"id":"zai-org/GLM-4.5-Air","name":"GLM 4.5 Air","description":"Efficient GLM model for fast reasoning, coding, and agent workflows","family":"glm-air","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":131072},"cost":{"input":0.1,"output":0.5},"canonical_model_id":"zhipuai/glm-4.5-air"},"deepseek-ai/DeepSeek-R1-0528":{"id":"deepseek-ai/DeepSeek-R1-0528","name":"DeepSeek R1 0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":75000,"output":163840},"cost":{"input":0.2,"output":0.8},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek-ai/DeepSeek-V3.1":{"id":"deepseek-ai/DeepSeek-V3.1","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":75000,"output":163840},"cost":{"input":0.2,"output":0.8},"canonical_model_id":"deepseek/deepseek-v3.1"},"deepseek-ai/DeepSeek-V3-0324":{"id":"deepseek-ai/DeepSeek-V3-0324","name":"DeepSeek V3 0324","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":75000,"output":163840},"cost":{"input":0.2,"output":0.8},"canonical_model_id":"deepseek/deepseek-v3-0324"},"Qwen/Qwen3-235B-A22B-Thinking-2507":{"id":"Qwen/Qwen3-235B-A22B-Thinking-2507","name":"Qwen3 235B A22B Thinking 2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.2,"output":0.6},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"Qwen/Qwen3-235B-A22B-Instruct-2507":{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.2,"output":0.3},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8":{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8","name":"Qwen3 Coder 480B A35B Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.2,"output":0.8},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"}}},"openrouter":{"id":"openrouter","env":["OPENROUTER_API_KEY"],"npm":"@openrouter/ai-sdk-provider","api":"https://openrouter.ai/api/v1","name":"OpenRouter","doc":"https://openrouter.ai/models","models":{"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.3-contributor":{"id":"meta/muse-spark-1.3-contributor","name":"Muse Spark 1.3 Contributor","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/muse-glimmer-30b":{"id":"meta/muse-glimmer-30b","name":"Muse Glimmer 30B","description":"Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-08-10","last_updated":"2026-08-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.3,"output":1.2,"cache_read":0.04},"canonical_model_id":"meta/muse-glimmer-30b"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"meta/muse-wonder":{"id":"meta/muse-wonder","name":"Muse Wonder","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":4,"output":20,"cache_read":0.2}},"meta/muse-spark-1.2-contributor":{"id":"meta/muse-spark-1.2-contributor","name":"Muse Spark 1.2 Contributor","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-08-21","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.1,"output":0.2,"cache_read":0.002}},"meta/muse-spark-open":{"id":"meta/muse-spark-open","name":"Muse Spark Open","description":"Open Llama multimodal model for image understanding and text reasoning","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"~unbiased/pareto-latest":{"id":"~unbiased/pareto-latest","name":"Pareto Latest","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.8,"output":3.2,"cache_read":0.03}},"sakana/sakana-namazu":{"id":"sakana/sakana-namazu","name":"Sakana Namazu","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"sakana-namazu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.95,"output":4,"cache_read":0.15},"canonical_model_id":"sakana/sakana-namazu"},"sakana/fugu-ultra-v2":{"id":"sakana/fugu-ultra-v2","name":"Fugu Ultra v2","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-08-28","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra-v2"},"sakana/fugu-max":{"id":"sakana/fugu-max","name":"Fugu Max","description":"Multi-agent model for routing expert agents across complex analytical tasks","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"sakana/fugu-max"},"sakana/fugu-ultra":{"id":"sakana/fugu-ultra","name":"Fugu Ultra","description":"Quality-first multi-agent model for hard research, analysis, and competitions","family":"fugu","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"sakana/fugu-ultra"},"z-ai/glm-4.5-air":{"id":"z-ai/glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.13,"output":0.85,"cache_read":0.025},"canonical_model_id":"zhipuai/glm-4.5-air"},"z-ai/glm-4.5v":{"id":"z-ai/glm-4.5v","name":"GLM-4.5V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-08-11","last_updated":"2025-08-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":16384},"cost":{"input":0.6,"output":1.8,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5v"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-5.3-prime":{"id":"z-ai/glm-5.3-prime","name":"GLM 5.3 Prime","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2.8,"output":8.8,"cache_read":0.56}},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.54,"output":1.98,"cache_read":0.099},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-5-turbo":{"id":"z-ai/glm-5-turbo","name":"GLM-5-Turbo","description":"Faster GLM-5 lane for coding agents that need lower latency","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":false,"temperature":true,"release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5-turbo"},"z-ai/glm-4.5":{"id":"z-ai/glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"status":"deprecated","cost":{"input":0.6,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.5"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943717},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":0.0341,"output":7,"cache_read":0.033},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-4.7-flash":{"id":"z-ai/glm-4.7-flash","name":"GLM-4.7-Flash","description":"Budget GLM lane for fast coding help, routing, and everyday automation","family":"glm-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-01-19","last_updated":"2026-01-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":117964},"cost":{"input":0.0605,"output":0.4},"canonical_model_id":"zhipuai/glm-4.7-flash"},"z-ai/glm-5.3-flashx":{"id":"z-ai/glm-5.3-flashx","name":"GLM 5.3 FlashX","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.37,"output":1.25,"cache_read":0.09}},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.2},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-4.6":{"id":"z-ai/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.55,"output":2.2,"cache_read":0.11},"canonical_model_id":"zhipuai/glm-4.6"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0.7,"output":2.2,"cache_read":0.13},"canonical_model_id":"zhipuai/glm-5.3"},"z-ai/glm-5v-turbo":{"id":"z-ai/glm-5v-turbo","name":"GLM-5V-Turbo","description":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-04-01","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":202752,"output":131072},"cost":{"input":1.2,"output":4,"cache_read":0.24},"canonical_model_id":"zhipuai/glm-5v-turbo"},"z-ai/glm-4.6v":{"id":"z-ai/glm-4.6v","name":"GLM-4.6V","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.3,"output":0.9,"cache_read":0.055},"canonical_model_id":"zhipuai/glm-4.6v"},"moonshotai/kimi-k2-thinking":{"id":"moonshotai/kimi-k2-thinking","name":"Kimi K2 Thinking","description":"Thinking Kimi model for slower research passes, planning, and hard technical questions","family":"kimi-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-11-06","last_updated":"2025-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.6,"output":2.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2-thinking"},"moonshotai/kimi-k2.6":{"id":"moonshotai/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.65,"output":3.41,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2.6"},"moonshotai/kimi-k2-0905":{"id":"moonshotai/kimi-k2-0905","name":"Kimi K2 Instruct 0905","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-09-05","last_updated":"2025-09-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":98304},"cost":{"input":0.6,"output":2.5},"canonical_model_id":"moonshotai/kimi-k2-0905"},"moonshotai/kimi-k2.7-code":{"id":"moonshotai/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.6712,"output":3.35,"cache_read":0.18},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":0.239,"output":11.936,"cache_read":0.21},"canonical_model_id":"moonshotai/kimi-k3"},"moonshotai/kimi-k2.5":{"id":"moonshotai/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.5,"output":2.5,"cache_read":0.15},"canonical_model_id":"moonshotai/kimi-k2.5"},"moonshotai/kimi-k2":{"id":"moonshotai/kimi-k2","name":"Kimi K2 0711","description":"Kimi model for long-context chat, coding, and agentic reasoning","family":"kimi-k2","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-12-31","release_date":"2025-07-11","last_updated":"2025-07-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.57,"output":2.3},"canonical_model_id":"moonshotai/kimi-k2"},"mistralai/devstral-2512":{"id":"mistralai/devstral-2512","name":"Devstral 2","description":"Mistral coding agent model for repository tasks and software engineering workflows","family":"devstral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-12","release_date":"2025-12-09","last_updated":"2025-12-09","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/devstral-2512"},"mistralai/mistral-saba":{"id":"mistralai/mistral-saba","name":"Saba","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-02-17","last_updated":"2025-02-17","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":26214},"cost":{"input":0.2,"output":0.6,"cache_read":0.02},"canonical_model_id":"mistral/mistral-saba"},"mistralai/ministral-3b-2512":{"id":"mistralai/ministral-3b-2512","name":"Ministral 3 3B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":104857},"cost":{"input":0.1,"output":0.1,"cache_read":0.01},"canonical_model_id":"mistral/ministral-3b-2512"},"mistralai/mixtral-8x22b-instruct":{"id":"mistralai/mixtral-8x22b-instruct","name":"Mixtral 8x22B Instruct","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-01-31","release_date":"2024-04-17","last_updated":"2024-04-17","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":52428},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mixtral-8x22b-instruct"},"mistralai/mistral-medium-3":{"id":"mistralai/mistral-medium-3","name":"Mistral Medium 3","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-05-07","last_updated":"2025-05-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/mistral-medium-3"},"mistralai/mistral-small-3.1-24b-instruct":{"id":"mistralai/mistral-small-3.1-24b-instruct","name":"Mistral Small 3.1 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-03-17","last_updated":"2025-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":102400},"cost":{"input":0.351,"output":0.555},"canonical_model_id":"mistral/mistral-small-3.1-24b-instruct"},"mistralai/ministral-8b-2512":{"id":"mistralai/ministral-8b-2512","name":"Ministral 3 8B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.15,"output":0.15,"cache_read":0.015},"canonical_model_id":"mistral/ministral-8b-2512"},"mistralai/mistral-small-2603":{"id":"mistralai/mistral-small-2603","name":"Mistral Small 4","description":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","family":"mistral-small","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-16","last_updated":"2026-03-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"mistral/mistral-small-2603"},"mistralai/voxtral-small-24b-2507":{"id":"mistralai/voxtral-small-24b-2507","name":"Voxtral Small 24B 2507","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"voxtral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-15","last_updated":"2025-07-15","modalities":{"input":["text","audio","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":26214},"cost":{"input":0.1,"output":0.3,"cache_read":0.01},"canonical_model_id":"mistral/voxtral-small-24b-2507"},"mistralai/mistral-large-4-0":{"id":"mistralai/mistral-large-4-0","name":"Mistral Large 4","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":262144},"cost":{"input":1.36,"output":4.18,"cache_read":0.14},"canonical_model_id":"mistral/mistral-large-4"},"mistralai/codestral-2508":{"id":"mistralai/codestral-2508","name":"Codestral 25.08","description":"Mistral coding model for code completion, generation, and developer workflows","family":"codestral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-30","last_updated":"2025-07-30","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":204800},"cost":{"input":0.3,"output":0.9,"cache_read":0.03},"canonical_model_id":"mistral/codestral-2508"},"mistralai/mistral-medium-3.1":{"id":"mistralai/mistral-medium-3.1","name":"Mistral Medium 3.1","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-30","release_date":"2025-08-12","last_updated":"2025-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":104857},"cost":{"input":0.4,"output":2,"cache_read":0.04},"canonical_model_id":"mistral/mistral-medium-3.1"},"mistralai/ministral-14b-2512":{"id":"mistralai/ministral-14b-2512","name":"Ministral 3 14B 2512","description":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","family":"ministral","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.2,"output":0.2,"cache_read":0.02},"canonical_model_id":"mistral/ministral-14b-2512"},"mistralai/mistral-nemo":{"id":"mistralai/mistral-nemo","name":"Mistral Nemo","description":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","family":"mistral-nemo","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.029,"output":0.03},"canonical_model_id":"mistral/mistral-nemo"},"mistralai/mistral-large-2407":{"id":"mistralai/mistral-large-2407","name":"Mistral Large 2407","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-03-31","release_date":"2024-07-24","last_updated":"2024-07-24","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":104857},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mistral-large-2407"},"mistralai/mistral-large":{"id":"mistralai/mistral-large","name":"Mistral Large","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11-30","release_date":"2024-02-26","last_updated":"2024-02-26","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":102400},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"mistral/mistral-large"},"mistralai/mistral-small-3.2-24b-instruct":{"id":"mistralai/mistral-small-3.2-24b-instruct","name":"Mistral Small 3.2 24B","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-06-20","last_updated":"2025-06-20","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":16384},"cost":{"input":0.09375,"output":0.25},"canonical_model_id":"mistral/mistral-small-3.2-24b-instruct"},"mistralai/mistral-medium-3-5":{"id":"mistralai/mistral-medium-3-5","name":"Mistral Medium 3.5","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral-medium","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-30","last_updated":"2026-04-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":209715},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"mistral/mistral-medium-3-5"},"mistralai/mistral-small-24b-instruct-2501":{"id":"mistralai/mistral-small-24b-instruct-2501","name":"Mistral Small 3","description":"Efficient Mistral model for fast chat, extraction, and production assistants","family":"mistral-small","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-10-31","release_date":"2025-01-30","last_updated":"2025-01-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.05,"output":0.08},"canonical_model_id":"mistral/mistral-small-24b-instruct-2501"},"mistralai/mistral-large-2512":{"id":"mistralai/mistral-large-2512","name":"Mistral Large 3","description":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","family":"mistral-large","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-11","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":209715},"cost":{"input":0.5,"output":1.5,"cache_read":0.05},"canonical_model_id":"mistral/mistral-large-2512"},"bytedance-seed/seed-2-1-turbo":{"id":"bytedance-seed/seed-2-1-turbo","name":"Seed 2.1 Turbo","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-23","last_updated":"2026-06-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":235929},"cost":{"input":0.5,"output":2.5},"canonical_model_id":"bytedance-seed/seed-2.1-turbo"},"bytedance-seed/seed-1.6-flash":{"id":"bytedance-seed/seed-1.6-flash","name":"Seed 1.6 Flash","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.075,"output":0.3,"tiers":[{"input":0.1,"output":0.8,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-6-flash"},"bytedance-seed/seed-1.6":{"id":"bytedance-seed/seed-1.6","name":"Seed 1.6","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.25,"output":2,"tiers":[{"input":0.5,"output":4,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-1-6"},"bytedance-seed/seed-2.0-mini":{"id":"bytedance-seed/seed-2.0-mini","name":"Seed 2.0 Mini","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.1,"output":0.4,"tiers":[{"input":0.2,"output":0.8,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-mini"},"bytedance-seed/seed-2.0-code":{"id":"bytedance-seed/seed-2.0-code","name":"Seed 2.0 Code","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"status":"deprecated","cost":{"input":0.5,"output":3,"tiers":[{"input":1,"output":6,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-code"},"bytedance-seed/seed-2.0-lite":{"id":"bytedance-seed/seed-2.0-lite","name":"Seed 2.0 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"seed","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-14","last_updated":"2026-02-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.25,"output":2,"tiers":[{"input":0.5,"output":4,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"bytedance-seed/seed-2.0-lite"},"gryphe/mythomax-l2-13b":{"id":"gryphe/mythomax-l2-13b","name":"MythoMax L2 13B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2023-08-10","last_updated":"2023-08-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":3686},"cost":{"input":0.08,"output":0.11},"canonical_model_id":"gryphe/mythomax-l2-13b"},"undi95/remm-slerp-l2-13b":{"id":"undi95/remm-slerp-l2-13b","name":"ReMM SLERP 13B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-06-30","release_date":"2023-07-22","last_updated":"2023-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":6144,"output":5529},"cost":{"input":0.35,"output":0.65},"canonical_model_id":"undi95/remm-slerp-l2-13b"},"inclusionai/ling-3.0-flash":{"id":"inclusionai/ling-3.0-flash","name":"Ling 3.0 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-23","last_updated":"2026-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.021,"output":0.063,"cache_read":0.0042},"canonical_model_id":"inclusionai/ling-3.0-flash"},"inclusionai/ling-3.0-flash-vl":{"id":"inclusionai/ling-3.0-flash-vl","name":"Ling 3.0 Flash VL","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.06,"output":0.18,"cache_read":0.012},"canonical_model_id":"inclusionai/ling-3.0-flash-vl"},"inclusionai/ling-3.1-flash":{"id":"inclusionai/ling-3.1-flash","name":"Ling 3.1 Flash","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"ling","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.075,"output":0.225,"cache_read":0.015},"canonical_model_id":"inclusionai/ling-3.1-flash"},"liquid/lfm-2.5-2.6b:free":{"id":"liquid/lfm-2.5-2.6b:free","name":"LFM2.5-2.6B (free)","description":"Free provider route for experiments, demos, and cost-sensitive chat workloads","family":"liquid","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":8192},"cost":{"input":0,"output":0}},"liquid/lfm-2.5-24b-a2b":{"id":"liquid/lfm-2.5-24b-a2b","name":"LFM2.5-24B-A2B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"liquid","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.05,"output":0.7,"cache_read":0.025}},"poolside/laguna-xs-2.1:free":{"id":"poolside/laguna-xs-2.1:free","name":"Laguna XS 2.1 (free)","description":"Free provider route for experiments, demos, and cost-sensitive chat workloads","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-02","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-xs-2.1"},"poolside/laguna-xs-2.1":{"id":"poolside/laguna-xs-2.1","name":"Laguna XS 2.1","description":"Agentic coding model from Poolside in the XS size class for local deployment","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-02","last_updated":"2026-07-02","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0.06,"output":0.12,"cache_read":0.03},"canonical_model_id":"poolside/laguna-xs-2.1"},"poolside/laguna-s-2.1":{"id":"poolside/laguna-s-2.1","name":"Laguna S 2.1","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":0.09,"output":0.18,"cache_read":0.009},"canonical_model_id":"poolside/laguna-s-2.1"},"poolside/laguna-s-2.1:free":{"id":"poolside/laguna-s-2.1:free","name":"Laguna S 2.1 (free)","description":"Free provider route for experiments, demos, and cost-sensitive chat workloads","family":"laguna","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"status":"deprecated","cost":{"input":0,"output":0},"canonical_model_id":"poolside/laguna-s-2.1"},"thinkingmachines/inkling:free":{"id":"thinkingmachines/inkling:free","name":"Inkling (free)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/inkling-small":{"id":"thinkingmachines/inkling-small","name":"Inkling Small","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":262144},"cost":{"input":0.45,"output":1.2,"cache_read":0.1},"canonical_model_id":"thinkingmachines/inkling-small"},"thinkingmachines/inkling":{"id":"thinkingmachines/inkling","name":"Inkling","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":471859},"cost":{"input":1,"output":4.05,"cache_read":0.17},"canonical_model_id":"thinkingmachines/inkling"},"thinkingmachines/inkling-small:free":{"id":"thinkingmachines/inkling-small:free","name":"Inkling Small (free)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"ling","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-30","last_updated":"2026-07-30","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0,"output":0},"canonical_model_id":"thinkingmachines/inkling-small"},"cognitivecomputations/dolphin-mistral-24b-venice-edition":{"id":"cognitivecomputations/dolphin-mistral-24b-venice-edition","name":"Uncensored","description":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","family":"mistral","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-04-30","release_date":"2025-07-09","last_updated":"2025-07-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0.2,"output":0.9},"canonical_model_id":"cognitivecomputations/dolphin-mistral-24b-venice-edition"},"thedrummer/unslopnemo-12b":{"id":"thedrummer/unslopnemo-12b","name":"UnslopNemo 12B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04-30","release_date":"2024-11-08","last_updated":"2024-11-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1024000,"output":819200},"cost":{"input":0.4,"output":0.4},"canonical_model_id":"thedrummer/unslopnemo-12b"},"thedrummer/skyfall-36b-v2":{"id":"thedrummer/skyfall-36b-v2","name":"Skyfall 36B V2","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2025-03-10","last_updated":"2025-03-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":29491},"cost":{"input":0.55,"output":0.8,"cache_read":0.25},"canonical_model_id":"thedrummer/skyfall-36b-v2"},"thedrummer/cydonia-24b-v4.1":{"id":"thedrummer/cydonia-24b-v4.1","name":"Cydonia 24B V4.1","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04-30","release_date":"2025-09-27","last_updated":"2025-09-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.3,"output":0.5,"cache_read":0.15},"canonical_model_id":"thedrummer/cydonia-24b-v4.1"},"morph/morph-v3-large":{"id":"morph/morph-v3-large","name":"Morph V3 Large","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-07","last_updated":"2025-07-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.9,"output":1.9},"canonical_model_id":"morph/morph-v3-large"},"morph/morph-v3-fast":{"id":"morph/morph-v3-fast","name":"Morph V3 Fast","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"morph","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-07-07","last_updated":"2025-07-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":0.8,"output":1.2},"canonical_model_id":"morph/morph-v3-fast"},"aion-labs/aion-3.5":{"id":"aion-labs/aion-3.5","name":"Aion 3.5","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":3,"output":6,"cache_read":0.75}},"aion-labs/aion-3.0-mini":{"id":"aion-labs/aion-3.0-mini","name":"Aion-3.0-Mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-14","last_updated":"2026-05-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.7,"output":1.4,"cache_read":0.18},"canonical_model_id":"aion-labs/aion-3.0-mini"},"aion-labs/aion-2.0":{"id":"aion-labs/aion-2.0","name":"Aion-2.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-21","last_updated":"2025-12-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.8,"output":1.6,"cache_read":0.2},"canonical_model_id":"aion-labs/aion-2.0"},"aion-labs/aion-rp-llama-3.1-8b":{"id":"aion-labs/aion-rp-llama-3.1-8b","name":"Aion-RP 1.0 (8B)","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12-31","release_date":"2025-02-04","last_updated":"2025-02-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":29491},"cost":{"input":0.8,"output":1.6},"canonical_model_id":"aion-labs/aion-rp-llama-3.1-8b"},"aion-labs/aion-3.5-mini":{"id":"aion-labs/aion-3.5-mini","name":"Aion 3.5 Mini","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":32768},"cost":{"input":0.7,"output":1.4,"cache_read":0.18}},"aion-labs/aion-3.0":{"id":"aion-labs/aion-3.0","name":"Aion-3.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-05-05","last_updated":"2026-05-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":3,"output":6,"cache_read":0.75},"canonical_model_id":"aion-labs/aion-3.0"},"inception/mercury-2.5":{"id":"inception/mercury-2.5","name":"Mercury 2.5","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":260000,"output":65536},"cost":{"input":0.04,"output":0.15,"cache_read":0.004},"canonical_model_id":"inception/mercury-2.5"},"inception/mercury-3":{"id":"inception/mercury-3","name":"Mercury 3","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.1,"output":0.35,"cache_read":0.01}},"inception/mercury-2":{"id":"inception/mercury-2","name":"Mercury 2","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"mercury","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-24","last_updated":"2026-02-24","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":50000},"cost":{"input":0.25,"output":0.75,"cache_read":0.025},"canonical_model_id":"inception/mercury-2"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-4o-2024-05-13":{"id":"openai/gpt-4o-2024-05-13","name":"GPT-4o (2024-05-13)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-05-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":5,"output":15},"canonical_model_id":"openai/gpt-4o-2024-05-13"},"openai/gpt-6-sol-pro":{"id":"openai/gpt-6-sol-pro","name":"GPT-6 Sol Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}}},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-audio-mini":{"id":"openai/gpt-audio-mini","type":"chat","name":"GPT-Audio Mini","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-01","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.6,"output":2.4},"canonical_model_id":"openai/gpt-audio-mini"},"openai/gpt-oss-20b":{"id":"openai/gpt-oss-20b","name":"GPT OSS 20B","description":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.018,"output":0.09,"cache_read":0.009},"canonical_model_id":"openai/gpt-oss-20b"},"openai/gpt-6.1-sol-pro":{"id":"openai/gpt-6.1-sol-pro","name":"GPT-6.1 Sol Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}}},"openai/gpt-4":{"id":"openai/gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8191,"output":4096},"cost":{"input":30,"output":60},"canonical_model_id":"openai/gpt-4"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-6-luna-pro":{"id":"openai/gpt-6-luna-pro","name":"GPT-6 Luna Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}}},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/o1-pro":{"id":"openai/o1-pro","name":"o1-pro","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2025-03-19","last_updated":"2025-03-19","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":150,"output":600},"canonical_model_id":"openai/o1-pro"},"openai/gpt-chat-latest":{"id":"openai/gpt-chat-latest","name":"GPT Chat Latest","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-05-05","last_updated":"2026-05-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5}},"openai/gpt-5.2-chat":{"id":"openai/gpt-5.2-chat","name":"GPT-5.2 Chat","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":32000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-chat-latest"},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/gpt-5.4-image-2":{"id":"openai/gpt-5.4-image-2","type":"image","name":"GPT-5.4 Image 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":false,"structured_output":true,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["image","text","pdf"],"output":["image","text"]},"open_weights":false,"limit":{"context":272000,"output":128000},"cost":{"input":8,"output":15,"cache_read":2},"canonical_model_id":"openai/gpt-5.4-image-2"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-3.5-turbo-instruct":{"id":"openai/gpt-3.5-turbo-instruct","name":"GPT-3.5 Turbo Instruct","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2021-09-30","release_date":"2023-09-28","last_updated":"2023-09-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4095,"output":3685},"cost":{"input":1.5,"output":2}},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-oss-safeguard-20b":{"id":"openai/gpt-oss-safeguard-20b","name":"GPT OSS Safeguard 20B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-29","last_updated":"2025-10-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-oss-safeguard-20b"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-audio":{"id":"openai/gpt-audio","type":"chat","name":"GPT-Audio","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-10-01","release_date":"2025-08-28","last_updated":"2025-08-28","modalities":{"input":["text","audio"],"output":["text","audio"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10},"canonical_model_id":"openai/gpt-audio"},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-3.5-turbo-0613":{"id":"openai/gpt-3.5-turbo-0613","name":"GPT-3.5 Turbo (older v0613)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2021-09-30","release_date":"2024-01-25","last_updated":"2024-01-25","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":4095,"output":3685},"cost":{"input":1,"output":2}},"openai/gpt-6-astra-pro":{"id":"openai/gpt-6-astra-pro","name":"GPT-6 Astra Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}}},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":65536},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-5.6-luna-pro":{"id":"openai/gpt-5.6-luna-pro","name":"GPT-5.6 Luna Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-5-image-mini":{"id":"openai/gpt-5-image-mini","type":"image","name":"GPT-5 Image Mini","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-10-16","last_updated":"2025-10-16","modalities":{"input":["pdf","image","text"],"output":["image","text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":2.5,"output":2,"cache_read":0.25},"canonical_model_id":"openai/gpt-5-image-mini"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4o-2024-11-20":{"id":"openai/gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"openai/gpt-5.6-sol-pro":{"id":"openai/gpt-5.6-sol-pro","name":"GPT-5.6 Sol Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-4o-2024-08-06":{"id":"openai/gpt-4o-2024-08-06","name":"GPT-4o (2024-08-06)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-08-06"},"openai/gpt-3.5-turbo-16k":{"id":"openai/gpt-3.5-turbo-16k","name":"GPT-3.5 Turbo 16k","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2021-09-30","release_date":"2023-08-28","last_updated":"2023-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":3,"output":4}},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/gpt-5-image":{"id":"openai/gpt-5-image","type":"image","name":"GPT-5 Image","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-10-01","release_date":"2025-10-14","last_updated":"2025-10-14","modalities":{"input":["image","text","pdf"],"output":["image","text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":10,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-5-image"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5-turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":0.5,"output":1.5},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/o1":{"id":"openai/o1","name":"o1","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-09","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":15,"output":60,"cache_read":7.5},"canonical_model_id":"openai/o1"},"openai/gpt-5.1-codex-max":{"id":"openai/gpt-5.1-codex-max","name":"GPT-5.1 Codex Max","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex-max"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.03},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/o4-mini-high":{"id":"openai/o4-mini-high","name":"o4 Mini High","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-06-30","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.275}},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/o3-pro":{"id":"openai/o3-pro","name":"o3-pro","description":"High-effort o3 tier for difficult technical reasoning and careful answers","family":"o-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-06-10","last_updated":"2025-06-10","modalities":{"input":["text","pdf","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":20,"output":80},"canonical_model_id":"openai/o3-pro"},"openai/gpt-4o-mini-2024-07-18":{"id":"openai/gpt-4o-mini-2024-07-18","name":"GPT-4o-mini (2024-07-18)","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.13},"canonical_model_id":"openai/gpt-5.1-codex"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2 Pro","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["image","text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/o3-mini-high":{"id":"openai/o3-mini-high","name":"o3 Mini High","description":"O-series reasoning model for hard analysis, math, coding, and planning","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2023-10-31","release_date":"2025-02-12","last_updated":"2025-02-12","modalities":{"input":["text","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"cost":{"input":1.1,"output":4.4,"cache_read":0.55}},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.6-terra-pro":{"id":"openai/gpt-5.6-terra-pro","name":"GPT-5.6 Terra Pro","description":"Frontier GPT model for professional reasoning, coding, and multimodal work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-r1-0528":{"id":"deepseek/deepseek-r1-0528","name":"DeepSeek R1 0528","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-05-28","last_updated":"2025-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.5,"output":2.15,"cache_read":0.35},"canonical_model_id":"deepseek/deepseek-r1-0528"},"deepseek/deepseek-v3.2-exp":{"id":"deepseek/deepseek-v3.2-exp","name":"DeepSeek V3.2 Exp","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":147456},"cost":{"input":0.27,"output":0.41}},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943717},"cost":{"input":0.091,"output":0.182,"cache_read":0.0182},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v3.1-terminus":{"id":"deepseek/deepseek-v3.1-terminus","name":"DeepSeek V3.1 Terminus","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-22","last_updated":"2025-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":147456},"cost":{"input":0.27,"output":1}},"deepseek/deepseek-v3.2":{"id":"deepseek/deepseek-v3.2","name":"DeepSeek V3.2","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":65536},"cost":{"input":0.28,"output":0.42,"cache_read":0.028},"canonical_model_id":"deepseek/deepseek-v3.2"},"deepseek/deepseek-v4.1-flash-1017":{"id":"deepseek/deepseek-v4.1-flash-1017","name":"DeepSeek V4.1 Flash 1017","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.15,"output":0.6,"cache_read":0.003},"canonical_model_id":"deepseek/deepseek-v4.1-flash-1017"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-chat-v3-0324":{"id":"deepseek/deepseek-chat-v3-0324","name":"DeepSeek V3 0324","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-07-31","release_date":"2025-03-24","last_updated":"2025-03-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":115200},"cost":{"input":0.29,"output":1.14,"cache_read":0.11},"canonical_model_id":"deepseek/deepseek-chat-v3-0324"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":262144},"cost":{"input":0.2156,"output":0.6468,"cache_read":0.00686},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","high","xhigh"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":384000},"cost":{"input":0.38019,"output":0.76038,"cache_read":0.031683},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-chat":{"id":"deepseek/deepseek-chat","name":"DeepSeek Chat","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":16000},"cost":{"input":0.2574,"output":1.0287},"canonical_model_id":"deepseek/deepseek-chat"},"deepseek/deepseek-v4.1-flash":{"id":"deepseek/deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.0342,"output":0.1368,"cache_read":0.000684},"canonical_model_id":"deepseek/deepseek-v4.1-flash"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.04444,"output":0.13332,"cache_read":0.001414},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"deepseek/deepseek-chat-v3.1":{"id":"deepseek/deepseek-chat-v3.1","name":"DeepSeek V3.1","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-21","last_updated":"2025-08-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":32768},"cost":{"input":0.25,"output":0.95,"cache_read":0.13},"canonical_model_id":"deepseek/deepseek-chat-v3.1"},"anthropic/claude-opus-4.1":{"id":"anthropic/claude-opus-4.1","name":"Claude Opus 4.1 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":32000},"cost":{"input":15,"output":75,"cache_read":1.5,"cache_write":18.75},"canonical_model_id":"anthropic/claude-opus-4-1"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25,"tiers":[{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":10,"output":37.5,"cache_read":1,"cache_write":12.5}},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75,"tiers":[{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":6,"output":22.5,"cache_read":0.6,"cache_write":7.5}},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"~x-ai/grok-latest":{"id":"~x-ai/grok-latest","name":"Grok Latest","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}}},"nousresearch/hermes-3-llama-3.1-405b":{"id":"nousresearch/hermes-3-llama-3.1-405b","name":"Hermes 3 405B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-16","last_updated":"2024-08-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":1,"output":1},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-405b"},"nousresearch/hermes-4-405b":{"id":"nousresearch/hermes-4-405b","name":"Hermes 4 405B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"hermes","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":1,"output":3},"canonical_model_id":"nousresearch/hermes-4-405b"},"nousresearch/hermes-3-llama-3.1-70b":{"id":"nousresearch/hermes-3-llama-3.1-70b","name":"Hermes 3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"nousresearch","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-18","last_updated":"2024-08-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"status":"deprecated","cost":{"input":0.7,"output":0.7},"canonical_model_id":"nousresearch/hermes-3-llama-3.1-70b"},"~z-ai/glm-latest":{"id":"~z-ai/glm-latest","name":"GLM Latest","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-19","last_updated":"2026-08-19","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.079,"output":1.4,"cache_read":0.078}},"~z-ai/glm-flash-latest":{"id":"~z-ai/glm-flash-latest","name":"GLM Flash Latest","description":"GLM vision model for visual reasoning, documents, and multimodal agents","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.0413,"output":0.0773,"cache_read":0.03}},"tencent/hy-mt2-30b-a3b":{"id":"tencent/hy-mt2-30b-a3b","name":"Hy-MT2-30B-A3B","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.074,"output":0.295},"canonical_model_id":"tencent/hy-mt2-30b-a3b"},"tencent/hunyuan-a13b-instruct":{"id":"tencent/hunyuan-a13b-instruct","name":"Hunyuan A13B Instruct","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"hunyuan","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-07-08","last_updated":"2025-07-08","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.14,"output":0.57}},"tencent/hy3-preview":{"id":"tencent/hy3-preview","name":"Hy3 preview","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.18,"output":0.6,"cache_read":0.06},"canonical_model_id":"tencent/hy3-preview"},"tencent/hy-mt2-7b":{"id":"tencent/hy-mt2-7b","name":"Hy-MT2-7B","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.074,"output":0.295},"canonical_model_id":"tencent/hy-mt2-7b"},"tencent/hy4-preview":{"id":"tencent/hy4-preview","name":"Hy4 preview","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-28","last_updated":"2026-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":64000},"cost":{"input":0.7506,"output":2.2509,"cache_read":0.0378},"canonical_model_id":"tencent/hy4-preview"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":192000,"output":128000},"cost":{"input":0.0825,"output":0.33,"cache_read":0.020625},"canonical_model_id":"tencent/hy3"},"tencent/hy-mt2-1.8b":{"id":"tencent/hy-mt2-1.8b","name":"Hy-MT2-1.8B","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":4096},"cost":{"input":0.044,"output":0.177},"canonical_model_id":"tencent/hy-mt2-1.8b"},"ibm-granite/granite-4.0-h-micro":{"id":"ibm-granite/granite-4.0-h-micro","name":"Granite 4.0 Micro","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"granite","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-10-20","last_updated":"2025-10-20","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":117900},"cost":{"input":0.017,"output":0.112}},"ibm-granite/granite-4.2-8b":{"id":"ibm-granite/granite-4.2-8b","name":"Granite 4.2 8B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"granite","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-25","last_updated":"2026-08-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.06,"output":0.25,"cache_read":0.015},"canonical_model_id":"ibm/granite-4.2-8b"},"moondream/moondream3.1-9b-a2b":{"id":"moondream/moondream3.1-9b-a2b","name":"Moondream 3.1","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.3,"output":2.5}},"moondream/moondream3-preview":{"id":"moondream/moondream3-preview","name":"Moondream 3 Preview","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":4096},"cost":{"input":0.3,"output":2.5}},"meta-llama/llama-3.1-8b-instruct":{"id":"meta-llama/llama-3.1-8b-instruct","name":"Llama-3.1-8B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.05,"output":0.08,"cache_read":0.025},"canonical_model_id":"meta/llama-3.1-8b-instruct"},"meta-llama/llama-3.2-3b-instruct":{"id":"meta-llama/llama-3.2-3b-instruct","name":"Llama 3.2 3B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":72000},"cost":{"input":0.0509,"output":0.335},"canonical_model_id":"meta/llama-3.2-3b-instruct"},"meta-llama/llama-guard-4-12b":{"id":"meta-llama/llama-guard-4-12b","name":"Llama Guard 4 12B","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"llama","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-29","last_updated":"2025-04-29","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":16384},"cost":{"input":0.18,"output":0.18},"canonical_model_id":"meta/llama-guard-4-12b"},"meta-llama/llama-3.1-70b-instruct":{"id":"meta-llama/llama-3.1-70b-instruct","name":"Llama-3.1-70B-Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-07-23","last_updated":"2024-07-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.4,"output":0.4},"canonical_model_id":"meta/llama-3.1-70b-instruct"},"meta-llama/llama-3.2-1b-instruct":{"id":"meta-llama/llama-3.2-1b-instruct","name":"Llama 3.2 1B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2024-09-25","last_updated":"2024-09-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":60000,"output":54000},"cost":{"input":0.027,"output":0.201},"canonical_model_id":"meta/llama-3.2-1b-instruct"},"meta-llama/llama-3.3-70b-instruct":{"id":"meta-llama/llama-3.3-70b-instruct","name":"Llama-3.3-70B-Instruct","description":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-12","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.1,"output":0.32},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"meta-llama/llama-4-scout":{"id":"meta-llama/llama-4-scout","name":"Llama 4 Scout","description":"Open multimodal Llama model for long-context analysis and efficient agents","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1310720,"output":16384},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"meta/llama-4-scout-17b-instruct"},"meta-llama/llama-4-maverick":{"id":"meta-llama/llama-4-maverick","name":"Llama 4 Maverick","description":"Open multimodal Llama model for strong reasoning and fast responses","family":"llama","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-04-05","last_updated":"2025-04-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":524288,"output":16384},"cost":{"input":0.1875,"output":0.6525,"cache_read":0.05},"canonical_model_id":"meta/llama-4-maverick-17b-instruct"},"xiaomi/mimo-v2.6-pro-ultraspeed":{"id":"xiaomi/mimo-v2.6-pro-ultraspeed","name":"MiMo-V2.6-Pro-UltraSpeed","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":4.35,"output":8.7,"cache_read":0.036},"canonical_model_id":"xiaomi/mimo-v2.6-pro-ultraspeed"},"xiaomi/mimo-v2.5":{"id":"xiaomi/mimo-v2.5","name":"MiMo-V2.5","description":"Open MiMo model for multimodal coding agents and long-context automation","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":945000},"status":"deprecated","cost":{"input":0.119,"output":0.238,"cache_read":0.00255},"canonical_model_id":"xiaomi/mimo-v2.5"},"xiaomi/mimo-v2.5-pro":{"id":"xiaomi/mimo-v2.5-pro","name":"MiMo-V2.5-Pro","description":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","family":"mimo","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-12","release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":945000},"status":"deprecated","cost":{"input":0.3045,"output":0.609,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.5-pro"},"xiaomi/mimo-v2.6-pro":{"id":"xiaomi/mimo-v2.6-pro","name":"MiMo-V2.6-Pro","description":"MiMo pro model for strong multimodal reasoning and agent execution","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":131072},"cost":{"input":0.435,"output":0.87,"cache_read":0.0036},"canonical_model_id":"xiaomi/mimo-v2.6-pro"},"xiaomi/mimo-v2.6-flash":{"id":"xiaomi/mimo-v2.6-flash","name":"MiMo-V2.6-Flash","description":"MiMo flash model for fast multimodal assistance and agent workflows","family":"mimo","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":true,"limit":{"context":1050000,"output":131072},"cost":{"input":0.14,"output":0.28,"cache_read":0.0028},"canonical_model_id":"xiaomi/mimo-v2.6-flash"},"~google/gemini-pro-latest":{"id":"~google/gemini-pro-latest","name":"Gemini Pro Latest","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["audio","pdf","image","text","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":2,"output":10,"reasoning":10,"cache_read":0.1,"cache_write":0.1875,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}}},"~google/gemini-flash-latest":{"id":"~google/gemini-flash-latest","name":"Gemini Flash Latest","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01-01","release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667}},"anthracite-org/magnum-v4-72b":{"id":"anthracite-org/magnum-v4-72b","name":"Magnum v4 72B","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-10-22","last_updated":"2024-10-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":4096},"cost":{"input":2.5,"output":5},"canonical_model_id":"anthracite-org/magnum-v4-72b"},"stepfun/step-3.7-flash":{"id":"stepfun/step-3.7-flash","name":"Step 3.7 Flash","description":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03-01","release_date":"2026-05-29","last_updated":"2026-05-29","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"input":256000,"output":230400},"cost":{"input":0.2,"output":1.15,"cache_read":0.04},"canonical_model_id":"stepfun/step-3.7-flash"},"stepfun/step-5-preview":{"id":"stepfun/step-5-preview","name":"Step 5 Preview","description":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-16","last_updated":"2026-09-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"input":1000000,"output":64000},"cost":{"input":1,"output":2.7,"cache_read":0.1},"canonical_model_id":"stepfun/step-5-preview"},"stepfun/step-3.5-flash":{"id":"stepfun/step-3.5-flash","name":"Step 3.5 Flash","description":"StepFun flash lane for quick multimodal reasoning and coding assistance","family":"step","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-01-29","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.1,"output":0.3},"canonical_model_id":"stepfun/step-3.5-flash"},"~openai/gpt-terra-latest":{"id":"~openai/gpt-terra-latest","name":"GPT Terra Latest","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}}},"~openai/gpt-mini-latest":{"id":"~openai/gpt-mini-latest","name":"GPT Mini Latest","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"output":128000},"cost":{"input":0.75,"output":4.5,"cache_read":0.075}},"~openai/gpt-astra-latest":{"id":"~openai/gpt-astra-latest","name":"GPT Astra Latest","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}}},"~openai/gpt-luna-latest":{"id":"~openai/gpt-luna-latest","name":"GPT Luna Latest","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}}},"~openai/gpt-sol-latest":{"id":"~openai/gpt-sol-latest","name":"GPT Sol Latest","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-09-11","last_updated":"2026-09-11","modalities":{"input":["pdf","image","text"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}}},"reflection/beam-1.0-501b-a23b":{"id":"reflection/beam-1.0-501b-a23b","name":"Beam 1.0 501B A23B","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":943718},"cost":{"input":0.3,"output":1,"cache_read":0.04},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"elorian/glimpse":{"id":"elorian/glimpse","name":"Glimpse","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-17","last_updated":"2026-10-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":256000},"cost":{"input":0.6,"output":3,"cache_read":0.06}},"~deepseek/deepseek-v4-flash-latest":{"id":"~deepseek/deepseek-v4-flash-latest","name":"DeepSeek V4 Flash Latest","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-01","last_updated":"2026-08-01","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.0422,"output":0.1729,"cache_read":0.0275}},"~deepseek/deepseek-pro-latest":{"id":"~deepseek/deepseek-pro-latest","name":"DeepSeek Pro Latest","description":"Flagship DeepSeek model for coding, reasoning, and agentic work","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-14","last_updated":"2026-09-14","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":393216},"cost":{"input":0.1307,"output":8,"cache_read":0.13}},"~deepseek/deepseek-flash-latest":{"id":"~deepseek/deepseek-flash-latest","name":"DeepSeek Flash Latest","description":"Fast DeepSeek model for efficient chat, coding help, and agent loops","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-14","last_updated":"2026-09-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.07,"output":0.5,"cache_read":0.008}},"~anthropic/claude-fable-latest":{"id":"~anthropic/claude-fable-latest","name":"Claude Fable Latest","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5}},"~anthropic/claude-haiku-latest":{"id":"~anthropic/claude-haiku-latest","name":"Claude Haiku Latest","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]}},"~anthropic/claude-opus-latest":{"id":"~anthropic/claude-opus-latest","name":"Claude Opus Latest","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5}},"~anthropic/claude-sonnet-latest":{"id":"~anthropic/claude-sonnet-latest","name":"Claude Sonnet Latest","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5}},"writer/palmyra-x5":{"id":"writer/palmyra-x5","name":"Palmyra X5","description":"General-purpose chat model for instruction following, writing, and analysis","family":"palmyra","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1040000,"output":8192},"cost":{"input":0.6,"output":6},"canonical_model_id":"writer/palmyra-x5"},"fireworks/ember-1":{"id":"fireworks/ember-1","name":"Ember-1","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"fireworks/ember-1"},"amazon/nova-2-lite-v1":{"id":"amazon/nova-2-lite-v1","name":"Nova 2 Lite","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"nova","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-10","release_date":"2025-12-02","last_updated":"2025-12-02","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65535},"cost":{"input":0.3,"output":2.5},"canonical_model_id":"amazon/nova-2-lite-v1"},"amazon/nova-pro-v1":{"id":"amazon/nova-pro-v1","name":"Nova Pro 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova-pro","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":5120},"cost":{"input":0.8,"output":3.2},"canonical_model_id":"amazon/nova-pro-v1"},"amazon/nova-lite-v1":{"id":"amazon/nova-lite-v1","name":"Nova Lite 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-lite","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":300000,"output":5120},"cost":{"input":0.06,"output":0.24},"canonical_model_id":"amazon/nova-lite-v1"},"amazon/nova-premier-v1":{"id":"amazon/nova-premier-v1","name":"Nova Premier 1.0","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"nova","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10","release_date":"2025-04-30","last_updated":"2025-04-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32000},"status":"deprecated","cost":{"input":2.5,"output":12.5,"cache_read":0.625},"canonical_model_id":"amazon/nova-premier"},"amazon/nova-micro-v1":{"id":"amazon/nova-micro-v1","name":"Nova Micro 1.0","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"nova-micro","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-10-31","release_date":"2024-12-05","last_updated":"2024-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":5120},"cost":{"input":0.035,"output":0.14},"canonical_model_id":"amazon/nova-micro-v1"},"~moonshotai/kimi-latest":{"id":"~moonshotai/kimi-latest","name":"Kimi Latest","description":"Kimi multimodal agent model for visual understanding, coding, and planning","family":"kimi","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":0.238,"output":16,"cache_read":0.228}},"perplexity/sonar-reasoning-pro":{"id":"perplexity/sonar-reasoning-pro","name":"Sonar Reasoning Pro","description":"Web-grounded reasoning model for multi-step research and cited answers","family":"sonar-reasoning","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-07","last_updated":"2025-03-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":115200},"cost":{"input":2,"output":8}},"perplexity/sonar-deep-research":{"id":"perplexity/sonar-deep-research","name":"Sonar Deep Research","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar-deep-research","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-07","last_updated":"2025-03-07","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":115200},"cost":{"input":2,"output":8,"reasoning":3}},"perplexity/sonar-pro-search":{"id":"perplexity/sonar-pro-search","name":"Sonar Pro Search","description":"Advanced Sonar search model for deeper research and cited synthesis","family":"sonar-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2025-10-30","last_updated":"2025-10-30","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8000},"cost":{"input":3,"output":15}},"perplexity/sonar-pro":{"id":"perplexity/sonar-pro","name":"Sonar Pro","description":"Advanced Sonar search model for deeper research and cited synthesis","family":"sonar-pro","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-03-07","last_updated":"2025-03-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":8000},"cost":{"input":3,"output":15}},"perplexity/sonar":{"id":"perplexity/sonar","name":"Sonar","description":"Sonar search model for current answers, retrieval, and citation-backed chat","family":"sonar","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2025-01-27","last_updated":"2025-01-27","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":127072,"output":114364},"cost":{"input":1,"output":1}},"xingchen-agi/xing4.0-29b-a4b:free":{"id":"xingchen-agi/xing4.0-29b-a4b:free","name":"Xing4.0 29B A4B (free)","description":"Free provider route for experiments, demos, and cost-sensitive chat workloads","family":"agi","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-16","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0,"output":0}},"relace/relace-search":{"id":"relace/relace-search","name":"Relace Search","description":"Tool-capable chat model for instruction following and agentic application workflows","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-12-08","last_updated":"2025-12-08","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":1,"output":3},"canonical_model_id":"relace/relace-search"},"relace/relace-apply-3":{"id":"relace/relace-apply-3","name":"Relace Apply 3","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2025-09-26","last_updated":"2025-09-26","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":128000},"cost":{"input":0.85,"output":1.25},"canonical_model_id":"relace/relace-apply-3"},"perceptron/perceptron-mk1":{"id":"perceptron/perceptron-mk1","name":"Perceptron Mk1","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-05-12","last_updated":"2026-05-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":0.15,"output":1.5},"canonical_model_id":"perceptron/perceptron-mk1"},"perceptron/perceptron-mk1.5":{"id":"perceptron/perceptron-mk1.5","name":"Perceptron Mk1.5","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-25","last_updated":"2026-09-25","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":36864,"output":8192},"cost":{"input":0.15,"output":1.5}},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m2.1":{"id":"minimax/minimax-m2.1","name":"MiniMax-M2.1","description":"Earlier MiniMax agent model for practical coding and productivity tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":false,"temperature":true,"release_date":"2025-12-23","last_updated":"2025-12-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.1"},"minimax/minimax-m1":{"id":"minimax/minimax-m1","name":"MiniMax M1","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2024-06-30","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":40000},"cost":{"input":0.55,"output":2.2}},"minimax/minimax-01":{"id":"minimax/minimax-01","name":"MiniMax-01","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2024-03-31","release_date":"2025-01-15","last_updated":"2025-01-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000192,"output":40000},"cost":{"input":0.2,"output":1.1}},"minimax/minimax-m2-her":{"id":"minimax/minimax-m2-her","name":"MiniMax-M2 Her","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-01-23","last_updated":"2026-01-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":2048},"cost":{"input":0.3,"output":1.2,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2-Her"},"minimax/minimax-m2":{"id":"minimax/minimax-m2","name":"MiniMax-M2","description":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2025-10-27","last_updated":"2025-10-27","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.255,"output":1.02},"canonical_model_id":"minimax/MiniMax-M2"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m3.1-flash:free":{"id":"minimax/minimax-m3.1-flash:free","name":"MiniMax M3.1 Flash (free)","description":"MiniMax multimodal coding model for long-context reasoning and agent tasks","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-19","last_updated":"2026-10-19","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":524288},"cost":{"input":0,"output":0},"canonical_model_id":"minimax/MiniMax-M3.1-Flash"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":32768},"cost":{"input":0.27,"output":0.95,"cache_read":0.03},"canonical_model_id":"minimax/MiniMax-M2.5"},"upstage/solar-mini4":{"id":"upstage/solar-mini4","name":"Solar Mini 4","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"solar-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.05,"output":0.2,"cache_read":0.005},"canonical_model_id":"upstage/solar-mini4"},"upstage/solar-pro-3":{"id":"upstage/solar-pro-3","name":"Solar Pro 3","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07","release_date":"2026-01-26","last_updated":"2026-03-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":117964},"cost":{"input":0.15,"output":0.6,"cache_read":0.015},"canonical_model_id":"upstage/solar-pro3"},"upstage/solar-pro4":{"id":"upstage/solar-pro4","name":"Solar Pro 4","description":"Flagship model for demanding analysis, coding, and production agent workflows","family":"solar-pro","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02","release_date":"2026-08-06","last_updated":"2026-08-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":524288,"output":131072},"cost":{"input":0.15,"output":0.6,"cache_read":0.03},"canonical_model_id":"upstage/solar-pro4"},"inference-net/schematron-v2-small":{"id":"inference-net/schematron-v2-small","name":"Schematron V2 Small","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":0.05,"output":0.23,"cache_read":0.05},"canonical_model_id":"inference-net/schematron-v2-small"},"inference-net/schematron-v2-turbo":{"id":"inference-net/schematron-v2-turbo","name":"Schematron V2 Turbo","description":"Efficient model for low-latency assistance, extraction, and routine automation","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":8192},"cost":{"input":0.03,"output":0.15,"cache_read":0.03},"canonical_model_id":"inference-net/schematron-v2-turbo"},"scaledown/summarize":{"id":"scaledown/summarize","name":"Summarize","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":4000},"cost":{"input":0.04,"output":0}},"scaledown/extract":{"id":"scaledown/extract","name":"Extract","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":900000},"cost":{"input":0.04,"output":0}},"microsoft/phi-4":{"id":"microsoft/phi-4","name":"Phi 4","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","family":"phi","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2025-01-10","last_updated":"2025-01-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":14745},"cost":{"input":0.07,"output":0.14}},"nvidia/nemotron-3.5-lightning":{"id":"nvidia/nemotron-3.5-lightning","name":"Nemotron 3.5 Lightning 30B A3B","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.07,"output":0.2,"cache_read":0.04},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-3-super-120b-a12b":{"id":"nvidia/nemotron-3-super-120b-a12b","name":"Nemotron 3 Super 120B A12B","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.3,"output":2.4,"cache_read":0.15},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/nemotron-3.5-content-safety:free":{"id":"nvidia/nemotron-3.5-content-safety:free","name":"Nemotron 3.5 Content Safety (free)","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":8192},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"nvidia/nemotron-3-ultra-550b-a55b":{"id":"nvidia/nemotron-3-ultra-550b-a55b","name":"Nemotron 3 Ultra 550B A55B","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":182520},"cost":{"input":0.6,"output":2.4,"cache_read":0.12},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/nemotron-3.5-lightning:free":{"id":"nvidia/nemotron-3.5-lightning:free","name":"Nemotron 3.5 Lightning (free)","description":"Nemotron model for efficient reasoning, coding, and specialized AI agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-08-11","last_updated":"2026-08-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-lightning"},"nvidia/nemotron-3-nano-30b-a3b":{"id":"nvidia/nemotron-3-nano-30b-a3b","name":"Nemotron 3 Nano 30B A3B","description":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06-25","release_date":"2025-12-15","last_updated":"2025-12-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.06,"output":0.24},"canonical_model_id":"nvidia/nemotron-3-nano-30b-a3b"},"nvidia/nemotron-3.5-super-vl":{"id":"nvidia/nemotron-3.5-super-vl","name":"Nemotron 3.5 Super VL","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":235929},"cost":{"input":0.15,"output":0.5,"cache_read":0.02},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/nemotron-3.5-super-vl:free":{"id":"nvidia/nemotron-3.5-super-vl:free","name":"Nemotron 3.5 Super VL (free)","description":"Nemotron multimodal model for visual reasoning and agentic AI workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3.5-super-vl-120b-a12b"},"nvidia/nemotron-3-super-120b-a12b:free":{"id":"nvidia/nemotron-3-super-120b-a12b:free","name":"Nemotron 3 Super (free)","description":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-06","release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-super-120b-a12b"},"nvidia/nemotron-3-ultra-550b-a55b:free":{"id":"nvidia/nemotron-3-ultra-550b-a55b:free","name":"Nemotron 3 Ultra (free)","description":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","family":"nemotron","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]},{"type":"budget_tokens"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09","release_date":"2026-06-04","last_updated":"2026-06-04","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-ultra-550b-a55b"},"nvidia/nemotron-3.5-content-safety":{"id":"nvidia/nemotron-3.5-content-safety","name":"Nemotron 3.5 Content Safety","description":"Safety model for policy screening, moderation, and risk-aware routing workflows","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.2,"output":0.2},"canonical_model_id":"nvidia/nemotron-3.5-content-safety"},"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free":{"id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free","name":"Nemotron 3 Nano Omni (free)","description":"Open Nemotron omni model combining reasoning with text, vision, and audio","family":"nemotron","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-28","last_updated":"2026-04-28","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning"},"mancer/weaver":{"id":"mancer/weaver","name":"Weaver (alpha)","description":"General-purpose chat model for instruction following, writing, and analysis","family":"alpha","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2023-06-30","release_date":"2023-08-02","last_updated":"2023-08-02","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8000,"output":6000},"cost":{"input":0.4,"output":0.75},"canonical_model_id":"mancer/weaver"},"bytedance/ui-tars-1.5-7b":{"id":"bytedance/ui-tars-1.5-7b","name":"UI-TARS 7B ","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01-31","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":2048},"cost":{"input":0.1,"output":0.2,"cache_read":0.1}},"sao10k/l3.1-euryale-70b":{"id":"sao10k/l3.1-euryale-70b","name":"Llama 3.1 Euryale 70B v2.2","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-28","last_updated":"2024-08-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.85,"output":0.85},"canonical_model_id":"sao10k/l3.1-euryale-70b"},"sao10k/l3-lunaris-8b":{"id":"sao10k/l3-lunaris-8b","name":"Llama 3 8B Lunaris","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2023-12-31","release_date":"2024-08-13","last_updated":"2024-08-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":7372},"cost":{"input":0.04,"output":0.05},"canonical_model_id":"sao10k/l3-lunaris-8b"},"meituan/longcat-2.0":{"id":"meituan/longcat-2.0","name":"LongCat-2.0","description":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","family":"longcat","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048756,"output":262144},"cost":{"input":0.3,"output":1.2,"cache_read":0.006},"canonical_model_id":"meituan/longcat-2.0"},"nex-agi/nex-n2.5-pro":{"id":"nex-agi/nex-n2.5-pro","name":"Nex-N2.5-Pro","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"agi","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.075,"output":0.25,"cache_read":0.015}},"nex-agi/nex-n2.5-mini":{"id":"nex-agi/nex-n2.5-mini","name":"Nex-N2.5-Mini","description":"Multimodal reasoning model for visual analysis, planning, and tool use","family":"agi","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"release_date":"2026-09-08","last_updated":"2026-09-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.025,"output":0.1,"cache_read":0.0025}},"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":81920},"cost":{"input":0.14,"output":1,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3-vl-30b-a3b-instruct":{"id":"qwen/qwen3-vl-30b-a3b-instruct","name":"Qwen3 VL 30B A3B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-04","last_updated":"2025-10-04","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"alibaba/qwen3-vl-30b-a3b-instruct"},"qwen/qwen3-30b-a3b":{"id":"qwen/qwen3-30b-a3b","name":"Qwen3 30B A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.5},"canonical_model_id":"alibaba/qwen3-30b-a3b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.475,"output":4.425,"cache_read":0.295,"cache_write":1.84375},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.32,"output":1.28,"cache_read":0.064,"cache_write":0.4,"tiers":[{"input":0.96,"output":3.84,"cache_read":0.192,"cache_write":1.2,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.96,"output":3.84,"cache_read":0.192,"cache_write":1.2}},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3-14b":{"id":"qwen/qwen3-14b","name":"Qwen3 14B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":40960,"output":16384},"cost":{"input":0.12,"output":0.24}},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":65536},"cost":{"input":0.04,"output":1.35,"cache_read":0.018},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.8-max-prime":{"id":"qwen/qwen3.8-max-prime","name":"Qwen 3.8 Max Prime","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.8-max","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-23","last_updated":"2026-09-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":4,"output":12,"cache_read":0.5},"canonical_model_id":"alibaba/qwen3.8-max-prime"},"qwen/qwen3.6-35b-a3b":{"id":"qwen/qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.1,"output":0.95,"cache_read":0.05},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.26,"output":2.08},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3.8-2.4t-a95b":{"id":"qwen/qwen3.8-2.4t-a95b","name":"Qwen3.8 2.4T A95B","description":"Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":909000},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-2.4t-a95b"},"qwen/qwen3-coder-next":{"id":"qwen/qwen3-coder-next","name":"Qwen3 Coder Next","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09","release_date":"2026-02-03","last_updated":"2026-02-03","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.12,"output":0.8,"cache_read":0.07},"canonical_model_id":"alibaba/qwen3-coder-next"},"qwen/qwen3-30b-a3b-instruct-2507":{"id":"qwen/qwen3-30b-a3b-instruct-2507","name":"Qwen3 30B-A3B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.09,"output":0.3},"canonical_model_id":"alibaba/qwen3-30b-a3b-instruct-2507"},"qwen/qwen-2.5-7b-instruct":{"id":"qwen/qwen-2.5-7b-instruct","name":"Qwen2.5 7B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-10-16","last_updated":"2024-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":29491},"cost":{"input":0.1,"output":0.2}},"qwen/qwen3.8-max-0902":{"id":"qwen/qwen3.8-max-0902","name":"Qwen3.8 Max 0902","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max-0902"},"qwen/qwen-2.5-72b-instruct":{"id":"qwen/qwen-2.5-72b-instruct","name":"Qwen2.5 72B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-09-19","last_updated":"2024-09-19","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":16384},"cost":{"input":0.36,"output":0.4}},"qwen/qwen3.5-flash-02-23":{"id":"qwen/qwen3.5-flash-02-23","name":"Qwen3.5-Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-25","last_updated":"2026-02-25","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.065,"output":0.26}},"qwen/qwen3-vl-8b-instruct":{"id":"qwen/qwen3-vl-8b-instruct","name":"Qwen3 VL 8B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-10-14","last_updated":"2025-10-14","modalities":{"input":["image","text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.25,"output":0.75,"cache_read":0.12}},"qwen/qwen3.5-plus-02-15":{"id":"qwen/qwen3.5-plus-02-15","name":"Qwen3.5 Plus 2026-02-15","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.26,"output":1.56,"tiers":[{"input":0.325,"output":1.95,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.325,"output":1.95}}},"qwen/qwen2.5-vl-72b-instruct":{"id":"qwen/qwen2.5-vl-72b-instruct","name":"Qwen2.5-VL 72B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-01-26","last_updated":"2025-01-26","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":115200},"cost":{"input":0.8,"output":1,"cache_read":0.4},"canonical_model_id":"alibaba/qwen2-5-vl-72b-instruct"},"qwen/qwen3.7-flash":{"id":"qwen/qwen3.7-flash","name":"Qwen3.7 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":65536},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.038,"tiers":[{"input":0.1,"output":0.4,"cache_read":0.02,"cache_write":0.125,"tier":{"type":"context","size":32000}},{"input":0.2,"output":0.8,"cache_read":0.04,"cache_write":0.25,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.27,"output":2.16,"cache_read":0.27},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1875,"output":1.125,"cache_write":0.234375,"tiers":[{"input":0.75,"output":3,"cache_write":0.9375,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.75,"output":3,"cache_write":0.9375}},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.39,"output":2.34},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3-coder-30b-a3b-instruct":{"id":"qwen/qwen3-coder-30b-a3b-instruct","name":"Qwen3-Coder 30B-A3B Instruct","description":"Smaller Qwen coder for efficient local agents and repo-level fixes","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-31","last_updated":"2025-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.07,"output":0.28},"canonical_model_id":"alibaba/qwen3-coder-30b-a3b-instruct"},"qwen/qwen3-next-80b-a3b-thinking":{"id":"qwen/qwen3-next-80b-a3b-thinking","name":"Qwen3-Next 80B-A3B (Thinking)","description":"Efficient Qwen thinking model for local reasoning, math, and coding agents","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.15,"output":1.2},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-thinking"},"qwen/qwen3.8-omni-flash":{"id":"qwen/qwen3.8-omni-flash","name":"Qwen3.8 Omni Flash","description":"Qwen omni model for text, vision, audio, and multimodal agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image","audio","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-omni-flash"},"qwen/qwen3.8-flash":{"id":"qwen/qwen3.8-flash","name":"Qwen3.8 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":0.15,"output":0.47,"cache_read":0.016},"canonical_model_id":"alibaba/qwen3.8-flash"},"qwen/qwen3-next-80b-a3b-instruct":{"id":"qwen/qwen3-next-80b-a3b-instruct","name":"Qwen3-Next 80B-A3B Instruct","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09","last_updated":"2025-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.09,"output":1.1},"canonical_model_id":"alibaba/qwen3-next-80b-a3b-instruct"},"qwen/qwen3-vl-30b-a3b-thinking":{"id":"qwen/qwen3-vl-30b-a3b-thinking","name":"Qwen3 VL 30B A3B Thinking","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.29,"output":1}},"qwen/qwen3-235b-a22b-2507":{"id":"qwen/qwen3-235b-a22b-2507","name":"Qwen3 235B A22B Instruct 2507","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-07-21","last_updated":"2025-07-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.09,"output":0.55},"canonical_model_id":"alibaba/qwen3-235b-a22b-instruct-2507"},"qwen/qwen3-32b":{"id":"qwen/qwen3-32b","name":"Qwen3 32B","description":"Dense open Qwen model for self-hosted chat, reasoning, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-04-28","last_updated":"2025-04-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.08,"output":0.28},"canonical_model_id":"alibaba/qwen3-32b"},"qwen/qwen3-235b-a22b-thinking-2507":{"id":"qwen/qwen3-235b-a22b-thinking-2507","name":"Qwen3 235B-A22B Thinking 2507","description":"Qwen reasoning model for deliberate problem solving, math, and coding","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2025-07-25","last_updated":"2025-07-25","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":16384},"cost":{"input":0.45,"output":3.5},"canonical_model_id":"alibaba/qwen3-235b-a22b-thinking-2507"},"qwen/qwen3.5-9b":{"id":"qwen/qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":0.08,"output":0.13,"cache_read":0.04},"canonical_model_id":"alibaba/qwen3.5-9b"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.325,"output":1.95,"cache_write":0.40625,"tiers":[{"input":1.3,"output":3.9,"cache_write":1.625,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.3,"output":3.9,"cache_write":1.625}},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen/qwen3-vl-235b-a22b-instruct":{"id":"qwen/qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.21,"output":1.9,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"qwen/qwen-2.5-coder-32b-instruct":{"id":"qwen/qwen-2.5-coder-32b-instruct","name":"Qwen2.5-Coder-32B-Instruct","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2024-11-12","last_updated":"2024-11-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32768,"output":29491},"cost":{"input":0.66,"output":1},"canonical_model_id":"alibaba/qwen2.5-coder-32b-instruct"},"qwen/qwen-plus":{"id":"qwen/qwen-plus","name":"Qwen Plus","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2024-01-25","last_updated":"2025-09-11","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"cost":{"input":0.26,"output":0.78,"cache_read":0.052,"cache_write":0.325,"tiers":[{"input":0.78,"output":2.34,"cache_read":0.156,"cache_write":0.975,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.78,"output":2.34,"cache_read":0.156,"cache_write":0.975}},"canonical_model_id":"alibaba/qwen-plus"},"qwen/qwen3.5-plus-20260420":{"id":"qwen/qwen3.5-plus-20260420","name":"Qwen3.5 Plus 2026-04-20","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.5","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":1.8,"cache_write":0.375,"tiers":[{"input":0.375,"output":2.25,"cache_write":0.46875,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.375,"output":2.25,"cache_write":0.46875}}},"qwen/qwen3-coder-flash":{"id":"qwen/qwen3-coder-flash","name":"Qwen3 Coder Flash","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.195,"output":0.975,"cache_read":0.039,"cache_write":0.24375,"tiers":[{"input":0.325,"output":1.625,"cache_read":0.065,"cache_write":0.40625,"tier":{"type":"context","size":32000}},{"input":0.52,"output":2.6,"cache_read":0.104,"cache_write":0.65,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3-coder-flash"},"qwen/qwen3-coder":{"id":"qwen/qwen3-coder","name":"Qwen3 Coder 480B A35B","description":"Qwen coding model for software agents, repository edits, and code reasoning","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-22","last_updated":"2025-07-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.3,"output":1,"cache_read":0.1},"canonical_model_id":"alibaba/qwen3-coder-480b-a35b-instruct"},"qwen/qwen3.6-27b":{"id":"qwen/qwen3.6-27b","name":"Qwen3.6 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-22","last_updated":"2026-04-22","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"alibaba/qwen3.6-27b"},"cohere/command-r7b-12-2024":{"id":"cohere/command-r7b-12-2024","name":"Command R7B","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-12-13","last_updated":"2024-12-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.0375,"output":0.15},"canonical_model_id":"cohere/command-r7b-12-2024"},"cohere/command-r-plus-08-2024":{"id":"cohere/command-r-plus-08-2024","name":"Command R+","description":"Cohere's RAG workhorse for long-context enterprise search and tool use","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-r-plus-08-2024"},"cohere/north-mini-code:free":{"id":"cohere/north-mini-code:free","name":"North Mini Code (free)","description":"Cohere coding model for practical software engineering and agentic edits","family":"north","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-09-23","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":64000},"cost":{"input":0,"output":0},"canonical_model_id":"cohere/north-mini-code-1-0"},"cohere/command-a-plus":{"id":"cohere/command-a-plus","name":"Command A+","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04-01","release_date":"2026-05-20","last_updated":"2026-06-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":192000,"output":64000},"cost":{"input":0.3,"output":1.5,"cache_read":0.15},"canonical_model_id":"cohere/command-a-plus-05-2026"},"cohere/command-a":{"id":"cohere/command-a","name":"Command A","description":"Cohere command model for multilingual enterprise agents, tools, and chat","family":"command-a","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08-31","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"output":8192},"cost":{"input":2.5,"output":10},"canonical_model_id":"cohere/command-a"},"cohere/command-r-08-2024":{"id":"cohere/command-r-08-2024","name":"Command R","description":"Cohere retrieval model for long-context chat and enterprise RAG workflows","family":"command-r","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06-01","release_date":"2024-08-30","last_updated":"2024-08-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":128000,"output":4000},"cost":{"input":0.15,"output":0.6},"canonical_model_id":"cohere/command-r-08-2024"},"recursive/flint-v1.0:free":{"id":"recursive/flint-v1.0:free","name":"Flint (free)","description":"Free provider route for experiments, demos, and cost-sensitive chat workloads","family":"flint","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-21","last_updated":"2026-10-21","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0,"output":0}},"unbiased/pareto-26.10":{"id":"unbiased/pareto-26.10","name":"Pareto 26.10","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":0.8,"output":3.2,"cache_read":0.03}},"unbiased/pareto-26.9":{"id":"unbiased/pareto-26.9","name":"Pareto 26.9","description":"Multimodal model for analyzing text, images, documents, and rich media","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.25}},"prism-ml/ternary-bonsai-2-27b":{"id":"prism-ml/ternary-bonsai-2-27b","name":"Ternary Bonsai 2 27B","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-18","last_updated":"2026-09-18","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.075,"output":0.5,"cache_read":0.0375}},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.825,"output":4.125,"reasoning":4.125,"cache_read":0.0825,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"status":"deprecated","cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Gemini multimodal model for text, image, audio, video, and document tasks","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":943718},"cost":{"input":2,"output":10,"reasoning":10,"cache_read":0.1,"cache_write":0.1875},"canonical_model_id":"google/gemini-4-argon"},"google/gemma-4-26b-a4b-it:free":{"id":"google/gemma-4-26b-a4b-it:free","name":"Gemma 4 26B A4B (free)","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemini-3.1-flash-lite-preview":{"id":"google/gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"reasoning":1.5,"cache_read":0.025,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"google/gemma-3-27b-it":{"id":"google/gemma-3-27b-it","name":"Gemma 3 27B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.08,"output":0.35,"cache_read":0.04},"canonical_model_id":"google/gemma-3-27b-it"},"google/gemini-3.1-flash-lite-image":{"id":"google/gemini-3.1-flash-lite-image","type":"image","name":"Nano Banana 2 Lite","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":58982},"cost":{"input":0.25,"output":1.5},"canonical_model_id":"google/gemini-3.1-flash-lite-image"},"google/gemini-3-pro-image-preview":{"id":"google/gemini-3-pro-image-preview","type":"image","name":"Nano Banana Pro Preview","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-11-20","last_updated":"2025-11-20","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3-pro-image-preview"},"google/gemma-3-4b-it":{"id":"google/gemma-3-4b-it","name":"Gemma 3 4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":117964},"cost":{"input":0.04,"output":0.05,"cache_read":0.02},"canonical_model_id":"google/gemma-3-4b-it"},"google/gemini-2.5-flash-image":{"id":"google/gemini-2.5-flash-image","type":"image","name":"Nano Banana","description":"Nano Banana image model for fast generation, edits, and character-consistent assets","family":"gemini-flash","attachment":true,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-26","last_updated":"2025-08-26","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":32768,"output":8192},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.083333},"canonical_model_id":"google/gemini-2.5-flash-image"},"google/gemma-4-31b-it:free":{"id":"google/gemma-4-31b-it:free","name":"Gemma 4 31B (free)","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0,"output":0},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-2.5-pro-preview":{"id":"google/gemini-2.5-pro-preview","name":"Gemini 2.5 Pro Preview 06-05","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01-31","release_date":"2025-06-05","last_updated":"2025-06-05","modalities":{"input":["pdf","image","text","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.25,"output":10,"reasoning":10,"cache_read":0.125,"cache_write":0.375,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro-preview"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"reasoning":2.5,"cache_read":0.03,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-3.1-flash-image-preview":{"id":"google/gemini-3.1-flash-image-preview","type":"image","name":"Nano Banana 2 Preview","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-26","last_updated":"2026-02-26","modalities":{"input":["image","text"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":58982},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image-preview"},"google/gemini-nano-banana-2.1":{"id":"google/gemini-nano-banana-2.1","type":"image","name":"Nano Banana 2.1","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-10-06","last_updated":"2026-10-06","modalities":{"input":["image","text"],"output":["text","image"]},"open_weights":false,"limit":{"context":65536,"output":58982},"cost":{"input":1.5,"output":7.5},"canonical_model_id":"google/gemini-nano-banana-2.1"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":235929},"cost":{"input":0.0765,"output":0.255,"cache_read":0.0425},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"reasoning":3.75,"cache_read":0.075,"cache_write":0.041667},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemma-3-12b-it":{"id":"google/gemma-3-12b-it","name":"Gemma 3 12B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-08","release_date":"2025-03-12","last_updated":"2025-03-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":16384},"cost":{"input":0.05,"output":0.15},"canonical_model_id":"google/gemma-3-12b-it"},"google/gemma-2-27b-it":{"id":"google/gemma-2-27b-it","name":"Gemma 2 27B","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":false,"reasoning":false,"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2024-06-30","release_date":"2024-07-13","last_updated":"2024-07-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":8192,"output":2048},"cost":{"input":0.65,"output":0.65},"canonical_model_id":"google/gemma-2-27b-it"},"google/lyria-3-clip-preview":{"id":"google/lyria-3-clip-preview","name":"Lyria 3 Clip Preview","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"lyria","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-25","last_updated":"2026-03-25","modalities":{"input":["text","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/lyria-3-clip-preview"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"reasoning":9,"cache_read":0.15,"cache_write":0.083333},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-3.1-flash-image":{"id":"google/gemini-3.1-flash-image","type":"image","name":"Nano Banana 2","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","high"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["image","text"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":0.5,"output":3},"canonical_model_id":"google/gemini-3.1-flash-image"},"google/lyria-3-pro-preview":{"id":"google/lyria-3-pro-preview","name":"Lyria 3 Pro Preview","description":"Speech generation model for controllable voice, narration, and audio delivery","family":"lyria","attachment":true,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"release_date":"2026-03-25","last_updated":"2026-03-25","modalities":{"input":["text","image"],"output":["text","audio"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0,"output":0},"canonical_model_id":"google/lyria-3-pro-preview"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","minimal","low","medium","high"]}],"tool_call":true,"interleaved":{"field":"reasoning_details"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"reasoning":3,"cache_read":0.05,"cache_write":0.083333},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":16384},"cost":{"input":0.2,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65535},"status":"deprecated","cost":{"input":0.1,"output":0.4,"reasoning":0.4,"cache_read":0.01,"cache_write":0.083333},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-3-pro-image":{"id":"google/gemini-3-pro-image","type":"image","name":"Nano Banana Pro","description":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image"],"output":["text","image"]},"open_weights":false,"limit":{"context":131072,"output":32768},"cost":{"input":2,"output":12,"reasoning":12,"cache_read":0.2,"cache_write":0.375},"canonical_model_id":"google/gemini-3-pro-image"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":128,"max":32768}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"status":"deprecated","cost":{"input":1.25,"output":10,"reasoning":10,"cache_read":0.125,"cache_write":0.375,"tiers":[{"input":2.5,"output":15,"cache_read":0.25,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":15,"cache_read":0.25}},"canonical_model_id":"google/gemini-2.5-pro"},"dots-studio/dots-3-note-preview:free":{"id":"dots-studio/dots-3-note-preview:free","name":"Dots3-Note Preview (free)","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":512000,"output":460800},"status":"deprecated","cost":{"input":0,"output":0}},"x-ai/grok-4.5":{"id":"x-ai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"x-ai/grok-4.6":{"id":"x-ai/grok-4.6","name":"Grok 4.6","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"x-ai/grok-4.7":{"id":"x-ai/grok-4.7","name":"Grok 4.7","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":450000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"x-ai/grok-4.20-multi-agent":{"id":"x-ai/grok-4.20-multi-agent","name":"Grok 4.20 Multi-Agent","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-09-01","release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":1800000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"x-ai/grok-4.20":{"id":"x-ai/grok-4.20","name":"Grok 4.20","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-09-01","release_date":"2026-03-31","last_updated":"2026-03-31","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":1800000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}}},"x-ai/grok-4.3":{"id":"x-ai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":900000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"x-ai/grok-build-0.1":{"id":"x-ai/grok-build-0.1","name":"Grok Build 0.1","description":"Fast Grok coding model tuned for agentic engineering and iterative edits","family":"grok-build","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":256000,"output":230400},"cost":{"input":1,"output":2,"cache_read":0.2,"tiers":[{"input":2,"output":4,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2,"output":4,"cache_read":0.4}},"canonical_model_id":"xai/grok-build-0.1"},"openrouter/free":{"id":"openrouter/free","name":"Free Models Router","description":"Multimodal reasoning model for visual analysis, planning, and tool use","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-01","last_updated":"2026-02-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"input":200000,"output":8000},"cost":{"input":0,"output":0}},"openrouter/bodybuilder":{"id":"openrouter/bodybuilder","name":"Body Builder (beta)","description":"Preview model for early access evaluation, prototyping, and compatibility testing","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2025-12-05","last_updated":"2025-12-05","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":128000}},"openrouter/fusion":{"id":"openrouter/fusion","name":"Fusion","description":"General-purpose chat model for instruction following, writing, and analysis","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000}},"openrouter/pareto-code":{"id":"openrouter/pareto-code","name":"Pareto Code Router","description":"Coding model for repository understanding, refactors, and agentic engineering tasks","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":false,"release_date":"2026-04-21","last_updated":"2026-04-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":200000}},"openrouter/auto":{"id":"openrouter/auto","name":"Auto Router","description":"Image model for prompt-driven generation, editing, and visual design workflows","family":"auto","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2023-11-08","last_updated":"2023-11-08","modalities":{"input":["text","image","audio","pdf","video"],"output":["text","image"]},"open_weights":false,"limit":{"context":2000000,"output":2000000}},"arcee-ai/trinity-large-thinking":{"id":"arcee-ai/trinity-large-thinking","name":"Trinity Large Thinking","description":"Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use","family":"trinity","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-01","last_updated":"2026-05-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":80000},"cost":{"input":0.25,"output":0.8,"cache_read":0.06},"canonical_model_id":"arcee-ai/trinity-large-thinking"},"rekaai/reka-edge":{"id":"rekaai/reka-edge","name":"Reka Edge","description":"Multimodal model for analyzing text, images, documents, and rich media","family":"reka","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-11","last_updated":"2026-03-11","modalities":{"input":["image","text","video"],"output":["text"]},"open_weights":true,"limit":{"context":16384,"output":14745},"cost":{"input":0.1,"output":0.1},"canonical_model_id":"rekaai/reka-edge"},"rekaai/reka-flash-3":{"id":"rekaai/reka-flash-3","name":"Reka Flash 3","description":"Efficient model for low-latency assistance, extraction, and routine automation","family":"reka","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":true,"knowledge":"2025-01-31","release_date":"2025-03-10","last_updated":"2025-03-10","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":65536,"output":58982},"cost":{"input":0.1,"output":0.2},"canonical_model_id":"rekaai/reka-flash-3"}}},"wafer.ai":{"id":"wafer.ai","env":["WAFER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://pass.wafer.ai/v1","name":"Wafer","doc":"https://docs.wafer.ai/wafer-pass","models":{"GLM-5.2":{"id":"GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"},"MiniMax-M3":{"id":"MiniMax-M3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"status":"deprecated","cost":{"input":0.33,"output":1.32,"cache_read":0.07,"cache_write":0,"tiers":[{"input":0.66,"output":2.64,"cache_read":0.13,"cache_write":0,"tier":{"type":"context","size":512000}}],"context_over_200k":{"input":0.66,"output":2.64,"cache_read":0.13,"cache_write":0}},"canonical_model_id":"minimax/MiniMax-M3"},"Kimi-K2.6":{"id":"Kimi-K2.6","name":"Kimi K2.6","description":"Kimi K2.6 sparse MoE model with a 262K context window. Available serverless and not included in standard Wafer Pass. Non-ZDR only: requests with `Wafer-ZDR: required` are rejected.","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"status":"deprecated","cost":{"input":1.14,"output":4.8,"cache_read":0.19,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.6"},"GLM-5.1":{"id":"GLM-5.1","name":"GLM-5.1","description":"General Language Model 5.1 — high-quality bilingual (EN/ZH) generation with strong coding and reasoning capabilities.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"status":"deprecated","cost":{"input":1,"output":3.2,"cache_read":0.1,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"glm5.2-fast":{"id":"glm5.2-fast","name":"GLM5.2-Fast","description":"The same model served for high TPS.","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["none","low","medium","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"status":"deprecated","cost":{"input":3,"output":10.25,"cache_read":0.5,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.2"}}},"orcarouter":{"id":"orcarouter","env":["ORCAROUTER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.orcarouter.ai/v1","name":"OrcaRouter","doc":"https://docs.orcarouter.ai","models":{"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-1.2"},"z-ai/glm-5.3-flash-free":{"id":"z-ai/glm-5.3-flash-free","name":"GLM-5.3-Flash (free)","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-4.5-air":{"id":"z-ai/glm-4.5-air","name":"GLM-4.5-Air","description":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","family":"glm-air","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.2,"output":1.1,"cache_read":0.03,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5-air"},"z-ai/glm-5.1":{"id":"z-ai/glm-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":200000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5.1"},"z-ai/glm-4.7":{"id":"z-ai/glm-4.7","name":"GLM-4.7","description":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7"},"z-ai/glm-4.5":{"id":"z-ai/glm-4.5","name":"GLM-4.5","description":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-07-28","last_updated":"2025-07-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":98304},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.5"},"z-ai/glm-5.3-flash":{"id":"z-ai/glm-5.3-flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":128000},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"z-ai/glm-5.2":{"id":"z-ai/glm-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"z-ai/glm-5":{"id":"z-ai/glm-5","name":"GLM-5","description":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":1,"output":3.2,"cache_read":0.26,"cache_write":0},"canonical_model_id":"zhipuai/glm-5"},"z-ai/glm-4.6":{"id":"z-ai/glm-4.6","name":"GLM-4.6","description":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-30","last_updated":"2025-09-30","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.2,"cache_read":0.11,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.6"},"z-ai/glm-5.3":{"id":"z-ai/glm-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":131072},"cost":{"input":1.26,"output":3.96,"cache_read":0.234},"canonical_model_id":"zhipuai/glm-5.3"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-4o-2024-05-13":{"id":"openai/gpt-4o-2024-05-13","name":"GPT-4o (2024-05-13)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-05-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":5,"output":15},"canonical_model_id":"openai/gpt-4o-2024-05-13"},"openai/gpt-5-pro":{"id":"openai/gpt-5-pro","name":"GPT-5 Pro","description":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-10-06","last_updated":"2025-10-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":272000},"cost":{"input":15,"output":120},"canonical_model_id":"openai/gpt-5-pro"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":100000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-5-chat-latest":{"id":"openai/gpt-5-chat-latest","name":"GPT-5 Chat (latest)","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":false,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":100000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5-chat-latest"},"openai/gpt-4":{"id":"openai/gpt-4","name":"GPT-4","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":false,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-11","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":8192,"output":8192},"cost":{"input":30,"output":60},"canonical_model_id":"openai/gpt-4"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.3-codex":{"id":"openai/gpt-5.3-codex","name":"GPT-5.3 Codex","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-05","last_updated":"2026-02-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.3-codex"},"openai/gpt-5.2-codex":{"id":"openai/gpt-5.2-codex","name":"GPT-5.2 Codex","description":"Code-specialist GPT for repository edits, reviews, and long-running software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-codex"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":30,"output":180,"tiers":[{"input":60,"output":270,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":60,"output":270}},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.2":{"id":"openai/gpt-5.2","name":"GPT-5.2","description":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":12.5,"output":75,"cache_read":1.25},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":5,"output":30,"cache_read":0.5,"tiers":[{"input":10,"output":45,"cache_read":1,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":10,"output":45,"cache_read":1}},"canonical_model_id":"openai/gpt-5.5"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":5,"output":30,"cache_read":0.5},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":2.5,"output":15,"cache_read":0.25,"tiers":[{"input":5,"output":22.5,"cache_read":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":5,"output":22.5,"cache_read":0.5}},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"status":"deprecated","cost":{"input":0.03,"output":0.17},"canonical_model_id":"openai/gpt-oss-120b"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04}},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0.15,"output":0.6,"cache_read":0.075},"canonical_model_id":"openai/gpt-4o-mini"},"openai/gpt-4o-2024-11-20":{"id":"openai/gpt-4o-2024-11-20","name":"GPT-4o (2024-11-20)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-11-20","last_updated":"2024-11-20","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-11-20"},"openai/gpt-4o-2024-08-06":{"id":"openai/gpt-4o-2024-08-06","name":"GPT-4o (2024-08-06)","description":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-08-06","last_updated":"2024-08-06","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":2.5,"output":10,"cache_read":1.25},"canonical_model_id":"openai/gpt-4o-2024-08-06"},"openai/gpt-4-turbo":{"id":"openai/gpt-4-turbo","name":"GPT-4 Turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"knowledge":"2023-12","release_date":"2023-11-06","last_updated":"2024-04-09","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":4096},"cost":{"input":10,"output":30},"canonical_model_id":"openai/gpt-4-turbo"},"openai/gpt-3.5-turbo":{"id":"openai/gpt-3.5-turbo","name":"GPT-3.5-turbo","description":"Compact GPT model for low-latency assistance and high-volume workloads","family":"gpt","attachment":false,"reasoning":false,"tool_call":false,"structured_output":false,"temperature":true,"knowledge":"2021-09-01","release_date":"2023-03-01","last_updated":"2023-11-06","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":16385,"output":4096},"cost":{"input":0.5,"output":1.5,"cache_read":0},"canonical_model_id":"openai/gpt-3.5-turbo"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5.1-codex-mini":{"id":"openai/gpt-5.1-codex-mini","name":"GPT-5.1 Codex mini","description":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5.1-codex-mini"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.1-codex":{"id":"openai/gpt-5.1-codex","name":"GPT-5.1 Codex","description":"Codex GPT for repository edits, code review, and practical software agents","family":"gpt-codex","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-codex"},"openai/gpt-5.2-pro":{"id":"openai/gpt-5.2-pro","name":"GPT-5.2 Pro","description":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":21,"output":168},"canonical_model_id":"openai/gpt-5.2-pro"},"openai/gpt-5.2-chat-latest":{"id":"openai/gpt-5.2-chat-latest","name":"GPT-5.2 Chat","description":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-08-31","release_date":"2025-12-11","last_updated":"2025-12-11","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.75,"output":14,"cache_read":0.175},"canonical_model_id":"openai/gpt-5.2-chat-latest"},"openai/gpt-5.1-chat-latest":{"id":"openai/gpt-5.1-chat-latest","name":"GPT-5.1 Chat","description":"Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1-chat-latest"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":1.5,"output":9,"cache_read":0.15},"provider":{"body":{"service_tier":"priority"}}}}},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":400000,"input":272000,"output":128000},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-reasoner":{"id":"deepseek/deepseek-reasoner","name":"DeepSeek Reasoner","description":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.242,"output":0.726,"cache_read":0.0077},"canonical_model_id":"deepseek/deepseek-reasoner"},"deepseek/deepseek-v4-flash":{"id":"deepseek/deepseek-v4-flash","name":"DeepSeek V4 Flash","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.22,"output":0.66,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-v4-pro-0813":{"id":"deepseek/deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","description":"DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-13","last_updated":"2026-08-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro-0813"},"deepseek/deepseek-v4-flash-vision-exp":{"id":"deepseek/deepseek-v4-flash-vision-exp","name":"DeepSeek V4 Flash Vision Exp","description":"Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-21","last_updated":"2026-09-01","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.22,"output":0.66,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4-flash-vision-exp"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.66,"output":1.98,"cache_read":0.022},"canonical_model_id":"deepseek/deepseek-v4-pro"},"deepseek/deepseek-v4-flash-free":{"id":"deepseek/deepseek-v4-flash-free","name":"DeepSeek V4 Flash (free)","description":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0,"output":0},"canonical_model_id":"deepseek/deepseek-v4-flash"},"deepseek/deepseek-chat":{"id":"deepseek/deepseek-chat","name":"DeepSeek Chat","description":"DeepSeek chat model for instruction following, coding, and analysis","family":"deepseek","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-09","release_date":"2025-12-01","last_updated":"2026-02-28","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.22,"output":0.66,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-chat"},"deepseek/deepseek-v4-flash-0731":{"id":"deepseek/deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","description":"Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding","family":"deepseek-flash","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-07-31","last_updated":"2026-07-31","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1000000,"output":384000},"cost":{"input":0.22,"output":0.66,"cache_read":0.007},"canonical_model_id":"deepseek/deepseek-v4-flash-0731"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"experimental":{"modes":{"fast":{"cost":{"input":30,"output":150,"cache_read":3,"cache_write":37.5},"provider":{"body":{"speed":"fast"},"headers":{"anthropic-beta":"fast-mode-2026-02-01"}}}}},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":10},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"status":"deprecated","cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"tencent/hy3-free":{"id":"tencent/hy3-free","name":"Hy3 (free)","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0,"output":0},"canonical_model_id":"tencent/hy3"},"tencent/hy3":{"id":"tencent/hy3","name":"Hy3","description":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","family":"Hy","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-07-06","last_updated":"2026-07-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":256000,"input":192000,"output":128000},"cost":{"input":0.18,"output":0.59,"cache_read":0.059},"canonical_model_id":"tencent/hy3"},"minimax/minimax-m2.5-highspeed":{"id":"minimax/minimax-m2.5-highspeed","name":"MiniMax-M2.5-highspeed","description":"High-speed MiniMax model for low-latency coding and agent workflows","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-13","last_updated":"2026-02-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5-highspeed"},"minimax/minimax-m2.7":{"id":"minimax/minimax-m2.7","name":"MiniMax-M2.7","description":"Open MiniMax flagship for coding agents, office automation, and complex environments","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7"},"minimax/minimax-m3":{"id":"minimax/minimax-m3","name":"MiniMax-M3","description":"MiniMax multimodal model for long-context coding, perception, and agent planning","family":"minimax","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-06-01","last_updated":"2026-06-01","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":512000},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M3"},"minimax/minimax-m2.5":{"id":"minimax/minimax-m2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.03,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.5"},"minimax/minimax-m2.7-highspeed":{"id":"minimax/minimax-m2.7-highspeed","name":"MiniMax-M2.7-highspeed","description":"Low-latency M2.7 variant for interactive coding plans and agent loops","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"temperature":true,"release_date":"2026-03-18","last_updated":"2026-03-18","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0.6,"output":2.4,"cache_read":0.06,"cache_write":0.375},"canonical_model_id":"minimax/MiniMax-M2.7-highspeed"},"kimi/kimi-k2.6":{"id":"kimi/kimi-k2.6","name":"Kimi K2.6","description":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-20","last_updated":"2026-04-20","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.95,"output":4,"cache_read":0.16},"canonical_model_id":"moonshotai/kimi-k2.6"},"kimi/kimi-k2.7-code":{"id":"kimi/kimi-k2.7-code","name":"Kimi K2.7 Code","description":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-06-12","last_updated":"2026-06-12","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.95,"output":4,"cache_read":0.19},"canonical_model_id":"moonshotai/kimi-k2.7-code"},"kimi/kimi-k3":{"id":"kimi/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"kimi/kimi-k2.5":{"id":"kimi/kimi-k2.5","name":"Kimi K2.5","description":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","family":"kimi-k2","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":false,"knowledge":"2025-01","release_date":"2026-01-27","last_updated":"2026-01-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.6,"output":3,"cache_read":0.1,"cache_write":0},"canonical_model_id":"moonshotai/kimi-k2.5"},"grok/grok-4.5":{"id":"grok/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"grok/grok-4.6":{"id":"grok/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"grok/grok-4.3":{"id":"grok/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2,"tiers":[{"input":2.5,"output":5,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":2.5,"output":5,"cache_read":0.4}},"canonical_model_id":"xai/grok-4.3"},"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.057,"output":0.459,"tiers":[{"input":0.229,"output":1.835,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3.7-max":{"id":"qwen/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.25,"output":3.75,"cache_read":0.25,"cache_write":1.563},"canonical_model_id":"alibaba/qwen3.7-max"},"qwen/qwen3.7-plus":{"id":"qwen/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.35,"output":1.42,"cache_read":0.071,"tiers":[{"input":1.06,"output":4.25,"cache_read":0.21,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.06,"output":4.25,"cache_read":0.21}},"canonical_model_id":"alibaba/qwen3.7-plus"},"qwen/qwen3-max":{"id":"qwen/qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":0.359,"output":1.434,"tiers":[{"input":0.574,"output":2.294,"tier":{"type":"context","size":32000}},{"input":1.004,"output":4.014,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3-max"},"qwen/qwen3.8-27b":{"id":"qwen/qwen3.8-27b","name":"Qwen3.8 27B","description":"Dense 27B vision-language model for coding, agent tasks, and image and video understanding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.33,"output":2.4},"canonical_model_id":"alibaba/qwen3.8-27b"},"qwen/qwen3.6-35b-a3b":{"id":"qwen/qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B","description":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.248,"output":1.485},"canonical_model_id":"alibaba/qwen3.6-35b-a3b"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.115,"output":0.917,"tiers":[{"input":0.287,"output":2.294,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3.5-flash":{"id":"qwen/qwen3.5-flash","name":"Qwen3.5 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":1,"max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1,"output":0.4},"canonical_model_id":"alibaba/qwen3.5-flash"},"qwen/qwen3-vl-235b-a22b-thinking":{"id":"qwen/qwen3-vl-235b-a22b-thinking","name":"Qwen3 VL 235B A22B Thinking","description":"Qwen vision-language thinking model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024,"max":81920}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":40960},"cost":{"input":0.4,"output":4},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-thinking"},"qwen/qwen3.7-flash":{"id":"qwen/qwen3.7-flash","name":"Qwen3.7 Flash","description":"Lightweight multimodal Qwen model for high-throughput text, image, and video tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-15","last_updated":"2026-07-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":991808,"output":131072},"cost":{"input":0.03,"output":0.13,"cache_read":0.006,"cache_write":0.038,"tiers":[{"input":0.1,"output":0.4,"cache_read":0.02,"tier":{"type":"context","size":32000}},{"input":0.2,"output":0.8,"cache_read":0.04,"tier":{"type":"context","size":256000}}]},"canonical_model_id":"alibaba/qwen3.7-flash"},"qwen/qwen3.8-max":{"id":"qwen/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25,"cache_write":2.5},"canonical_model_id":"alibaba/qwen3.8-max"},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.086,"output":0.688,"tiers":[{"input":0.258,"output":2.064,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3.6-flash":{"id":"qwen/qwen3.6-flash","name":"Qwen3.6 Flash","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen3.6","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-27","last_updated":"2026-04-27","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5,"tiers":[{"input":1,"output":4,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1,"output":4}},"canonical_model_id":"alibaba/qwen3.6-flash"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.172,"output":1.032,"tiers":[{"input":0.43,"output":2.58,"tier":{"type":"context","size":128000}}]},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3.5-plus":{"id":"qwen/qwen3.5-plus","name":"Qwen3.5 Plus","description":"Qwen vision-language model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-02-16","last_updated":"2026-02-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.4,"output":2.4,"tiers":[{"input":0.5,"output":3,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":0.5,"output":3}},"canonical_model_id":"alibaba/qwen3.5-plus"},"qwen/qwen3.6-plus":{"id":"qwen/qwen3.6-plus","name":"Qwen3.6 Plus","description":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.276,"output":1.651,"cache_read":0.0276,"cache_write":0.345,"tiers":[{"input":1.101,"output":6.602,"cache_read":0.1101,"cache_write":1.37625,"tier":{"type":"context","size":256000}}],"context_over_200k":{"input":1.101,"output":6.602,"cache_read":0.1101,"cache_write":1.37625}},"canonical_model_id":"alibaba/qwen3.6-plus"},"qwen/qwen3-vl-235b-a22b-instruct":{"id":"qwen/qwen3-vl-235b-a22b-instruct","name":"Qwen3 VL 235B A22B Instruct","description":"Qwen vision-language instruct model for visual reasoning, documents, and agent tasks","family":"qwen","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03-31","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":32768},"cost":{"input":0.4,"output":1.6},"canonical_model_id":"alibaba/qwen3-vl-235b-a22b-instruct"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"input_audio":1},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-3.1-pro-preview-customtools":{"id":"google/gemini-3.1-pro-preview-customtools","name":"Gemini 3.1 Pro Preview Custom Tools","description":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":4,"output":18,"cache_read":0.4},"canonical_model_id":"google/gemini-3.1-pro-preview-customtools"},"google/gemini-3.1-flash-lite-preview":{"id":"google/gemini-3.1-flash-lite-preview","name":"Gemini 3.1 Flash Lite Preview","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-03-03","last_updated":"2026-03-03","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-3.1-flash-lite-preview"},"google/gemini-robotics-er-1.6-preview":{"id":"google/gemini-robotics-er-1.6-preview","name":"Gemini Robotics-ER 1.6 Preview","description":"Vision-language model for embodied reasoning: spatial understanding, task planning, and physical-world agentic robotics","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens","min":0}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-14","last_updated":"2026-04-14","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":false,"limit":{"context":131072,"output":65536},"cost":{"input":1,"output":5},"canonical_model_id":"google/gemini-robotics-er-1.6-preview"},"google/gemini-flash-latest":{"id":"google/gemini-flash-latest","name":"Gemini Flash Latest","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.1},"canonical_model_id":"google/gemini-flash-latest"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-3.1-pro-preview":{"id":"google/gemini-3.1-pro-preview","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"input_audio":2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"},"google/gemma-4-26b-a4b-it":{"id":"google/gemma-4-26b-a4b-it","name":"Gemma 4 26B A4B IT","description":"Open Gemma instruction model for efficient chat and self-hosted deployments","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.06,"output":0.33,"cache_read":0.0075},"canonical_model_id":"google/gemma-4-26b-a4b-it"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-flash-lite-latest":{"id":"google/gemini-flash-lite-latest","name":"Gemini Flash-Lite Latest","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.025,"input_audio":0.5},"canonical_model_id":"google/gemini-flash-lite-latest"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15,"cache_write":0.08333,"input_audio":3},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-3-flash-preview":{"id":"google/gemini-3-flash-preview","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05,"input_audio":1},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemma-4-31b-it":{"id":"google/gemma-4-31b-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.13,"output":0.38,"cache_read":0.02},"canonical_model_id":"google/gemma-4-31b-it"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01,"input_audio":0.3},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"google/gemini-2.5-pro"},"orcarouter/fusion-mini":{"id":"orcarouter/fusion-mini","name":"OrcaRouter Fusion Mini","description":"Leaner two-model Fusion panel that runs Claude Opus 4.8 and GPT-5.5 in parallel on hard requests, then has a Claude Opus 4.8 judge return the strongest single answer verbatim. Easy requests fall through to a cheaper default and bill as one call.","family":"model-router","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000}},"orcarouter/free":{"id":"orcarouter/free","name":"OrcaRouter Free","description":"Built-in router over the free tier that scores each request's difficulty and sends light work to the smaller free model and harder work to the stronger one. Priced at zero and never falls back to a paid model.","family":"auto","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-27","last_updated":"2026-08-27","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":65536,"output":32768},"cost":{"input":0,"output":0}},"orcarouter/fusion":{"id":"orcarouter/fusion","name":"OrcaRouter Fusion","description":"Curated fan-out router that runs Claude Opus 4.8, GPT-5.5 and Gemini 3.1 Pro in parallel on hard requests, then has a Claude Opus 4.8 judge return the strongest single answer verbatim. Easy requests fall through to a cheaper default and bill as one call.","family":"model-router","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000}},"orcarouter/fusion-flash":{"id":"orcarouter/fusion-flash","name":"OrcaRouter Fusion Flash","description":"Budget Fusion panel that runs Gemini 3.5 Flash, MiniMax M2.7 and GLM 5.1 in parallel on hard requests, then has a Claude Opus 4.8 judge return the strongest single answer verbatim. Cost-sensitive fan-out over a 200K window.","family":"model-router","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-06-15","last_updated":"2026-06-15","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":128000}},"orcarouter/auto":{"id":"orcarouter/auto","name":"OrcaRouter Auto","description":"Automatic model router for matching prompts to suitable backends and budgets","family":"auto","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2025-01-01","last_updated":"2026-05-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"cost":{"input":0,"output":0}}}},"friendli":{"id":"friendli","env":["FRIENDLI_TOKEN"],"npm":"@ai-sdk/openai-compatible","api":"https://api.friendli.ai/serverless/v1","name":"Friendli","doc":"https://friendli.ai/docs/guides/serverless_endpoints/introduction","models":{"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax-M2.5","description":"Prior MiniMax coding model for agent workflows, office edits, and automation","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196608,"output":131072},"cost":{"input":0.3,"output":1.2,"cache_read":0.06},"canonical_model_id":"minimax/MiniMax-M2.5"},"zai-org/GLM-5.2":{"id":"zai-org/GLM-5.2","name":"GLM-5.2","description":"Open flagship GLM for long-horizon coding agents and million-token context work","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-06-13","last_updated":"2026-06-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.2"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":1.26,"output":3.96,"cache_read":0.234},"canonical_model_id":"zhipuai/glm-5.3"},"zai-org/GLM-5.1":{"id":"zai-org/GLM-5.1","name":"GLM-5.1","description":"Strong GLM coding model for agentic engineering, terminals, and repository generation","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-04-07","last_updated":"2026-04-07","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":202752,"output":131072},"cost":{"input":1.4,"output":4.4,"cache_read":0.26},"canonical_model_id":"zhipuai/glm-5.1"},"zai-org/GLM-5.3-Flash":{"id":"zai-org/GLM-5.3-Flash","name":"GLM-5.3-Flash","description":"Native multimodal GLM model for efficient coding and long-horizon agent tasks","family":"glm-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high","max"]},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"release_date":"2026-08-26","last_updated":"2026-08-26","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":0.15,"output":0.5,"cache_read":0.03},"canonical_model_id":"zhipuai/glm-5.3-flash"},"deepseek-ai/DeepSeek-V3.2":{"id":"deepseek-ai/DeepSeek-V3.2","name":"DeepSeek V3.2","description":"Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use","family":"deepseek","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2024-07","release_date":"2025-12-01","last_updated":"2025-12-01","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":163840,"output":64000},"cost":{"input":0.5,"output":1.5,"cache_read":0.25},"canonical_model_id":"deepseek/deepseek-v3.2"},"google/gemma-4-31B-it":{"id":"google/gemma-4-31B-it","name":"Gemma 4 31B IT","description":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","family":"gemma","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-04-02","last_updated":"2026-04-02","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":32768},"cost":{"input":0.14,"output":0.4},"canonical_model_id":"google/gemma-4-31b-it"}}},"coralbricks":{"id":"coralbricks","env":["CORAL_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://inference.coralbricks.ai/v1","name":"CoralBricks","doc":"https://www.coralbricks.ai/docs","models":{"deepseek-v4.1-flash-fast":{"id":"deepseek-v4.1-flash-fast","name":"DeepSeek V4.1 Flash","description":"DeepSeek V4.1 Flash model for reasoning and agentic coding","family":"deepseek-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-09-10","last_updated":"2026-09-10","modalities":{"input":["text","image"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":393216},"cost":{"input":0.25,"output":1.2,"cache_read":0.005,"cache_write":0},"canonical_model_id":"deepseek/deepseek-v4.1-flash"}}},"cloudferro-sherlock":{"id":"cloudferro-sherlock","env":["CLOUDFERRO_SHERLOCK_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api-sherlock.cloudferro.com/openai/v1/","name":"CloudFerro Sherlock","doc":"https://docs.sherlock.cloudferro.com/","models":{"MiniMaxAI/MiniMax-M2.5":{"id":"MiniMaxAI/MiniMax-M2.5","name":"MiniMax-M2.5","description":"MiniMax model for chat, coding, office work, and agentic tasks","family":"minimax","attachment":false,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01","release_date":"2026-02-12","last_updated":"2026-02-12","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":196000,"input":180000,"output":16000},"cost":{"input":0.3,"output":1.2},"canonical_model_id":"minimax/MiniMax-M2.5"},"openai/gpt-oss-120b":{"id":"openai/gpt-oss-120b","name":"OpenAI GPT OSS 120B","description":"Open GPT reasoning model for self-hosted agents and controllable deployments","family":"gpt-oss","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-06","release_date":"2025-08-05","last_updated":"2025-08-05","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131000,"output":131000},"cost":{"input":2.92,"output":2.92},"canonical_model_id":"openai/gpt-oss-120b"},"speakleash/Bielik-11B-v3.0-Instruct":{"id":"speakleash/Bielik-11B-v3.0-Instruct","name":"Bielik 11B v3.0 Instruct","description":"Open-weight instruction model for adaptable chat and self-hosted production workloads","attachment":false,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-03","release_date":"2025-03-13","last_updated":"2025-03-13","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":32000,"output":32000},"cost":{"input":0.67,"output":0.67}},"meta-llama/Llama-3.3-70B-Instruct":{"id":"meta-llama/Llama-3.3-70B-Instruct","name":"Llama 3.3 70B Instruct","description":"Open Llama instruction model for multilingual chat, reasoning, and coding","family":"llama","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2024-10-09","release_date":"2024-12-06","last_updated":"2024-12-06","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":70000,"output":70000},"cost":{"input":2.92,"output":2.92},"canonical_model_id":"meta/llama-3.3-70b-instruct"},"zai-org/GLM-5.3":{"id":"zai-org/GLM-5.3","name":"GLM-5.3","description":"Flagship GLM model for long-horizon coding, agents, and complex project delivery","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-08-14","last_updated":"2026-08-14","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0.34,"output":4.48},"canonical_model_id":"zhipuai/glm-5.3"}}},"mixlayer":{"id":"mixlayer","env":["MIXLAYER_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://models.mixlayer.ai/v1","name":"Mixlayer","doc":"https://docs.mixlayer.com","models":{"qwen/qwen3.5-35b-a3b":{"id":"qwen/qwen3.5-35b-a3b","name":"Qwen3.5 35B A3B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.25,"output":1.3},"canonical_model_id":"alibaba/qwen3.5-35b-a3b"},"qwen/qwen3.5-122b-a10b":{"id":"qwen/qwen3.5-122b-a10b","name":"Qwen3.5 122B A10B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.4,"output":3.2},"canonical_model_id":"alibaba/qwen3.5-122b-a10b"},"qwen/qwen3.5-27b":{"id":"qwen/qwen3.5-27b","name":"Qwen3.5 27B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.3,"output":2.4},"canonical_model_id":"alibaba/qwen3.5-27b"},"qwen/qwen3.5-397b-a17b":{"id":"qwen/qwen3.5-397b-a17b","name":"Qwen3.5 397B A17B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"qwen/qwen3.5-9b":{"id":"qwen/qwen3.5-9b","name":"Qwen3.5 9B","description":"Qwen instruction model for multilingual chat, reasoning, and tool use","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"temperature":true,"release_date":"2026-02-23","last_updated":"2026-02-23","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":262144},"cost":{"input":0.1,"output":0.4},"canonical_model_id":"alibaba/qwen3.5-9b"}}},"kuae-cloud-coding-plan":{"id":"kuae-cloud-coding-plan","env":["KUAE_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://coding-plan-endpoint.kuaecloud.net/v1","name":"KUAE Cloud Coding Plan","doc":"https://docs.mthreads.com/kuaecloud/kuaecloud-doc-online/coding_plan/","models":{"GLM-4.7":{"id":"GLM-4.7","name":"GLM-4.7","description":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","family":"glm","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"knowledge":"2025-04","release_date":"2025-12-22","last_updated":"2025-12-22","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":204800,"output":131072},"cost":{"input":0,"output":0,"cache_read":0,"cache_write":0},"canonical_model_id":"zhipuai/glm-4.7"}}},"cloudflare-ai-gateway":{"id":"cloudflare-ai-gateway","env":["CLOUDFLARE_API_TOKEN","CLOUDFLARE_ACCOUNT_ID","CLOUDFLARE_GATEWAY_ID"],"npm":"ai-gateway-provider","name":"Cloudflare AI Gateway","doc":"https://developers.cloudflare.com/ai-gateway/","models":{"meta/muse-wonder-1.0":{"id":"meta/muse-wonder-1.0","name":"Muse Wonder","description":"Muse Wonder is a frontier multimodal reasoning model from Meta for personal and professional agents, long-horizon coding, and computer use. It combines Active Memory, which carries relevant context across conversations, with Active Focus, which keeps it working through blockers toward a goal.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-10-12","last_updated":"2026-10-12","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":4,"output":20,"cache_read":0.2},"canonical_model_id":"meta/muse-wonder-1.0"},"meta/muse-spark-1.3":{"id":"meta/muse-spark-1.3","name":"Muse Spark 1.3","description":"Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.16},"canonical_model_id":"meta/muse-spark-1.3"},"meta/muse-spark-1.1":{"id":"meta/muse-spark-1.1","name":"Muse Spark 1.1","description":"Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-04-08","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf","video"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.16},"canonical_model_id":"meta/muse-spark-1.1"},"meta/muse-spark-1.2":{"id":"meta/muse-spark-1.2","name":"Muse Spark 1.2","description":"Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-08-05","last_updated":"2026-08-05","modalities":{"input":["text","image","video","pdf","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.16},"canonical_model_id":"meta/muse-spark-1.2"},"meta/muse-spark-open":{"id":"meta/muse-spark-open","name":"Muse Spark Open","description":"Muse Spark Open is an 816B total / 42B active open-weight multimodal reasoning model from Meta for coding and long-horizon agentic work, with 1M-token context.","family":"muse","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-01-04","release_date":"2026-10-15","last_updated":"2026-10-15","modalities":{"input":["text","image","video","audio"],"output":["text"]},"open_weights":true,"limit":{"context":1040000,"output":131072},"cost":{"input":1.25,"output":4.25,"cache_read":0.15},"canonical_model_id":"meta/muse-spark-open"},"moonshotai/kimi-k3":{"id":"moonshotai/kimi-k3","name":"Kimi K3","description":"Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work","family":"kimi-k3","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"release_date":"2026-07-16","last_updated":"2026-07-16","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":1048576,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"moonshotai/kimi-k3"},"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-5.1":{"id":"openai/gpt-5.1","name":"GPT-5.1","description":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-09-30","release_date":"2025-11-13","last_updated":"2025-11-13","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5.1"},"openai/gpt-4.1-mini":{"id":"openai/gpt-4.1-mini","name":"GPT-4.1 mini","description":"Affordable GPT-4.1 lane for fast coding help and structured extraction","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.4,"output":1.6,"cache_read":0.1},"canonical_model_id":"openai/gpt-4.1-mini"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/o3-mini":{"id":"openai/o3-mini","name":"o3-mini","description":"Smaller o-series reasoner for economical coding, math, and planning tasks","family":"o-mini","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2024-12-20","last_updated":"2025-01-29","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":1.1,"output":4.4,"cache_read":0.55},"canonical_model_id":"openai/o3-mini"},"openai/gpt-5.5-pro":{"id":"openai/gpt-5.5-pro","name":"GPT-5.5 Pro","description":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.5-pro"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-5-mini":{"id":"openai/gpt-5-mini","name":"GPT-5 Mini","description":"Small GPT-5 for responsive agents, coding help, and everyday automation","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.25,"output":2,"cache_read":0.025},"canonical_model_id":"openai/gpt-5-mini"},"openai/gpt-5.4-pro":{"id":"openai/gpt-5.4-pro","name":"GPT-5.4 Pro","description":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","family":"gpt-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["medium","high","xhigh"]}],"tool_call":true,"structured_output":false,"temperature":false,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":30,"output":180},"canonical_model_id":"openai/gpt-5.4-pro"},"openai/gpt-5-nano":{"id":"openai/gpt-5-nano","name":"GPT-5 Nano","description":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.05,"output":0.4,"cache_read":0.005},"canonical_model_id":"openai/gpt-5-nano"},"openai/gpt-5.5":{"id":"openai/gpt-5.5","name":"GPT-5.5","description":"Default frontier GPT for coding, computer use, research, and knowledge work","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2025-12-01","release_date":"2026-04-23","last_updated":"2026-04-23","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":5,"output":30},"canonical_model_id":"openai/gpt-5.5"},"openai/o3":{"id":"openai/o3","name":"o3","description":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","family":"o","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/o3"},"openai/gpt-5.4":{"id":"openai/gpt-5.4","name":"GPT-5.4","description":"Agent-ready GPT for coding and computer-use workflows at a lower cost","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-05","last_updated":"2026-03-05","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2.5,"output":15,"cache_read":0.25},"canonical_model_id":"openai/gpt-5.4"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":10,"cache_read":0.25,"cache_write":3.125},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25},"canonical_model_id":"openai/gpt-5.6-luna"},"openai/gpt-4o":{"id":"openai/gpt-4o","name":"GPT-4o","description":"Omni-era GPT for multimodal chat, practical coding, and general assistants","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-05-13","last_updated":"2024-08-06","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":1.25,"output":5,"cache_read":0.625},"canonical_model_id":"openai/gpt-4o"},"openai/gpt-4.1-nano":{"id":"openai/gpt-4.1-nano","name":"GPT-4.1 nano","description":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","family":"gpt-nano","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":32768},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.1,"output":0.4,"cache_read":0.025},"canonical_model_id":"openai/gpt-4.1-nano"},"openai/gpt-4o-mini":{"id":"openai/gpt-4o-mini","name":"GPT-4o mini","description":"Small omni GPT for cheap multimodal assistance and production-scale traffic","family":"gpt-mini","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2023-09","release_date":"2024-07-18","last_updated":"2024-07-18","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"output":16384},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.075,"output":0.3,"cache_read":0.0375},"canonical_model_id":"openai/gpt-4o-mini"},"openai/o4-mini":{"id":"openai/o4-mini","name":"o4-mini","description":"Fast o-series model for compact reasoning, coding, and tool use","family":"o-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-05","release_date":"2025-04-16","last_updated":"2025-04-16","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":100000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":1.1,"output":4.4,"cache_read":0.275},"canonical_model_id":"openai/o4-mini"},"openai/gpt-4.1":{"id":"openai/gpt-4.1","name":"GPT-4.1","description":"Long-lived GPT workhorse for coding, instruction following, and production apps","family":"gpt","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2024-04","release_date":"2025-04-14","last_updated":"2025-04-14","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1047576,"output":32768},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":2,"output":8,"cache_read":0.5},"canonical_model_id":"openai/gpt-4.1"},"openai/gpt-5":{"id":"openai/gpt-5","name":"GPT-5","description":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","family":"gpt","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2024-09-30","release_date":"2025-08-07","last_updated":"2025-08-07","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"openai/gpt-5"},"openai/gpt-5.4-mini":{"id":"openai/gpt-5.4-mini","name":"GPT-5.4 mini","description":"Strong small GPT for coding subagents, quick tool use, and high-volume work","family":"gpt-mini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.75,"output":4.5,"cache_read":0.075},"canonical_model_id":"openai/gpt-5.4-mini"},"openai/gpt-5.4-nano":{"id":"openai/gpt-5.4-nano","name":"GPT-5.4 nano","description":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","family":"gpt-nano","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-03-17","last_updated":"2026-03-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":128000,"input":272000,"output":128000},"provider":{"npm":"@ai-sdk/openai"},"cost":{"input":0.2,"output":1.25,"cache_read":0.02},"canonical_model_id":"openai/gpt-5.4-nano"},"deepseek/deepseek-v4-pro":{"id":"deepseek/deepseek-v4-pro","name":"DeepSeek V4 Pro","description":"Open MoE flagship with million-token context for coding and long agent runs","family":"deepseek-thinking","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2026-04-24","last_updated":"2026-04-24","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":131072,"output":384000},"cost":{"input":1.74,"output":3.48,"cache_read":0.145},"canonical_model_id":"deepseek/deepseek-v4-pro"},"anthropic/claude-sonnet-5.5":{"id":"anthropic/claude-sonnet-5.5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-opus-4.7":{"id":"anthropic/claude-opus-4.7","name":"Claude Opus 4.7","description":"Stronger Opus tier for advanced software work and high-stakes reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-04-16","last_updated":"2026-04-16","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-7"},"anthropic/claude-sonnet-4.6":{"id":"anthropic/claude-sonnet-4.6","name":"Claude Sonnet 4.6","description":"Claude workhorse for coding agents, careful analysis, and production cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-08-31","release_date":"2026-02-17","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-6"},"anthropic/claude-haiku-4.5":{"id":"anthropic/claude-haiku-4.5","name":"Claude Haiku 4.5 (latest)","description":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-02-28","release_date":"2025-10-15","last_updated":"2025-10-15","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":1,"output":5,"cache_read":0.1,"cache_write":1.25},"canonical_model_id":"anthropic/claude-haiku-4-5"},"anthropic/claude-fable-5":{"id":"anthropic/claude-fable-5","name":"Claude Fable 5","description":"Claude model for creative writing, analysis, and controlled agent workflows","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-09","last_updated":"2026-06-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5"},"anthropic/claude-opus-4.6":{"id":"anthropic/claude-opus-4.6","name":"Claude Opus 4.6","description":"High-end Claude for difficult coding, planning, and slower expert reasoning","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","max"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05-31","release_date":"2026-02-05","last_updated":"2026-03-13","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-6"},"anthropic/claude-opus-5":{"id":"anthropic/claude-opus-5","name":"Claude Opus 5","description":"Strongest Claude Opus model for coding, agents, and professional work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05","release_date":"2026-07-24","last_updated":"2026-07-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-5"},"anthropic/claude-opus-5.5":{"id":"anthropic/claude-opus-5.5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-opus-4.5":{"id":"anthropic/claude-opus-4.5","name":"Claude Opus 4.5 (latest)","description":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]},{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-05","release_date":"2025-11-24","last_updated":"2025-11-24","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-5"},"anthropic/claude-sonnet-4.5":{"id":"anthropic/claude-sonnet-4.5","name":"Claude Sonnet 4.5 (latest)","description":"Balanced Claude model for coding, analysis, agent workflows, and cost control","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"budget_tokens","min":1024}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-07-31","release_date":"2025-09-29","last_updated":"2025-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":200000,"output":64000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":3,"output":15,"cache_read":0.3,"cache_write":3.75},"canonical_model_id":"anthropic/claude-sonnet-4-5"},"anthropic/claude-opus-4.8":{"id":"anthropic/claude-opus-4.8","name":"Claude Opus 4.8","description":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01","release_date":"2026-05-28","last_updated":"2026-05-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":5,"output":25,"cache_read":0.5,"cache_write":6.25},"canonical_model_id":"anthropic/claude-opus-4-8"},"anthropic/claude-haiku-5.5":{"id":"anthropic/claude-haiku-5.5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-sonnet-5":{"id":"anthropic/claude-sonnet-5","name":"Claude Sonnet 5","description":"Everyday Claude agent model for coding, planning, browsing, and general work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-01-31","release_date":"2026-06-30","last_updated":"2026-06-30","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5"},"anthropic/claude-fable-5.1":{"id":"anthropic/claude-fable-5.1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"reflection-ai/beam-1.0-501b-a23b":{"id":"reflection-ai/beam-1.0-501b-a23b","name":"Beam 1.0 501B A23B","description":"Sparse mixture-of-experts reasoning model with 23B active parameters for coding and agentic workflows","family":"beam","attachment":false,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"structured_output":true,"temperature":true,"knowledge":"2026-06-30","release_date":"2026-10-05","last_updated":"2026-10-16","modalities":{"input":["text"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":131072},"cost":{"input":0,"output":0,"cache_read":0},"canonical_model_id":"reflection/beam-1.0-501b-a23b"},"fireworks/ember-1":{"id":"fireworks/ember-1","name":"Ember-1","description":"Specialized model from Fireworks built on Kimi K3, producing shorter reasoning traces with approximately 40% fewer tokens while maintaining comparable quality","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","high","max"]}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":1040000,"output":131072},"cost":{"input":3,"output":15,"cache_read":0.3},"canonical_model_id":"fireworks/ember-1"},"xai/grok-4.5":{"id":"xai/grok-4.5","name":"Grok 4.5","description":"xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-07-08","last_updated":"2026-07-08","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.3,"tiers":[{"input":4,"output":12,"cache_read":0.6,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":0.6}},"canonical_model_id":"xai/grok-4.5"},"xai/grok-4.6":{"id":"xai/grok-4.6","name":"Grok 4.6","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-02-01","release_date":"2026-08-12","last_updated":"2026-08-12","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.6"},"xai/grok-4.20-0309-non-reasoning":{"id":"xai/grok-4.20-0309-non-reasoning","name":"Grok 4.20 (Non-Reasoning)","description":"Grok model for agentic tool use, reasoning, coding, and live assistance","family":"grok","attachment":true,"reasoning":false,"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"xai/grok-4.20-0309-non-reasoning"},"xai/grok-4.7":{"id":"xai/grok-4.7","name":"Grok 4.7","description":"xAI's frontier model for long-running agents, coding, knowledge work, and visual projects","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-05","release_date":"2026-09-21","last_updated":"2026-09-21","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":500000,"output":500000},"cost":{"input":2,"output":6,"cache_read":0.5,"tiers":[{"input":4,"output":12,"cache_read":1,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":12,"cache_read":1}},"canonical_model_id":"xai/grok-4.7"},"xai/grok-4.3":{"id":"xai/grok-4.3","name":"Grok 4.3","description":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-04-17","last_updated":"2026-04-17","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":30000},"cost":{"input":1.25,"output":2.5,"cache_read":0.2},"canonical_model_id":"xai/grok-4.3"},"xai/grok-4.20-0309-reasoning":{"id":"xai/grok-4.20-0309-reasoning","name":"Grok 4.20 (Reasoning)","description":"Reasoning Grok for document-heavy analysis and long-horizon tool use","family":"grok","attachment":true,"reasoning":true,"reasoning_options":[],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-03-09","last_updated":"2026-03-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":2000000,"output":30000},"cost":{"input":2,"output":6,"cache_read":0.2},"canonical_model_id":"xai/grok-4.20-0309-reasoning"},"alibaba/qwen3.7-max":{"id":"alibaba/qwen3.7-max","name":"Qwen3.7 Max","description":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","family":"qwen","attachment":false,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"release_date":"2026-05-21","last_updated":"2026-05-21","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":1.25,"output":3.75,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.7-max"},"alibaba/qwen3.7-plus":{"id":"alibaba/qwen3.7-plus","name":"Qwen3.7 Plus","description":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2026-06-02","last_updated":"2026-06-02","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":64000},"cost":{"input":0.32,"output":1.28,"cache_read":0.064},"canonical_model_id":"alibaba/qwen3.7-plus"},"alibaba/qwen3-max":{"id":"alibaba/qwen3-max","name":"Qwen3 Max","description":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","family":"qwen","attachment":false,"reasoning":false,"tool_call":true,"temperature":true,"knowledge":"2025-04","release_date":"2025-09-23","last_updated":"2025-09-23","modalities":{"input":["text"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":65536},"cost":{"input":1.2,"output":6},"canonical_model_id":"alibaba/qwen3-max"},"alibaba/qwen3.8-max":{"id":"alibaba/qwen3.8-max","name":"Qwen3.8 Max","description":"2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","xhigh"]},{"type":"budget_tokens","min":0,"max":262144}],"tool_call":true,"interleaved":{"field":"reasoning_content"},"temperature":true,"release_date":"2026-08-03","last_updated":"2026-08-03","modalities":{"input":["text","image","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":131072},"cost":{"input":2,"output":6,"cache_read":0.25},"canonical_model_id":"alibaba/qwen3.8-max"},"alibaba/qwen3.5-397b-a17b":{"id":"alibaba/qwen3.5-397b-a17b","name":"Qwen3.5 397B-A17B","description":"Large open Qwen multimodal MoE for visual agents and long technical tasks","family":"qwen","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"budget_tokens"}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-02-15","last_updated":"2026-02-15","modalities":{"input":["text","image","video"],"output":["text"]},"open_weights":true,"limit":{"context":262144,"output":65536},"cost":{"input":0.6,"output":3.6},"canonical_model_id":"alibaba/qwen3.5-397b-a17b"},"unbiased/pareto":{"id":"unbiased/pareto","name":"Pareto","description":"Blended multimodal model that runs multiple language models in parallel and returns a synthesized answer","attachment":true,"reasoning":false,"tool_call":true,"structured_output":false,"temperature":true,"release_date":"2026-09-17","last_updated":"2026-09-17","modalities":{"input":["text","image"],"output":["text"]},"open_weights":false,"limit":{"context":262144,"output":131072},"cost":{"input":2.5,"output":7.5,"cache_read":0.25},"canonical_model_id":"unbiased/pareto"},"google/gemini-3.8-flash":{"id":"google/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.8-flash"},"google/gemini-3.1-flash-lite":{"id":"google/gemini-3.1-flash-lite","name":"Gemini 3.1 Flash Lite","description":"Low-latency Gemini model for high-volume multimodal and agent workloads","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-07","last_updated":"2026-05-07","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.25,"output":1.5,"cache_read":0.03},"canonical_model_id":"google/gemini-3.1-flash-lite"},"google/gemini-2.5-flash":{"id":"google/gemini-2.5-flash","name":"Gemini 2.5 Flash","description":"Fast Gemini workhorse for multimodal apps where latency and price matter","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03},"canonical_model_id":"google/gemini-2.5-flash"},"google/gemini-4-argon":{"id":"google/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":2,"output":10,"cache_read":0.1},"canonical_model_id":"google/gemini-4-argon"},"google/gemini-3-flash":{"id":"google/gemini-3-flash","name":"Gemini 3 Flash Preview","description":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-12-17","last_updated":"2025-12-17","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.5,"output":3,"cache_read":0.05},"canonical_model_id":"google/gemini-3-flash-preview"},"google/gemini-3.5-flash-lite":{"id":"google/gemini-3.5-flash-lite","name":"Gemini 3.5 Flash Lite","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.3,"output":2.5,"cache_read":0.03,"cache_write":0.03},"canonical_model_id":"google/gemini-3.5-flash-lite"},"google/gemini-3.6-flash":{"id":"google/gemini-3.6-flash","name":"Gemini 3.6 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-07-21","last_updated":"2026-07-21","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":7.5,"cache_read":0.15,"cache_write":1.5},"canonical_model_id":"google/gemini-3.6-flash"},"google/gemini-3.7-flash":{"id":"google/gemini-3.7-flash","name":"Gemini 3.7 Flash","description":"High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2026-03","release_date":"2026-08-13","last_updated":"2026-08-13","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075},"canonical_model_id":"google/gemini-3.7-flash"},"google/gemini-3.5-flash":{"id":"google/gemini-3.5-flash","name":"Gemini 3.5 Flash","description":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-05-19","last_updated":"2026-05-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":1.5,"output":9,"cache_read":0.15},"canonical_model_id":"google/gemini-3.5-flash"},"google/gemini-2.5-flash-lite":{"id":"google/gemini-2.5-flash-lite","name":"Gemini 2.5 Flash-Lite","description":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","family":"gemini-flash-lite","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":0.1,"output":0.4,"cache_read":0.01},"canonical_model_id":"google/gemini-2.5-flash-lite"},"google/gemini-2.5-pro":{"id":"google/gemini-2.5-pro","name":"Gemini 2.5 Pro","description":"Google's proven reasoning model for coding, math, and multimodal analysis","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["minimal","low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2025-06-17","last_updated":"2025-06-17","modalities":{"input":["text","image","audio","video","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":1.25,"output":10,"cache_read":0.125},"canonical_model_id":"google/gemini-2.5-pro"},"google/gemini-3.1-pro":{"id":"google/gemini-3.1-pro","name":"Gemini 3.1 Pro Preview","description":"Reasoning-first Gemini preview for agentic coding and complex problem solving","family":"gemini-pro","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"structured_output":true,"temperature":true,"knowledge":"2025-01","release_date":"2026-02-19","last_updated":"2026-02-19","modalities":{"input":["text","image","video","audio","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":65536},"cost":{"input":2,"output":12,"cache_read":0.2,"tiers":[{"input":4,"output":18,"cache_read":0.4,"tier":{"type":"context","size":200000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4}},"canonical_model_id":"google/gemini-3.1-pro-preview"}}},"aixy":{"id":"aixy","env":["AIXY_API_KEY"],"npm":"@ai-sdk/openai-compatible","api":"https://api.aixy-gateway.com/v1","name":"Aixy","doc":"https://docs.aixy-gateway.com/integrations/overview","models":{"openai/gpt-6-sol":{"id":"openai/gpt-6-sol","name":"GPT-6 Sol","description":"OpenAI model for complex coding and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-20","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.aixy-gateway.com/v1","shape":"responses"},"cost":{"input":2,"output":10,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-6-sol"},"openai/gpt-6-astra":{"id":"openai/gpt-6-astra","name":"GPT-6 Astra","description":"GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.","family":"gpt-astra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-04","last_updated":"2026-09-04","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.aixy-gateway.com/v1","shape":"responses"},"cost":{"input":10,"output":50,"cache_read":1,"cache_write":12.5,"tiers":[{"input":20,"output":75,"cache_read":2,"cache_write":25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":20,"output":75,"cache_read":2,"cache_write":25}},"canonical_model_id":"openai/gpt-6-astra"},"openai/gpt-5.6-terra":{"id":"openai/gpt-5.6-terra","name":"GPT-5.6 Terra","description":"Balanced GPT-5.6 model for capable, cost-efficient everyday work","family":"gpt-terra","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.aixy-gateway.com/v1","shape":"responses"},"cost":{"input":2,"output":12,"cache_read":0.2,"cache_write":2.5,"tiers":[{"input":4,"output":18,"cache_read":0.4,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":18,"cache_read":0.4,"cache_write":5}},"canonical_model_id":"openai/gpt-5.6-terra"},"openai/gpt-6-luna":{"id":"openai/gpt-6-luna","name":"GPT-6 Luna","description":"OpenAI's most efficient model for focused, high-volume tasks","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-05-18","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.aixy-gateway.com/v1","shape":"responses"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.2,"output":0.75,"cache_read":0.02,"cache_write":0.25}},"canonical_model_id":"openai/gpt-6-luna"},"openai/gpt-5.6-sol":{"id":"openai/gpt-5.6-sol","name":"GPT-5.6 Sol","description":"Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.aixy-gateway.com/v1","shape":"responses"},"cost":{"input":4,"output":20,"cache_read":0.4,"cache_write":5,"tiers":[{"input":8,"output":30,"cache_read":0.8,"cache_write":10,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":8,"output":30,"cache_read":0.8,"cache_write":10}},"canonical_model_id":"openai/gpt-5.6-sol"},"openai/gpt-6.1-sol":{"id":"openai/gpt-6.1-sol","name":"GPT-6.1 Sol","description":"Near-Astra performance for complex coding, computer use, and professional work at a lower cost.","family":"gpt-sol","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-04-30","release_date":"2026-09-29","last_updated":"2026-09-29","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.aixy-gateway.com/v1","shape":"responses"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5,"tiers":[{"input":4,"output":15,"cache_read":0.2,"cache_write":5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":4,"output":15,"cache_read":0.2,"cache_write":5}},"canonical_model_id":"openai/gpt-6.1-sol"},"openai/gpt-5.6-luna":{"id":"openai/gpt-5.6-luna","name":"GPT-5.6 Luna","description":"Cost-efficient GPT-5.6 model for fast, high-volume workloads","family":"gpt-luna","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["none","low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-02-16","release_date":"2026-07-09","last_updated":"2026-07-09","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1050000,"input":922000,"output":128000},"provider":{"npm":"@ai-sdk/openai","api":"https://api.aixy-gateway.com/v1","shape":"responses"},"cost":{"input":0.2,"output":1.2,"cache_read":0.02,"cache_write":0.25,"tiers":[{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5,"tier":{"type":"context","size":272000}}],"context_over_200k":{"input":0.4,"output":1.8,"cache_read":0.04,"cache_write":0.5}},"canonical_model_id":"openai/gpt-5.6-luna"},"anthropic/claude-sonnet-5-5":{"id":"anthropic/claude-sonnet-5-5","name":"Claude Sonnet 5.5","description":"Fast Claude model for everyday coding, agents, and knowledge work","family":"claude-sonnet","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-28","last_updated":"2026-09-28","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.aixy-gateway.com/v1"},"cost":{"input":2,"output":10,"cache_read":0.1,"cache_write":2.5},"canonical_model_id":"anthropic/claude-sonnet-5-5"},"anthropic/claude-haiku-5-5":{"id":"anthropic/claude-haiku-5-5","name":"Claude Haiku 5.5","description":"Fast Claude model for responsive assistance, classification, and lightweight agents","family":"claude-haiku","attachment":true,"reasoning":true,"reasoning_options":[{"type":"toggle"},{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-10-07","last_updated":"2026-10-07","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.aixy-gateway.com/v1"},"cost":{"input":0.1,"output":0.5,"cache_read":0.01,"cache_write":0.125,"tiers":[{"input":0.5,"output":2.5,"cache_read":0.05,"cache_write":0.625,"tier":{"type":"context","size":100000}}]},"canonical_model_id":"anthropic/claude-haiku-5-5"},"anthropic/claude-opus-5-5":{"id":"anthropic/claude-opus-5-5","name":"Claude Opus 5.5","description":"Claude model for long-running agentic coding and knowledge work","family":"claude-opus","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-22","last_updated":"2026-09-22","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.aixy-gateway.com/v1"},"cost":{"input":4,"output":20,"cache_read":0.2,"cache_write":5},"canonical_model_id":"anthropic/claude-opus-5-5"},"anthropic/claude-fable-5-1":{"id":"anthropic/claude-fable-5-1","name":"Claude Fable 5.1","description":"Claude model for demanding reasoning and long-horizon agentic work","family":"claude-fable","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high","xhigh","max"]}],"tool_call":true,"structured_output":true,"temperature":false,"knowledge":"2026-06","release_date":"2026-09-01","last_updated":"2026-09-01","modalities":{"input":["text","image","pdf"],"output":["text"]},"open_weights":false,"limit":{"context":1000000,"output":128000},"provider":{"npm":"@ai-sdk/anthropic","api":"https://api.aixy-gateway.com/v1"},"cost":{"input":10,"output":50,"cache_read":0.25,"cache_write":12.5},"canonical_model_id":"anthropic/claude-fable-5-1"},"gemini/gemini-3.8-flash":{"id":"gemini/gemini-3.8-flash","name":"Gemini 3.8 Flash","description":"Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows","family":"gemini-flash","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":true,"release_date":"2026-09-02","last_updated":"2026-09-02","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":65536},"cost":{"input":0.75,"output":3.75,"cache_read":0.075,"input_audio":0.75},"canonical_model_id":"google/gemini-3.8-flash"},"gemini/gemini-4-argon":{"id":"gemini/gemini-4-argon","name":"Gemini 4 Argon","description":"Google's flagship reasoning model for complex software engineering, deep math and science, and long-horizon autonomous agents","family":"gemini","attachment":true,"reasoning":true,"reasoning_options":[{"type":"effort","values":["low","medium","high"]}],"tool_call":true,"temperature":false,"knowledge":"2026-03","release_date":"2026-10-13","last_updated":"2026-10-13","modalities":{"input":["text","image","audio"],"output":["text"]},"open_weights":false,"limit":{"context":1048576,"output":32768},"cost":{"input":2,"output":10,"cache_read":0.1,"input_audio":2},"canonical_model_id":"google/gemini-4-argon"}}}} \ No newline at end of file diff --git a/packages/core/src/permission.ts b/packages/core/src/permission.ts index 1dc798807b6c..37d58111af08 100644 --- a/packages/core/src/permission.ts +++ b/packages/core/src/permission.ts @@ -24,40 +24,27 @@ export type ID = typeof ID.Type export const Source = Permission.Source export type Source = typeof Source.Type -const RequestFields = { - sessionID: Permission.Request.fields.sessionID, - action: Permission.Request.fields.action, - resources: Permission.Request.fields.resources, - save: Permission.Request.fields.save, - metadata: Permission.Request.fields.metadata, - source: Permission.Request.fields.source, -} - export const Request = Permission.Request export type Request = typeof Request.Type export const Reply = Permission.Reply export type Reply = typeof Reply.Type -export const AssertInput = Schema.Struct({ - id: ID.pipe(Schema.optional), - ...RequestFields, - agent: Agent.ID.pipe(Schema.optional), -}).annotate({ identifier: "Permission.AssertInput" }) -export type AssertInput = typeof AssertInput.Type +export interface AssertInput extends Omit { + readonly id?: ID + readonly agent?: Agent.ID +} + +export interface ReplyInput { + readonly requestID: ID + readonly reply: Reply + readonly message?: string +} -export const ReplyInput = Schema.Struct({ - requestID: ID, - reply: Reply, - message: Schema.String.pipe(Schema.optional), -}).annotate({ identifier: "Permission.ReplyInput" }) -export type ReplyInput = typeof ReplyInput.Type - -export const AskResult = Schema.Struct({ - id: ID, - effect: Permission.Effect, -}).annotate({ identifier: "Permission.AskResult" }) -export type AskResult = typeof AskResult.Type +export interface AskResult { + readonly id: ID + readonly effect: Permission.Effect +} export { Event } from "@opencode/schema/permission" diff --git a/packages/core/src/permission/saved.ts b/packages/core/src/permission/saved.ts index 78aa4124c848..58cd16bb6c84 100644 --- a/packages/core/src/permission/saved.ts +++ b/packages/core/src/permission/saved.ts @@ -4,7 +4,7 @@ import { eq } from "drizzle-orm" import { Cause, Context, DateTime, Effect, Layer, Schema } from "effect" import { isSqlError } from "effect/sql/SqlError" import type { EffectDrizzleQueryError } from "drizzle-orm/effect-core/errors" -import { Project } from "@opencode/schema/project" +import type { Project } from "@opencode/schema/project" import { Database } from "../database/database.js" import { makeGlobalNode } from "@opencode/util/effect/app-node" import { PermissionTable } from "./sql.js" @@ -16,24 +16,21 @@ export type ID = typeof ID.Type export const Info = PermissionSaved.Info export type Info = typeof Info.Type -export const ListInput = Schema.Struct({ - projectID: Project.ID.pipe(Schema.optional), -}).annotate({ identifier: "PermissionSaved.ListInput" }) -export type ListInput = typeof ListInput.Type +export interface ListInput { + readonly projectID?: Project.ID +} -export const AddInput = Schema.Struct({ - projectID: Project.ID, - action: Schema.String, - resources: Schema.Array(Schema.String), -}).annotate({ identifier: "PermissionSaved.AddInput" }) -export type AddInput = typeof AddInput.Type +export interface AddInput { + readonly projectID: Project.ID + readonly action: string + readonly resources: readonly string[] +} -export const UpdateInput = Schema.Struct({ - id: ID, - action: Schema.String.pipe(Schema.optional), - resource: Schema.String.pipe(Schema.optional), -}).annotate({ identifier: "PermissionSaved.UpdateInput" }) -export type UpdateInput = typeof UpdateInput.Type +export interface UpdateInput { + readonly id: ID + readonly action?: string + readonly resource?: string +} export class NotFoundError extends Schema.TaggedError()("PermissionSaved.NotFoundError", { id: ID, diff --git a/packages/core/src/permission/sql.ts b/packages/core/src/permission/sql.ts index 4283f13411af..a0967603ee25 100644 --- a/packages/core/src/permission/sql.ts +++ b/packages/core/src/permission/sql.ts @@ -2,7 +2,7 @@ import { sqliteTable, text, uniqueIndex } from "drizzle-orm/sqlite-core" import type { Project } from "@opencode/schema/project" import { Timestamps } from "../database/schema.sql.js" import { ProjectTable } from "../project/sql.js" -import type { PermissionSaved } from "./saved.js" +import type { PermissionSaved } from "@opencode/schema/permission-saved" export const PermissionTable = sqliteTable( "permission", diff --git a/packages/core/src/persistent-pty/daemon.ts b/packages/core/src/persistent-pty/daemon.ts index 0b4a4a7f9a95..21bcaab2a37f 100644 --- a/packages/core/src/persistent-pty/daemon.ts +++ b/packages/core/src/persistent-pty/daemon.ts @@ -4,6 +4,7 @@ import net from "node:net" import path from "node:path" import { Cause, Data, Duration, Effect, Schema, Semaphore } from "effect" import type { Handoff } from "@opencode/schema/persistent-pty" +import { Session } from "@opencode/schema/session" const ProtocolVersion = 7 const MaxFrameBytes = 8 * 1024 * 1024 @@ -19,7 +20,7 @@ export const WireTerminal = Schema.Struct({ pid: Schema.NullOr(Schema.Number), title: Schema.String, foreground_process: Schema.NullOr(Schema.String), - group_id: Schema.String, + group_id: Session.ID, command: Schema.Array(Schema.String), cwd: Schema.String, cols: Schema.Number, @@ -681,8 +682,10 @@ async function* decoder(socket: net.Socket) { if (pending.length !== 0) throw new Error("opencode-pty truncated frame") } +const decodeWireResponse = Schema.decodeUnknownSync(Schema.fromJsonString(WireResponse)) + function decode(payload: Uint8Array) { - return Schema.decodeUnknownSync(WireResponse)(JSON.parse(Buffer.from(payload).toString("utf8"))) + return decodeWireResponse(Buffer.from(payload).toString("utf8")) } function failure(kind: DaemonError["kind"], cause: unknown) { diff --git a/packages/core/src/persistent-pty/index.ts b/packages/core/src/persistent-pty/index.ts index 69c673b0142c..780880fb2523 100644 --- a/packages/core/src/persistent-pty/index.ts +++ b/packages/core/src/persistent-pty/index.ts @@ -337,7 +337,7 @@ const makeLayer = (options: Options = {}) => }, }) .pipe(Effect.mapError(unavailable)) - if (attachment.role === "controller") current.set(Session.ID.make(attachment.terminal.group_id), id) + if (attachment.role === "controller") current.set(attachment.terminal.group_id, id) return { info: toInfo(attachment.terminal), role: attachment.role, @@ -400,7 +400,7 @@ function toInfo(value: WireTerminal): Info { pid: value.pid ?? 0, ...(status === "exited" ? { exitCode: value.lifecycle.exit_code ?? undefined } : {}), }), - sessionID: Session.ID.make(value.group_id), + sessionID: value.group_id, foregroundProcess: value.foreground_process, size: { cols: value.cols, rows: value.rows }, output: { head: value.output_head, tail: value.output_tail }, diff --git a/packages/core/src/plugin.ts b/packages/core/src/plugin.ts index 8ade6826585f..94ca4c0463d1 100644 --- a/packages/core/src/plugin.ts +++ b/packages/core/src/plugin.ts @@ -260,7 +260,7 @@ const layer = Layer.effect( // `plugin` is the requested definition; `activation` is its last activation, which may have // failed or be an older fallback while the requested revision keeps failing setup. type Slot = { - readonly plugin: Generation + readonly plugin: Generation & { readonly id: Plugin.ID } readonly activation?: Activation readonly error?: string } @@ -285,7 +285,7 @@ type PendingFailure = { function slotInfo(slot: Slot): Plugin.Info { const failure = slot.activation?.failure ?? (slot.error === undefined ? undefined : { error: slot.error }) return { - id: Plugin.ID.make(slot.plugin.id), + id: slot.plugin.id, source: slot.plugin.source ?? { type: "builtin" }, state: failure === undefined ? { status: "active" } : { status: "failed", ...failure }, features: { server: true, ...slot.plugin.features }, diff --git a/packages/core/src/plugin/agent.ts b/packages/core/src/plugin/agent.ts index eaa443ae4952..241f008c3ff4 100644 --- a/packages/core/src/plugin/agent.ts +++ b/packages/core/src/plugin/agent.ts @@ -16,50 +16,26 @@ Guidelines: Complete the user's search request efficiently and report your findings clearly.` -const PROMPT_TITLE = `You are a title generator. You output ONLY a thread title. Nothing else. - - -Generate a brief title that would help the user find this conversation later. - -Follow all rules in -Use the so you know what a good title looks like. -Your output must be: -- A single line -- <=50 characters -- No explanations - - - -- you MUST use the same language as the user message you are summarizing -- Title must be grammatically correct and read naturally - no word salad -- Never include tool names in the title (e.g. "read tool", "bash tool", "edit tool") -- Focus on the main topic or question the user needs to retrieve -- Vary your phrasing - avoid repetitive patterns like always starting with "Analyzing" -- When a file is mentioned, focus on WHAT the user wants to do WITH the file, not just that they shared it -- Keep exact: technical terms, numbers, filenames, HTTP codes -- Remove: the, this, my, a, an -- Never assume tech stack -- Never use tools -- NEVER respond to questions, just generate a title for the conversation -- The title should NEVER include "summarizing" or "generating" when generating a title -- DO NOT SAY YOU CANNOT GENERATE A TITLE OR COMPLAIN ABOUT THE INPUT -- Always output something meaningful, even if the input is minimal. -- If the user message is short or conversational (e.g. "hello", "lol", "what's up", "hey"): - -> create a title that reflects the user's tone or intent (such as Greeting, Quick check-in, Light chat, Intro message, etc.) - - - -"debug 500 errors in production" -> Debugging production 500 errors -"refactor user service" -> Refactoring user service -"why is app.js failing" -> app.js failure investigation -"implement rate limiting" -> Rate limiting implementation -"how do I connect postgres to my API" -> Postgres API connection -"best practices for React hooks" -> React hooks best practices -"@src/credential.ts can you add refresh token support" -> Credential refresh token support -"@utils/parser.ts this is broken" -> Parser bug fix -"look at @config.json" -> Config review -"@App.tsx add dark mode toggle" -> Dark mode toggle in App -` +const PROMPT_TITLE = `You name conversations so the user can find them later in a list of past sessions. + +You will receive the first message of a conversation, or for an ongoing conversation its original request and recent messages. Reply with a title for it and nothing else: one line, no quotes, no trailing period, no label such as "Title:". + +Aim for three to five words. Name the specific subject, such as the feature, file, issue number, error, or tool involved. Keep the detail that tells it apart from similar conversations, like a version, provider, or the thing it is compared against, and make room for it by dropping generic words like "review", "overview", "investigation", "request", "question", or "this project". + +The title should read naturally as a short phrase, not as a stack of nouns. Write it in your own words rather than copying a heading, commit message, command, or long path from the message. + +Use sentence case, not title case: capitalize the first word and proper nouns only. Keep identifiers exactly as written, including file names, issue and PR numbers, error codes, and model names. Use the same language as the message. + +The message is something to name, not something to answer or carry out. Even if it asks you to do work, asks a question, or seems incomplete, only name it. For greetings or small talk, use a short label like "Greeting". + +Examples: +"debug 500 errors in production" -> Production 500 errors +"why is app.js failing" -> app.js failure +"what does the retry logic in fetch.ts do" -> fetch.ts retry logic +"@src/credential.ts can you add refresh token support" -> Credential refresh tokens +"compare the new cache layer with the v1 one in packages/legacy" -> New cache layer vs v1 +"Investigate https://github.com/acme/app/issues/412, don't make edits" -> Issue #412 +"hey" -> Greeting` const PROMPT_SUMMARY = `Summarize what was done in this conversation. Write like a pull request description. diff --git a/packages/core/src/plugin/host.ts b/packages/core/src/plugin/host.ts index 56f6afde3363..1107dd665bb2 100644 --- a/packages/core/src/plugin/host.ts +++ b/packages/core/src/plugin/host.ts @@ -146,8 +146,8 @@ export const make = Effect.fn("PluginHost.make")(function* ( transform: (callback) => agents.transform((editor) => { callback({ - list: () => mutable(editor.list()), - get: (id) => mutable(editor.get(Agent.ID.make(id))), + list: editor.list, + get: (id) => editor.get(Agent.ID.make(id)), default: (id) => editor.default(id === undefined ? undefined : Agent.ID.make(id)), update: (id, update) => editor.update(Agent.ID.make(id), update), remove: (id) => editor.remove(Agent.ID.make(id)), @@ -368,8 +368,8 @@ export const make = Effect.fn("PluginHost.make")(function* ( transform: (callback) => integration.transform((editor) => { callback({ - list: () => mutable(editor.list()), - get: (id) => mutable(editor.get(Integration.ID.make(id))), + list: editor.list, + get: (id) => editor.get(Integration.ID.make(id)), update: (id, update) => editor.update(Integration.ID.make(id), update), remove: (id) => editor.remove(Integration.ID.make(id)), method: { diff --git a/packages/core/src/plugin/internal.ts b/packages/core/src/plugin/internal.ts index 87237d352a89..0a5fbb130d7b 100644 --- a/packages/core/src/plugin/internal.ts +++ b/packages/core/src/plugin/internal.ts @@ -57,6 +57,7 @@ import { Npm } from "@opencode/util/npm" import { Permission } from "../permission.js" import { Reference } from "../reference.js" import { WebSearch } from "../websearch.js" +import { Voice } from "../voice.js" import { Ripgrep } from "../ripgrep.js" import { Session } from "../session.js" import { SessionCompaction } from "../session/compaction.js" @@ -152,6 +153,7 @@ const services = [ SkillDiscovery.Service, Tool.Service, ToolOutput.Service, + Voice.Service, Watcher.Service, WellKnown.Service, Worktree.Service, @@ -206,6 +208,7 @@ export const requirements = LayerNode.group([ SkillDiscovery.node, Tool.node, ToolOutput.node, + Voice.node, Watcher.node, WellKnown.node, Worktree.node, diff --git a/packages/core/src/plugin/provider/cloudflare-workers-ai.ts b/packages/core/src/plugin/provider/cloudflare-workers-ai.ts index c66d0b9270b5..668584df7eb3 100644 --- a/packages/core/src/plugin/provider/cloudflare-workers-ai.ts +++ b/packages/core/src/plugin/provider/cloudflare-workers-ai.ts @@ -2,7 +2,6 @@ import { Effect } from "effect" import { define } from "@opencode/plugin/effect/plugin" import { Form } from "@opencode/schema/form" import { Provider } from "../../provider.js" -import { iife } from "../../util/iife.js" import { configuredSettings } from "./configured.js" const providerID = Provider.ID.make("cloudflare-workers-ai") @@ -11,18 +10,18 @@ export const CloudflareWorkersAIPlugin = define({ id: "opencode.provider.cloudflare.workers.ai", effect: Effect.fn(function* (ctx) { const configured = yield* configuredSettings(providerID) - const form = iife(() => { - if (typeof configured?.baseURL === "string" || resolveAccountId(configured ?? {})) return - return Form.Fields.make([ - { - type: "string", - key: "accountId", - title: "Enter your Cloudflare Account ID", - placeholder: "e.g. 1234567890abcdef1234567890abcdef", - required: true, - }, - ]) - }) + const form = + typeof configured?.baseURL === "string" || resolveAccountId(configured ?? {}) + ? undefined + : Form.Fields.make([ + { + type: "string", + key: "accountId", + title: "Enter your Cloudflare Account ID", + placeholder: "e.g. 1234567890abcdef1234567890abcdef", + required: true, + }, + ]) yield* ctx.integration.transform((editor) => { editor.method.update({ integrationID: providerID, diff --git a/packages/core/src/plugin/provider/github-copilot.ts b/packages/core/src/plugin/provider/github-copilot.ts index 19de498c5e81..4b449d024f56 100644 --- a/packages/core/src/plugin/provider/github-copilot.ts +++ b/packages/core/src/plugin/provider/github-copilot.ts @@ -1,5 +1,6 @@ import type { IntegrationOAuthMethodRegistration } from "@opencode/plugin/effect/integration" import type { SessionRequestKind, SessionTitle } from "@opencode/plugin/effect/session" +import { isRecord } from "@opencode/ai/utils/record" import { Effect, Option, Schema, Semaphore, Stream } from "effect" import { IntegrationConnection } from "../../integration/connection.js" import { Credential } from "../../credential.js" @@ -519,16 +520,16 @@ export function interactionType(kind: SessionRequestKind, child: boolean) { type RequestMetadata = ReturnType function requestMetadata(url: string, body: unknown) { - if (!record(body)) return { agent: false, vision: false } + if (!isRecord(body)) return { agent: false, vision: false } if (Array.isArray(body.input)) { const last = body.input.at(-1) return { - agent: !record(last) || last.role !== "user", + agent: !isRecord(last) || last.role !== "user", vision: body.input.some( (item) => - record(item) && + isRecord(item) && Array.isArray(item.content) && - item.content.some((part) => record(part) && part.type === "input_image"), + item.content.some((part) => isRecord(part) && part.type === "input_image"), ), } } @@ -536,35 +537,31 @@ function requestMetadata(url: string, body: unknown) { const last = body.messages.at(-1) if (url.includes("completions")) { return { - agent: !record(last) || last.role !== "user", + agent: !isRecord(last) || last.role !== "user", vision: body.messages.some( (message) => - record(message) && + isRecord(message) && Array.isArray(message.content) && - message.content.some((part) => record(part) && part.type === "image_url"), + message.content.some((part) => isRecord(part) && part.type === "image_url"), ), } } - const content = record(last) && Array.isArray(last.content) ? last.content : [] + const content = isRecord(last) && Array.isArray(last.content) ? last.content : [] return { agent: - !record(last) || last.role !== "user" || !content.some((part) => record(part) && part.type !== "tool_result"), + !isRecord(last) || last.role !== "user" || !content.some((part) => isRecord(part) && part.type !== "tool_result"), vision: body.messages.some( (message) => - record(message) && + isRecord(message) && Array.isArray(message.content) && message.content.some( (part) => - record(part) && + isRecord(part) && (part.type === "image" || (part.type === "tool_result" && Array.isArray(part.content) && - part.content.some((nested) => record(nested) && nested.type === "image"))), + part.content.some((nested) => isRecord(nested) && nested.type === "image"))), ), ), } } - -function record(input: unknown): input is Record { - return typeof input === "object" && input !== null && !Array.isArray(input) -} diff --git a/packages/core/src/plugin/provider/opencode.ts b/packages/core/src/plugin/provider/opencode.ts index 48f47caa3ad1..329d9c3c6daf 100644 --- a/packages/core/src/plugin/provider/opencode.ts +++ b/packages/core/src/plugin/provider/opencode.ts @@ -1,4 +1,16 @@ -import { Duration, Effect, Equal, FiberHandle, Option, Schema, SchemaGetter, Scope, Semaphore, Stream } from "effect" +import { + Clock, + Duration, + Effect, + Equal, + FiberHandle, + Option, + Schema, + SchemaGetter, + Scope, + Semaphore, + Stream, +} from "effect" import type { IntegrationOAuthMethodRegistration } from "@opencode/plugin/effect/integration" import { define } from "@opencode/plugin/effect/plugin" import type { SessionHttpResponse } from "@opencode/plugin/effect/session" @@ -12,6 +24,7 @@ import { ManagedPolicy } from "../../managed-policy.js" import { Model } from "../../model.js" import { Provider } from "../../provider.js" import { Variant } from "../../variant.js" +import { Voice } from "../../voice.js" import { WebSearch } from "../../websearch.js" import { ConfigPolicy } from "@opencode/schema/config/policy" import { ConfigProvider } from "@opencode/schema/config/provider" @@ -52,6 +65,13 @@ const RemoteResponse = Schema.Struct({ Schema.Struct({ ...Mcp.RemoteConfig.fields, auth: Schema.Literal("console").pipe(Schema.optional) }), ), }).pipe(Schema.optional), + // Hosted voice models the Console starts for the caller's workspace, the inference base that creates them, and the + // workspace's display name, which older Consoles omit. + voice: Schema.Struct({ + baseURL: Schema.String, + models: Schema.Array(Schema.String), + workspace: Schema.String.pipe(Schema.optional), + }).pipe(Schema.optional), // Organization policy compiled for the authenticated caller; omitted when there is none. // `experimental.policies` remains accepted alongside top-level `policies`. policies: RemotePolicies, @@ -81,6 +101,10 @@ const SsoRequired = Schema.Struct({ connectionId: Schema.String, }) const InvalidGrant = Schema.Struct({ error: Schema.Literal("invalid_grant") }) +const LiveSession = Schema.fromJsonString( + Schema.Struct({ session: Schema.Struct({ id: Schema.String }), transport: Schema.Struct({ sdp: Schema.String }) }), +) +const InferenceError = Schema.fromJsonString(Schema.Struct({ error: Schema.Struct({ message: Schema.String }) })) // Console's 401 when the session behind the access token was revoked or expired. const Unauthorized = Schema.Struct({ _tag: Schema.Literal("Unauthorized") }) const signedOutMessage = "Reconnect OpenCode Console to continue" @@ -182,15 +206,19 @@ function oauth(http: HttpClient.HttpClient) { } satisfies IntegrationOAuthMethodRegistration } -export const OpencodePlugin = define({ +export const OpencodePlugin = define< + HttpClient.HttpClient | Bus.Service | ManagedPolicy.Service | Voice.Service | Scope.Scope +>({ id: "opencode.provider.opencode", effect: Effect.fn(function* (ctx) { + const disableFree = ["1", "true"].includes(process.env.OPENCODE_DISABLE_FREE_MODELS?.toLowerCase() ?? "") const bus = yield* Bus.Service const client = yield* HttpClient.HttpClient // Every request here goes to the Console, which reads the User-Agent to tell which OpenCode a member runs // and whether it evaluates the policies it is being sent. const http = HttpClient.mapRequest(client, HttpClientRequest.setHeader("User-Agent", App.useragent(ctx.app))) const managed = yield* ManagedPolicy.Service + const voice = yield* Voice.Service const scope = yield* Scope.Scope const loading = Semaphore.makeUnsafe(1) type ActiveConnection = Effect.Success> @@ -206,8 +234,11 @@ export const OpencodePlugin = define>[] = [] + // When the last Console config load started, so rejected inference can reuse a recent status. + let checked = Number.NEGATIVE_INFINITY const load = Effect.fn("OpencodePlugin.load")(function* () { + checked = yield* Clock.currentTimeMillis const connection = yield* ctx.integration.connection.active("opencode") if (!connection) { reported = undefined @@ -246,12 +277,105 @@ export const OpencodePlugin = define) { + const active = yield* ctx.integration.connection.active("opencode") + if (!active || IntegrationConnection.key(active) !== IntegrationConnection.key(connection)) + return yield* Effect.fail(new Error("OpenCode Console connection changed")) + const credential = yield* ctx.integration.connection.resolve(active) + if (!credential || credential.type === "external") + return yield* Effect.fail(new Error("OpenCode Console is not connected")) + return credential + }) + + // Hosted GPT-Live: the server's apps start sessions on the Console's account, billed to the workspace, with + // the credential of the connection whose config offered it. + const consoleVoice = ( + connection: NonNullable, + offer: NonNullable, + ): Voice.Host => { + const send = (path: string, body: unknown) => + Effect.gen(function* () { + const credential = yield* resolveActive(connection) + const response = yield* HttpClient.withScope(http).execute( + HttpClientRequest.post(`${offer.baseURL}${path}`).pipe( + HttpClientRequest.acceptJson, + HttpClientRequest.setHeaders(credentialHeaders(credential)), + HttpClientRequest.bodyJsonUnsafe(body), + ), + ) + return { status: response.status, text: yield* response.text } + }).pipe( + Effect.provideService(FetchHttpClient.RequestInit, { redirect: "error" }), + Effect.scoped, + Effect.timeoutOrElse({ + duration: Duration.seconds(20), + orElse: () => Effect.fail(new Error("OpenCode Console didn't answer in time")), + }), + Effect.mapError( + (cause) => + new Voice.UnavailableError({ + service: "opencode", + message: cause instanceof Error ? cause.message : String(cause), + }), + ), + Effect.flatMap((reply): Effect.Effect => { + if (reply.status >= 200 && reply.status < 300) return Effect.succeed(reply.text) + if (reply.status === 401) + return Effect.fail(new Voice.UnavailableError({ service: "opencode", message: `${signedOutMessage}.` })) + const message = + Option.getOrUndefined(Schema.decodeUnknownOption(InferenceError)(reply.text))?.error.message ?? + `OpenCode Console refused the voice request (HTTP ${reply.status})` + return Effect.fail( + reply.status === 400 || reply.status === 404 + ? new Voice.InvalidRequestError({ message }) + : new Voice.UnavailableError({ service: "opencode", message }), + ) + }), + ) + return { + integration: "opencode", + models: offer.models, + workspace: offer.workspace, + session: (input) => + send("/live/sessions", { session: { ...input.session, model: input.model }, transport: input.transport }).pipe( + Effect.flatMap((text) => + Schema.decodeUnknownEffect(LiveSession)(text).pipe( + Effect.mapError( + () => + new Voice.UnavailableError({ + service: "opencode", + message: "OpenCode Console returned an invalid voice session", + }), + ), + ), + ), + Effect.map((created) => ({ + transport: { type: "webrtc" as const, sdp: created.transport.sdp }, + id: created.session.id, + })), + ), + instruct: (input) => + send(`/live/sessions/${encodeURIComponent(input.id)}/instructions`, { + instructions: input.instructions, + }).pipe(Effect.asVoid), + } + } + + // Statements and hosted voice ride on the snapshot, so a credential switch, disconnect, or 404 replaces them too. const publish = (next: typeof snapshot) => - managed.set({ - statements: [...(next.config?.experimental?.policies ?? []), ...(next.config?.policies ?? [])], - organization: next.organization, - }) + Effect.all( + [ + managed.set({ + statements: [...(next.config?.experimental?.policies ?? []), ...(next.config?.policies ?? [])], + organization: next.organization, + }), + voice.host( + next.connection && next.config?.voice ? consoleVoice(next.connection, next.config.voice) : undefined, + ), + ], + { discard: true }, + ) yield* ctx.integration.transform((editor) => { editor.update("opencode", (integration) => { @@ -351,7 +475,7 @@ export const OpencodePlugin = define { - if (!hasKey) { + if (!hasKey && !disableFree) { provider.activation = "enabled" provider.settings = { ...provider.settings, apiKey: "public" } } @@ -383,18 +507,7 @@ export const OpencodePlugin = define Effect.gen(function* () { - const active = yield* ctx.integration.connection.active("opencode") - if ( - !active || - (connection.type === "credential" - ? active.type !== "credential" || active.id !== connection.id - : active.type !== "env" || active.name !== connection.name) - ) { - return yield* Effect.fail(new Error("OpenCode Console connection changed")) - } - const credential = yield* ctx.integration.connection.resolve(active) - if (!credential || credential.type === "external") - return yield* Effect.fail(new Error("OpenCode Console is not connected")) + const credential = yield* resolveActive(connection) const metadata = credential.metadata const orgID = typeof metadata?.orgID === "string" ? metadata.orgID : undefined const token = credential.type === "oauth" ? credential.access : credential.key @@ -454,17 +567,24 @@ export const OpencodePlugin = define loading.withPermit(load().pipe(Effect.andThen(apply))) // Annotated: `apply` re-registers the hooks that call `check`, which would otherwise make inference circular. - const check = (): Effect.Effect => + // With `maxAge`, a load that started within it is reused; checking inside the permit lets concurrent callers + // share the load that was in flight when they arrived. + const check = (maxAge = Duration.zero): Effect.Effect => loading.withPermit( - load().pipe(Effect.flatMap((next) => (Equal.equals(snapshot, next) ? Effect.void : apply(next)))), + Effect.gen(function* () { + if ((yield* Clock.currentTimeMillis) - checked < Duration.toMillis(maxAge)) return + const next = yield* load() + if (!Equal.equals(snapshot, next)) yield* apply(next) + }), ) // Console inference rejects a missing SSO sign-in with a generic 401/403. Re-check the Console config, which // reports the reason, and replace the response with instructions when the connection needs the user. + // A client retrying rejected inference would otherwise send one config request per failed response. const explain = (event: SessionHttpResponse) => Effect.gen(function* () { if (event.response.status !== 401 && event.response.status !== 403) return - yield* check() + yield* check(Duration.seconds(30)) const status = reported if (!status) return void event.response.body?.cancel() @@ -579,7 +699,10 @@ function credentialHeaders(value: Credential.Key | Credential.OAuth): Record(body: Readonly> | undefined) { diff --git a/packages/core/src/project.ts b/packages/core/src/project.ts index 2fe4d6d99414..1b984003c2d9 100644 --- a/packages/core/src/project.ts +++ b/packages/core/src/project.ts @@ -145,14 +145,14 @@ const layer = Layer.effect( const changed = replaced ?? unarchived if (changed) yield* bus.publish(ProjectSchema.Event.Updated, fromRow(changed)) if (!project.vcs) return project - const directories: Array<{ projectID: ID; directory: AbsolutePath; strategy?: string }> = [ + const directories: Array<{ projectID: ID; directory: AbsolutePath; strategy?: Worktree.StrategyID }> = [ { projectID: project.id, directory: project.canonical }, ] if (project.directory !== project.canonical) directories.push({ projectID: project.id, directory: project.directory, - strategy: project.vcs.type === "git" ? "git" : undefined, + strategy: project.vcs.type === "git" ? Worktree.StrategyID.make("git") : undefined, }) // A missing directory row means this directory's resolution is a new durable // fact. The row insert commits atomically with the event, so a crash between @@ -340,14 +340,13 @@ const layer = Layer.effect( const hgDiscover = Effect.fnUntraced(function* (dotHg: AbsolutePath) { const worktree = AbsolutePath.make(path.dirname(dotHg)) - const store = AbsolutePath.make(dotHg) - const previous = yield* cached(store) + const previous = yield* cached(dotHg) const id = previous ?? (yield* hgRoot(worktree)) return { previous, id: id ?? ID.global, directory: worktree, - vcs: { type: "hg" as const, store }, + vcs: { type: "hg" as const, store: dotHg }, } }) diff --git a/packages/core/src/provider.ts b/packages/core/src/provider.ts index 11154e7701d5..b4f88d649d1b 100644 --- a/packages/core/src/provider.ts +++ b/packages/core/src/provider.ts @@ -234,16 +234,21 @@ export function mergeOverlay( ): Record | undefined { if (base === undefined) return overlay && decodeJsonRecord({ ...overlay }) if (overlay === undefined) return decodeJsonRecord({ ...base }) - return decodeJsonRecord( - Object.fromEntries( - new Set([...Object.keys(base), ...Object.keys(overlay)]).values().map((key): [string, unknown] => { - const left = base[key] - const right = overlay[key] - if (right === undefined) return [key, left] - if (isRecord(left) && isRecord(right)) return [key, mergeOverlay(left, right) ?? {}] - return [key, right] - }), - ), + return decodeJsonRecord(mergeRecord(base, overlay)) +} + +function mergeRecord( + base: Readonly>, + overlay: Readonly>, +): Record { + return Object.fromEntries( + new Set([...Object.keys(base), ...Object.keys(overlay)]).values().map((key): [string, unknown] => { + const left = base[key] + const right = overlay[key] + if (right === undefined) return [key, left] + if (isRecord(left) && isRecord(right)) return [key, mergeRecord(left, right)] + return [key, right] + }), ) } diff --git a/packages/core/src/pty.ts b/packages/core/src/pty.ts index c5d7015e0594..6d4ac6d628fc 100644 --- a/packages/core/src/pty.ts +++ b/packages/core/src/pty.ts @@ -42,11 +42,11 @@ export type Info = Types.DeepMutable export const CreateInput = Pty.CreateInput -export type CreateInput = Types.DeepMutable +export type CreateInput = typeof CreateInput.Type export const UpdateInput = Pty.UpdateInput -export type UpdateInput = Types.DeepMutable +export type UpdateInput = typeof UpdateInput.Type export { Event } from "@opencode/schema/pty" @@ -275,7 +275,7 @@ const layer = Layer.effect( const from = input.cursor === -1 ? end - : typeof input.cursor === "number" && Number.isSafeInteger(input.cursor) + : input.cursor !== undefined && Number.isSafeInteger(input.cursor) ? Math.max(0, input.cursor) : 0 const replay = (() => { diff --git a/packages/core/src/reference/instructions.ts b/packages/core/src/reference/instructions.ts index 36d40887339f..5fdbaa161f52 100644 --- a/packages/core/src/reference/instructions.ts +++ b/packages/core/src/reference/instructions.ts @@ -35,8 +35,8 @@ const update = (previous: ReadonlyArray, current: ReadonlyA (reference) => reference.name, (before, after) => before.path !== after.path || before.description !== after.description, ) - // Additions and removals render as small deltas; anything else restates the full list. - if (diff.changed.length > 0 || (diff.added.length === 0 && diff.removed.length === 0)) + // Added, changed, and removed references render as deltas; the full list is restated only when no entry differs. + if (diff.added.length === 0 && diff.changed.length === 0 && diff.removed.length === 0) return [ "The available project references have changed. This list supersedes the previous reference list.", render(current), @@ -45,6 +45,12 @@ const update = (previous: ReadonlyArray, current: ReadonlyA ...(diff.added.length === 0 ? [] : ["New project references are available in addition to those previously listed:", ...entries(diff.added)]), + ...(diff.changed.length === 0 + ? [] + : [ + "The following project references have changed. Each entry replaces the previously listed reference with the same name:", + ...entries(diff.changed.map((change) => change.current)), + ]), ...(diff.removed.length === 0 ? [] : [ diff --git a/packages/core/src/rpc.ts b/packages/core/src/rpc.ts index 4e4c054ecd86..41abed377462 100644 --- a/packages/core/src/rpc.ts +++ b/packages/core/src/rpc.ts @@ -97,7 +97,7 @@ const layer = Layer.effect( const data = (yield* encode(event.schema, args[1])) as Readonly> return yield* bus .publish(registered.definition, data, { - location: Location.Ref.make({ directory: ref.directory, workspaceID: ref.workspaceID }), + location: ref, }) .pipe(Effect.asVoid) }), @@ -284,6 +284,6 @@ const logicalEvent = Effect.fn("Rpc.logicalEvent")(function* < ...payload, type: eventType(definition, name), data, - location: Location.Ref.make({ directory: ref.directory, workspaceID: ref.workspaceID }), + location: ref, } as Rpc.EventPayload }) diff --git a/packages/core/src/session.ts b/packages/core/src/session.ts index 46a6eb0b8e3c..cb21826e1c21 100644 --- a/packages/core/src/session.ts +++ b/packages/core/src/session.ts @@ -1,7 +1,7 @@ export * as Session from "./session.js" export * from "./session/schema.js" -import { DateTime, Effect, Layer, Schema, Context, Stream } from "effect" +import { DateTime, Effect, Layer, Context, Stream } from "effect" import { LLMClient } from "@opencode/ai" import { ListAnchor, ListCursor, ListFilter } from "@opencode/schema/session" import { and, desc, eq } from "drizzle-orm" @@ -105,6 +105,7 @@ export { AttachmentError, BusyError, CompactionConflictError, + ForkEmptyError, InboxConflictError, InvalidListCursorError, MessageDecodeError, @@ -267,7 +268,8 @@ const layer = Layer.effect( const environments = yield* SessionEnvironment.Service const locations = yield* LocationServiceMap.Service const sessions = yield* Session.make() - const isDurableSessionEvent = Schema.is(SessionEvent.Durable) + const isDurableSessionEvent = (item: Bus.LogItem): item is SessionEvent.DurableEvent => + Object.hasOwn(SessionEvent.Durable.cases, item.type) const result = Service.of({ create: Effect.fn("Session.create")(function* (input) { @@ -297,13 +299,7 @@ const layer = Layer.effect( // location, so host policies that read them treat the family uniformly. metadata: input.metadata ?? parent?.metadata, permissions: input.permissions ?? parent?.permissions, - model: input.model - ? { - id: Model.ID.make(input.model.id), - providerID: input.model.providerID, - variant: input.model.variant, - } - : undefined, + model: input.model, }, { location }, ) diff --git a/packages/core/src/session/compaction.ts b/packages/core/src/session/compaction.ts index 5053c24848cb..eb9648031d72 100644 --- a/packages/core/src/session/compaction.ts +++ b/packages/core/src/session/compaction.ts @@ -101,6 +101,8 @@ const RESERVE_MIN = 16_000 const RESERVE_SHARE = 0.15 /** A common window size, assumed for the compaction request when the model's window is unknown. */ const UNKNOWN_WINDOW = 200_000 +/** A summary gets at least this much output, or the room left if less, even when the estimate margin leaves less. */ +const SUMMARY_OUTPUT_FLOOR = 32_000 const IMAGE_TOKEN_ESTIMATE = 1_500 const PDF_TOKEN_ESTIMATE = 2_000 @@ -885,15 +887,20 @@ const hasMeasuredPrompt = (message: SessionMessage.Info, model: SessionContext.L * cannot shrink the conversation further. Lower the limit to the room the request being sent leaves in the window, * so the input and the requested output never exceed the window by our own estimate. * - * Like `budget` in `prepare`, the estimate counts as measured, without the margin primary requests add for estimates: - * `deliver` aims at about 90% of the window, so that margin would cut every summary near its target to the minimum. + * The limit leaves the same margin for a low estimate that primary requests do. `deliver` aims at about 90% of the + * window, so that margin alone would cut a summary near its target to the minimum; there it still gets up to + * `SUMMARY_OUTPUT_FLOOR` of the room left without the margin. */ const fitOutput = (request: LLMRequest, limit: SessionContext.Loaded["model"]["limit"]) => { const current = request.generation?.maxTokens if (current === undefined) return request + const estimated = estimateRequest(request) const maxTokens = Math.min( current, - SessionModelRequest.outputLimit(limit, "compaction", { measured: estimateRequest(request), estimated: 0 }), + Math.max( + SessionModelRequest.outputLimit(limit, { measured: 0, estimated }), + Math.min(SUMMARY_OUTPUT_FLOOR, SessionModelRequest.outputLimit(limit, { measured: estimated, estimated: 0 })), + ), ) if (maxTokens === current) return request return LLMRequest.update(request, { generation: { ...request.generation, maxTokens } }) diff --git a/packages/core/src/session/context.ts b/packages/core/src/session/context.ts index c23f6265272a..42ae72d053af 100644 --- a/packages/core/src/session/context.ts +++ b/packages/core/src/session/context.ts @@ -95,21 +95,27 @@ const layer = Layer.effect( const agent = yield* agents.get(Agent.ID.make("title")) if (!agent) return const primary = yield* resolveModel(session).pipe(Effect.orElseSucceed(() => undefined)) - const info = yield* Effect.gen(function* () { - if (agent.model) return yield* model.get(agent.model.providerID, agent.model.id) + const candidate = yield* Effect.gen(function* () { + if (agent.model) { + const info = yield* model.get(agent.model.providerID, agent.model.id) + return info && { info, variant: agent.model.variant ?? minimalVariant(info) } + } if (!primary) return - return yield* model.small(primary.ref.providerID) + const small = yield* model.small(primary.ref.providerID) + if (small) return { info: small, variant: minimalVariant(small) } + // Without a lower variant, the primary model at its selected variant is already the fallback. + const info = yield* model.get(primary.ref.providerID, primary.ref.id) + const variant = info && lowerVariant(info, primary.ref.variant) + return variant && { info, variant } }) - const variant = - agent.model?.variant ?? MINIMAL_REASONING_VARIANTS.find((id) => info?.variants.some((item) => item.id === id)) const preferred = - info && + candidate && (yield* resolveModel({ ...session, model: Model.Ref.make({ - providerID: info.providerID, - id: info.id, - ...(variant ? { variant } : {}), + providerID: candidate.info.providerID, + id: candidate.info.id, + ...(candidate.variant ? { variant: candidate.variant } : {}), }), }).pipe(Effect.orElseSucceed(() => undefined))) const selected = preferred ?? primary @@ -184,6 +190,18 @@ const layer = Layer.effect( /** Variant IDs that minimize reasoning output, in preference order. */ const MINIMAL_REASONING_VARIANTS = ["none", "minimal", "low"].map((id) => Model.VariantID.make(id)) +const minimalVariant = (info: Model.Info) => + MINIMAL_REASONING_VARIANTS.find((id) => info.variants.some((item) => item.id === id)) + +/** + * `none` and `minimal` always reduce reasoning. `low` can enable reasoning on a model whose default + * does not reason, so it is only used when it ranks below the variant the session selected. + */ +const lowerVariant = (info: Model.Info, selected: Model.VariantID | undefined) => { + const rank = (id: Model.VariantID | undefined) => info.variants.findIndex((item) => item.id === id) + return MINIMAL_REASONING_VARIANTS.find((id) => rank(id) !== -1 && (id !== "low" || rank(id) < rank(selected))) +} + export const node = makeLocationNode({ service: Service, layer, diff --git a/packages/core/src/session/execution.ts b/packages/core/src/session/execution.ts index eeed241ae14e..fb7a9d3a6443 100644 --- a/packages/core/src/session/execution.ts +++ b/packages/core/src/session/execution.ts @@ -27,15 +27,12 @@ export interface Interface { * Interrupt active work owned by this process. Idle interruption is a no-op. Resolves once * the interruption is accepted; cleanup settles asynchronously in the execution fiber. * Returns whether an active execution was interrupted. Compose with `awaitIdle` when - * settlement matters. `awaitSettlement` waits only for the interrupted execution, - * rather than fresh work admitted during its cleanup. + * settlement matters. */ readonly interrupt: ( sessionID: SessionSchema.ID, options?: { readonly resume?: boolean - readonly reason?: "user" | "inactivity" - readonly awaitSettlement?: boolean }, ) => Effect.Effect /** Resolves once this process owns no active execution for the Session. Returns immediately when idle and never starts work. */ @@ -45,7 +42,7 @@ export interface Interface { /** Routes execution from a Session ID to its selected instance's runner. */ export class Service extends Context.Service()("@opencode/SessionExecution") {} -type InterruptReason = "user" | "shutdown" | "inactivity" +type InterruptReason = "user" | "shutdown" export function terminal(exit: Exit.Exit, reason?: InterruptReason) { if (Exit.isSuccess(exit)) return { type: "succeeded" as const } @@ -152,7 +149,7 @@ export const layer = Layer.effect( isActive: coordinator.isActive, interrupt: (sessionID, options) => Effect.gen(function* () { - const interrupted = yield* coordinator.interrupt(sessionID, options?.reason ?? "user", options) + const interrupted = yield* coordinator.interrupt(sessionID, "user") if (!options?.resume) return interrupted // Resume steering input and between-turn control work from the interrupted // intent. Queued next-turn prompts stay parked: a steer-scoped drain never diff --git a/packages/core/src/session/history.ts b/packages/core/src/session/history.ts index 74159a27b2bb..e75bc39a3469 100644 --- a/packages/core/src/session/history.ts +++ b/packages/core/src/session/history.ts @@ -6,7 +6,7 @@ import { SessionLegacy } from "@opencode/schema/session-legacy" import { Database } from "../database/database.js" import { MessageDecodeError } from "./error.js" import { SessionMessage } from "./message.js" -import { SessionSchema } from "./schema.js" +import type { SessionSchema } from "./schema.js" import { Instructions } from "../instructions/index.js" import { InstructionState } from "./instruction-state.js" import { SessionRecentText } from "./recent-text.js" @@ -32,7 +32,8 @@ export const isNativeCompaction = ( /** A native compaction's window: canonical AI messages, stored as JSON. */ export const NativeWindow = Schema.toCodecJson(Schema.Array(Message)) -const decodesWindow = (messages: unknown) => Option.isSome(Schema.decodeUnknownOption(NativeWindow)(messages)) +const decodeNativeWindow = Schema.decodeUnknownOption(NativeWindow) +const decodesWindow = (messages: unknown) => Option.isSome(decodeNativeWindow(messages)) // Like other native data, a window only replays on the model that made it. One that no longer decodes, for example // a malformed one from a compaction hook, is skipped the same way instead of failing the session. @@ -44,7 +45,7 @@ const replayable = (message: SessionMessage.Info, boundary: Boundary) => message.model.id === boundary.id && decodesWindow(message.native.messages)) -export const latestCompaction = Effect.fnUntraced(function* ( +const latestCompaction = Effect.fnUntraced(function* ( db: DatabaseService, sessionID: SessionSchema.ID, boundary: Boundary, @@ -93,8 +94,8 @@ export const decodeMessageRow = (row: typeof SessionMessageTable.$inferSelect) = Effect.mapError( () => new MessageDecodeError({ - sessionID: SessionSchema.ID.make(row.session_id), - messageID: SessionMessage.ID.make(row.id), + sessionID: row.session_id, + messageID: row.id, }), ), ) @@ -154,13 +155,11 @@ const withRecentText = Effect.fnUntraced(function* ( const compaction = first?.message if (compaction?.type !== "compaction" || compaction.status !== "completed") return entries const keepID = compaction.metadata?.keepID - if (typeof keepID !== "string") return entries + if (!Schema.is(SessionMessage.ID)(keepID)) return entries const start = yield* db .select({ seq: SessionMessageTable.seq }) .from(SessionMessageTable) - .where( - and(eq(SessionMessageTable.session_id, sessionID), eq(SessionMessageTable.id, SessionMessage.ID.make(keepID))), - ) + .where(and(eq(SessionMessageTable.session_id, sessionID), eq(SessionMessageTable.id, keepID))) .get() .pipe(Effect.orDie) if (!start) return entries diff --git a/packages/core/src/session/inbox.ts b/packages/core/src/session/inbox.ts index 052a15525a76..31cd11f95d0e 100644 --- a/packages/core/src/session/inbox.ts +++ b/packages/core/src/session/inbox.ts @@ -75,8 +75,8 @@ function matches( const fromRow = (row: typeof SessionInboxTable.$inferSelect): Info => { const base = { - id: SessionMessage.ID.make(row.id), - sessionID: SessionSchema.ID.make(row.session_id), + id: row.id, + sessionID: row.session_id, time: { created: DateTime.makeUnsafe(row.time_created) }, } if (row.type === "compaction") @@ -131,15 +131,24 @@ const promotedFromMessage = Effect.fn("SessionInbox.promotedFromMessage")(functi return User.make({ ...base, type: "user", - payload: decodeUser(message), + payload: UserPayload.make({ + text: message.text, + files: message.files, + agents: message.agents, + skills: message.skills, + metadata: message.metadata, + }), }) if (message.type === "synthetic") return Synthetic.make({ ...base, type: "synthetic", - payload: decodeSynthetic(message), + payload: SyntheticPayload.make({ + text: message.text, + description: message.description, + metadata: message.metadata, + }), }) - return yield* new LifecycleConflict({ id }) }) export type Interface = Effect.Success> @@ -221,7 +230,7 @@ export const make = Effect.fn("SessionInbox.make")(function* () { return yield* admit({ id: input.id, sessionID: input.sessionID, - item: { type: "compaction", payload: {}, delivery: Delivery.make(input.delivery) }, + item: { type: "compaction", payload: {}, delivery: input.delivery }, }) }), ) @@ -464,17 +473,17 @@ const publish = Effect.fn("SessionInbox.publish")(function* ( yield* Effect.forEach( rows, (row) => { - const entry = fromRow(row) - if (entry.type === "compaction") return Effect.die(new LifecycleConflict({ id: entry.id })) + if (row.type !== "move" && row.type !== "user" && row.type !== "synthetic") + return Effect.die(new LifecycleConflict({ id: row.id })) return bus .publish(SessionEvent.InboxDelivered, { sessionID, - inboxID: entry.id, + inboxID: row.id, }) .pipe( Effect.catchDefect((defect) => defect instanceof LifecycleConflict - ? promotedFromMessage(db, sessionID, entry.id, entry.delivery).pipe( + ? promotedFromMessage(db, sessionID, row.id, row.delivery).pipe( Effect.flatMap((stored) => (stored !== undefined ? Effect.void : Effect.die(defect))), Effect.orDie, ) diff --git a/packages/core/src/session/info.ts b/packages/core/src/session/info.ts index 88ff26f5e42b..ab8a78c6c004 100644 --- a/packages/core/src/session/info.ts +++ b/packages/core/src/session/info.ts @@ -1,11 +1,7 @@ import { DateTime, Schema } from "effect" -import { Agent } from "@opencode/schema/agent" -import { Location } from "@opencode/schema/location" import { Model } from "@opencode/schema/model" -import { Project } from "@opencode/schema/project" import { Provider } from "@opencode/schema/provider" import { AbsolutePath, RelativePath } from "../schema.js" -import { Workspace } from "@opencode/schema/workspace" import { SessionSchema } from "./schema.js" import type { SessionTable } from "./sql.js" import { PersistedRevert } from "@opencode/schema/session-revert" @@ -15,18 +11,18 @@ const decodeRevert = Schema.decodeUnknownSync(PersistedRevert) export function fromRow(row: typeof SessionTable.$inferSelect): SessionSchema.Info { return SessionSchema.Info.make({ - id: SessionSchema.ID.make(row.id), - projectID: Project.ID.make(row.project_id), + id: row.id, + projectID: row.project_id, title: row.title ?? undefined, - parentID: row.parent_id ? SessionSchema.ID.make(row.parent_id) : undefined, + parentID: row.parent_id ?? undefined, fork: row.fork_session_id && row.fork_boundary ? { - sessionID: SessionSchema.ID.make(row.fork_session_id), + sessionID: row.fork_session_id, boundary: row.fork_boundary, } : undefined, - agent: row.agent ? Agent.ID.make(row.agent) : undefined, + agent: row.agent ?? undefined, model: row.model ? { id: Model.ID.make(row.model.id), @@ -44,10 +40,10 @@ export function fromRow(row: typeof SessionTable.$inferSelect): SessionSchema.In write: row.tokens_cache_write, }, }, - location: Location.Ref.make({ + location: { directory: AbsolutePath.make(row.directory), - workspaceID: row.workspace_id ? Workspace.ID.make(row.workspace_id) : undefined, - }), + workspaceID: row.workspace_id ?? undefined, + }, subpath: row.path ? RelativePath.make(row.path) : undefined, metadata: row.metadata ?? undefined, permissions: row.permission ?? undefined, diff --git a/packages/core/src/session/message-updater.ts b/packages/core/src/session/message-updater.ts index 8913b1a87ed5..24d9ce352809 100644 --- a/packages/core/src/session/message-updater.ts +++ b/packages/core/src/session/message-updater.ts @@ -19,6 +19,8 @@ export interface Adapter { type DraftAssistant = WritableDraft +const decodeAssistantContent = Schema.decodeUnknownSync(Schema.Array(SessionMessage.AssistantContent)) + const projectTerminalSnapshot = (draft: DraftAssistant, event: SessionEvent.Step.Ended | SessionEvent.Step.Failed) => { if (event.data.snapshot || event.data.files) draft.snapshot = { @@ -85,9 +87,7 @@ export function update(adapter: Adapter, event: SessionEvent.DurableEvent) { "session.unarchived": () => Effect.void, "session.message.content.updated": (event) => updateOwnedAssistant(event.data.messageID, (draft) => { - draft.content = castDraft( - Schema.decodeUnknownSync(Schema.Array(SessionMessage.AssistantContent))(event.data.content), - ) + draft.content = castDraft(decodeAssistantContent(event.data.content)) }), "session.usage.recorded": () => Effect.void, "session.agent.selected": (event) => diff --git a/packages/core/src/session/model-request.ts b/packages/core/src/session/model-request.ts index 15c7949ade4a..4bb8e47bae5d 100644 --- a/packages/core/src/session/model-request.ts +++ b/packages/core/src/session/model-request.ts @@ -48,8 +48,6 @@ const GENERATION_KEYS = new Set(Object.keys(GenerationOptions.fields)) const OUTPUT_TOKEN_FALLBACK = 32_000 // No reply needs more, however much the model allows. const OUTPUT_TOKEN_MAX = 256_000 -// A summary never needs more, and a request asking for more cannot be shrunk to fit a window the catalog overstates. -const SUMMARY_OUTPUT_MAX = 32_000 // Prompt text is estimated at about 4 characters per token, which can run low on dense text such as code. const ESTIMATE_ERROR = 0.15 // Never ask for less; only reachable with automatic compaction off, since it keeps the window from filling this far. @@ -84,13 +82,8 @@ export interface Input { } /** The default output limit: the catalog limit, fitted to the room the prompt leaves in the context window. */ -export const outputLimit = ( - limit: Model.Info["limit"], - kind: "primary" | "compaction", - inputTokens?: Input["inputTokens"], -) => { - const model = Math.min(limit.output > 0 ? limit.output : OUTPUT_TOKEN_FALLBACK, OUTPUT_TOKEN_MAX) - const requested = kind === "compaction" ? Math.min(model, SUMMARY_OUTPUT_MAX) : model +export const outputLimit = (limit: Model.Info["limit"], inputTokens?: Input["inputTokens"]) => { + const requested = Math.min(limit.output > 0 ? limit.output : OUTPUT_TOKEN_FALLBACK, OUTPUT_TOKEN_MAX) if (inputTokens === undefined || limit.context <= 0) return requested const room = limit.context - inputTokens.measured - Math.ceil(inputTokens.estimated * (1 + ESTIMATE_ERROR)) return Math.min(requested, Math.max(OUTPUT_TOKEN_MIN, room)) @@ -258,7 +251,7 @@ export const layer = Layer.effect( messages: input.messages, options: kind === "primary" || kind === "compaction" - ? { maxTokens: outputLimit(model.limit, kind, input.inputTokens) } + ? { maxTokens: outputLimit(model.limit, input.inputTokens) } : {}, }, Object.fromEntries(Array.from(given, ([d, t]) => [t.name, d])), @@ -304,7 +297,7 @@ export const layer = Layer.effect( const baseURL = base.model.route.endpoint.baseURL const modelHook = yield* hooks.trigger("session", "model.request", { ...scope, - baseURL: typeof baseURL === "string" ? baseURL : undefined, + baseURL, headers: { ...base.http?.headers }, }) const route = @@ -345,7 +338,7 @@ export const layer = Layer.effect( ), }) return HttpClientResponse.fromWeb(sent, after.response) - }).pipe(Effect.mapError((cause) => (cause instanceof Error ? cause : new Error(String(cause))))) + }) : undefined // HTTP hooks wrap every HTTP request, including the WebSocket fallback path. The route decides // which transport actually carries the request, so both hook families are always offered. diff --git a/packages/core/src/session/projector.ts b/packages/core/src/session/projector.ts index 63ccbd79c5a0..041bfe86bde7 100644 --- a/packages/core/src/session/projector.ts +++ b/packages/core/src/session/projector.ts @@ -6,13 +6,11 @@ import path from "path" import { Database } from "../database/database.js" import { Bus } from "../bus.js" import { makeGlobalNode } from "@opencode/util/effect/app-node" -import { Agent } from "@opencode/schema/agent" import { Model } from "@opencode/schema/model" import { SessionEvent } from "./event.js" import { SessionMessage } from "./message.js" import { SessionMessageUpdater } from "./message-updater.js" import { SessionInbox } from "./inbox.js" -import { Workspace } from "@opencode/schema/workspace" import { InstructionState } from "./instruction-state.js" import { SessionInboxTable, SessionMessageTable, SessionTable } from "./sql.js" import { InstructionEntry } from "./instruction-entry.js" @@ -225,16 +223,11 @@ const projectFork = Effect.fn("SessionProjector.projectFork")(function* ( } // A summary references the first message it kept, which the fork copied under a new ID. const keepID = row.type === "compaction" ? row.data.metadata?.keepID : undefined - if (typeof keepID !== "string") return copy + if (!Schema.is(SessionMessage.ID)(keepID)) return copy const kept = yield* db .select({ seq: SessionMessageTable.seq }) .from(SessionMessageTable) - .where( - and( - eq(SessionMessageTable.session_id, event.data.parentID), - eq(SessionMessageTable.id, SessionMessage.ID.make(keepID)), - ), - ) + .where(and(eq(SessionMessageTable.session_id, event.data.parentID), eq(SessionMessageTable.id, keepID))) .get() .pipe(Effect.orDie) if (!kept) return copy @@ -257,13 +250,13 @@ function run(db: DatabaseService, event: MessageEvent) { return Effect.gen(function* () { const updateMessage = (message: SessionMessage.Info) => { const encoded = encodeMessage(message) - const { id, type, ...data } = encoded + const { id: _, type, ...data } = encoded return db .update(SessionMessageTable) .set({ type, time_created: DateTime.toEpochMillis(message.time.created), data }) .where( and( - eq(SessionMessageTable.id, SessionMessage.ID.make(id)), + eq(SessionMessageTable.id, message.id), eq(SessionMessageTable.session_id, event.data.sessionID), ), ) @@ -280,7 +273,7 @@ function run(db: DatabaseService, event: MessageEvent) { .get() .pipe( Effect.orDie, - Effect.map((row) => (row?.agent ? Agent.ID.make(row.agent) : undefined)), + Effect.map((row) => row?.agent ?? undefined), ) }, getModel() { @@ -312,7 +305,7 @@ function run(db: DatabaseService, event: MessageEvent) { ? { location: { directory: AbsolutePath.make(row.directory), - workspaceID: row.workspaceID ? Workspace.ID.make(row.workspaceID) : undefined, + workspaceID: row.workspaceID ?? undefined, }, projectID: row.projectID, subpath: row.subpath === null ? undefined : RelativePath.make(row.subpath), @@ -411,11 +404,11 @@ function run(db: DatabaseService, event: MessageEvent) { function insertMessage(db: DatabaseService, event: SessionEvent.DurableEvent, message: SessionMessage.Info) { const encoded = encodeMessage(message) - const { id, type, ...data } = encoded + const { id: _, type, ...data } = encoded return db .insert(SessionMessageTable) .values({ - id: SessionMessage.ID.make(id), + id: message.id, session_id: event.data.sessionID, type, seq: event.durable.seq, @@ -468,7 +461,7 @@ const layer = Layer.effectDiscard( .values({ id: event.data.sessionID, project_id: event.data.projectID, - workspace_id: event.data.location.workspaceID ? Workspace.ID.make(event.data.location.workspaceID) : null, + workspace_id: event.data.location.workspaceID ?? null, parent_id: event.data.parentID, slug: event.data.slug, directory: event.data.location.directory, @@ -498,7 +491,7 @@ const layer = Layer.effectDiscard( directory: event.data.location.directory, path: event.data.subpath, ...(event.data.projectID ? { project_id: event.data.projectID } : {}), - workspace_id: event.data.location.workspaceID ? Workspace.ID.make(event.data.location.workspaceID) : null, + workspace_id: event.data.location.workspaceID ?? null, time_updated: event.created, }) .where(eq(SessionTable.id, event.data.sessionID)) @@ -773,7 +766,7 @@ const layer = Layer.effectDiscard( yield* db .update(SessionTable) .set({ - revert: { ...revert, files: revert.files ? [...revert.files] : undefined }, + revert, time_updated: event.created, }) .where(eq(SessionTable.id, event.data.sessionID)) diff --git a/packages/core/src/session/prompt.ts b/packages/core/src/session/prompt.ts index 80878f354dd6..d721a7454ed1 100644 --- a/packages/core/src/session/prompt.ts +++ b/packages/core/src/session/prompt.ts @@ -1,6 +1,6 @@ export * as SessionPrompt from "./prompt.js" -import { Base64, FileAttachment, Prompt } from "@opencode/schema/prompt" +import { Base64, FileAttachment } from "@opencode/schema/prompt" import { PromptInput } from "@opencode/schema/prompt-input" import { SessionInbox } from "@opencode/schema/session-inbox" import type { Session } from "@opencode/schema/session" @@ -77,15 +77,13 @@ export const prepare = Effect.fn("SessionPrompt.prepare")(function* (request: { return { type: "user", payload: SessionInbox.UserPayload.make({ - ...Prompt.make({ - text: input.text, - agents: input.agents, - files, - skills: selected?.length ? selected : undefined, - }), + text: input.text, + agents: input.agents, + files, + skills: selected?.length ? selected : undefined, metadata: event.metadata, }), - delivery: SessionInbox.Delivery.make(event.delivery), + delivery: event.delivery, } satisfies SessionInbox.Item }).pipe(instances.provide(request.session)) }) diff --git a/packages/core/src/session/revert.ts b/packages/core/src/session/revert.ts index bb08555f3018..6bffb3c4017d 100644 --- a/packages/core/src/session/revert.ts +++ b/packages/core/src/session/revert.ts @@ -30,9 +30,7 @@ export const stage = Effect.fn("SessionRevert.stage")(function* (input: { return yield* Effect.gen(function* () { const snapshot = yield* Snapshot.Service - const original = input.session.revert?.snapshot - ? Snapshot.ID.make(input.session.revert.snapshot) - : yield* snapshot.capture() + const original = input.session.revert?.snapshot ?? (yield* snapshot.capture()) const next = yield* plan(database.db, { sessionID: input.session.id, messageID: input.messageID }) const restore = new Map() if (original) { @@ -63,7 +61,7 @@ export const clear = Effect.fn("SessionRevert.clear")(function* (session: Sessio yield* Effect.gen(function* () { const snapshot = yield* Snapshot.Service if (!session.revert) return - const original = session.revert.snapshot ? Snapshot.ID.make(session.revert.snapshot) : undefined + const original = session.revert.snapshot if (original) yield* snapshot.restore({ files: new Map((session.revert.files ?? []).map((file) => [RelativePath.make(file.file), original])), @@ -108,7 +106,7 @@ const plan = Effect.fn("SessionRevert.plan")(function* (db: Database.Interface[" const message = SessionHistory.decodeMessageRowSync(row) if (message.type !== "assistant" || !message.snapshot?.start) continue for (const file of message.snapshot.files ?? []) - if (!files.has(file)) files.set(file, Snapshot.ID.make(message.snapshot.start)) + if (!files.has(file)) files.set(file, message.snapshot.start) } return files }) diff --git a/packages/core/src/session/run-coordinator.ts b/packages/core/src/session/run-coordinator.ts index 9c4ba4b09a97..a8e75319f1f1 100644 --- a/packages/core/src/session/run-coordinator.ts +++ b/packages/core/src/session/run-coordinator.ts @@ -17,14 +17,9 @@ export interface Coordinator { * Stops the active execution and clears its doorbell. No-op when idle. Resolves once the * interruption is accepted, not when cleanup settles: the execution fiber finishes its * finalizers and settled hook on its own time. Returns whether an active execution was - * interrupted. `awaitSettlement` waits for this execution's cleanup and settled hook, - * without following fresh work admitted during cleanup. `awaitIdle` follows successors too. + * interrupted. */ - readonly interrupt: ( - key: Key, - reason?: Reason, - options?: { readonly awaitSettlement?: boolean }, - ) => Effect.Effect + readonly interrupt: (key: Key, reason?: Reason) => Effect.Effect /** Resolves once no execution is active for the key. Returns immediately when already idle and never starts work. */ readonly awaitIdle: (key: Key) => Effect.Effect } @@ -175,22 +170,5 @@ export const make = (options: { return Deferred.await(execution.done).pipe(Effect.ignoreCause, Effect.andThen(awaitIdle(key))) }) - return { - active: Effect.sync(() => new Set(executions.keys())), - isActive, - run, - wake, - interrupt: (key, reason, options) => - Effect.suspend(() => { - const execution = executions.get(key) - return interrupt(key, reason).pipe( - Effect.tap(() => - options?.awaitSettlement && execution - ? Deferred.await(execution.done).pipe(Effect.ignoreCause) - : Effect.void, - ), - ) - }), - awaitIdle, - } + return { active: Effect.sync(() => new Set(executions.keys())), isActive, run, wake, interrupt, awaitIdle } }) diff --git a/packages/core/src/session/runner/publish-llm-event.ts b/packages/core/src/session/runner/publish-llm-event.ts index 16855713fb49..2e5d9a202e01 100644 --- a/packages/core/src/session/runner/publish-llm-event.ts +++ b/packages/core/src/session/runner/publish-llm-event.ts @@ -1,8 +1,10 @@ import { type LLMEvent, type ProviderMetadata, type ToolResultValue } from "@opencode/ai" +import { isRecord } from "@opencode/ai/utils/record" import type { Agent } from "@opencode/schema/agent" import type { Model } from "@opencode/schema/model" import type { RelativePath } from "@opencode/schema/schema" import type { Snapshot } from "@opencode/schema/snapshot" +import type { TokenUsage } from "@opencode/schema/token-usage" import { Effect, Fiber, Iterable } from "effect" import { isReadonlyArrayNonEmpty } from "effect/Array" import { Bus } from "../../bus.js" @@ -15,6 +17,7 @@ import { SessionUsage } from "../usage.js" import { contentFilterError } from "../to-session-error.js" import type { Tool } from "../../tool.js" import { KiloSessionUsage } from "../../kilocode/session/usage.js" // kilocode_change(gateway) +import { stringify } from "../../tool/runtime.js" type Input = { readonly sessionID: SessionSchema.ID @@ -27,8 +30,7 @@ type Input = { readonly assistantMessageID: SessionMessage.ID } -const asRecord = (value: unknown): Record => - typeof value === "object" && value !== null && !Array.isArray(value) ? (value as Record) : { value } +const asRecord = (value: unknown): Record => (isRecord(value) ? value : { value }) const INPUT_EXCERPT_LIMIT = 2048 @@ -68,7 +70,7 @@ export interface StepRecord { readonly finish: Extract["reason"]["normalized"] readonly rawFinish?: string readonly native?: SessionMessage.Native - readonly tokens: ReturnType + readonly tokens: TokenUsage.Info readonly providerCost?: Money.USD // kilocode_change(gateway) } readonly needsContinuation: boolean @@ -77,15 +79,6 @@ export interface StepRecord { /** Derives canonical model content from a provider-hosted tool result. */ type NonEmptyContent = readonly [Tool.Content, ...Tool.Content[]] -const stringify = (value: unknown) => { - if (typeof value === "string") return value - try { - return JSON.stringify(value) ?? String(value) - } catch { - return String(value) - } -} - const hostedContent = (result: ToolResultValue): NonEmptyContent => { if (result.type === "content" && isReadonlyArrayNonEmpty(result.value)) return result.value return [{ type: "text", text: stringify(result.value) }] @@ -155,7 +148,7 @@ export const createLLMEventPublisher = (bus: Pick, inp const native = (metadata: ProviderMetadata | undefined) => metadata?.[input.providerMetadataKey] const fragments = ( name: string, - ended: (id: string, value: string, ordinal: number, native?: Record) => Effect.Effect, + ended: (id: string, value: string, ordinal: number, native?: SessionMessage.Native) => Effect.Effect, delta?: (id: string, value: string, ordinal: number) => Effect.Effect, single = false, ) => { @@ -164,11 +157,11 @@ export const createLLMEventPublisher = (bus: Pick, inp readonly values: string[] pending: string timer?: Fiber.Fiber - native?: Record + native?: SessionMessage.Native } const chunks = new Map() let nextOrdinal = 0 - const start = (id: string, native?: Record) => + const start = (id: string, native?: SessionMessage.Native) => Effect.suspend(() => { if (chunks.has(id)) return Effect.die(new Error(`Duplicate ${name} start: ${id}`)) if (single && chunks.size > 0) return Effect.die(new Error(`${name} start before end: ${id}`)) @@ -187,7 +180,7 @@ export const createLLMEventPublisher = (bus: Pick, inp yield* delta(id, value, current.ordinal) return undefined }, Effect.uninterruptible) - const append = Effect.fnUntraced(function* (id: string, value: string, native?: Record) { + const append = Effect.fnUntraced(function* (id: string, value: string, native?: SessionMessage.Native) { const current = chunks.get(id) if (!current) return yield* Effect.die(new Error(`${name} delta before start: ${id}`)) current.values.push(value) @@ -211,7 +204,7 @@ export const createLLMEventPublisher = (bus: Pick, inp } return current.ordinal }) - const end = Effect.fnUntraced(function* (id: string, native?: Record, value?: string) { + const end = Effect.fnUntraced(function* (id: string, native?: SessionMessage.Native, value?: string) { const current = chunks.get(id) if (!current) return yield* Effect.die(new Error(`${name} end before start: ${id}`)) if (current.timer) yield* Fiber.interrupt(current.timer) @@ -421,7 +414,7 @@ export const createLLMEventPublisher = (bus: Pick, inp const publishStepFailure = Effect.fnUntraced(function* (details?: { readonly cost?: Money.USD - readonly tokens?: ReturnType + readonly tokens?: TokenUsage.Info readonly snapshot?: Snapshot.ID readonly files?: readonly RelativePath[] }) { diff --git a/packages/core/src/session/runner/to-llm-message.ts b/packages/core/src/session/runner/to-llm-message.ts index bf6cd1413aaa..3bb6bae19d07 100644 --- a/packages/core/src/session/runner/to-llm-message.ts +++ b/packages/core/src/session/runner/to-llm-message.ts @@ -103,6 +103,7 @@ const userAttachmentContent = (files: readonly FileAttachment[]) => { } const decodeToolInput = Schema.decodeUnknownOption(Schema.fromJsonString(Schema.Unknown)) +const decodeNativeWindow = Schema.decodeUnknownSync(SessionHistory.NativeWindow) const providerMetadata = ( provider: string, @@ -152,8 +153,8 @@ const toolResult = (tool: SessionMessage.AssistantTool, providerMetadata: Provid } const assistant = (message: SessionMessage.Assistant, model: Model.Ref, providerMetadataKey: string) => { - const sameProvider = String(message.model.providerID) === String(model.providerID) - const sameModel = sameProvider && String(message.model.id) === String(model.id) + const sameProvider = message.model.providerID === model.providerID + const sameModel = sameProvider && message.model.id === model.id const reuseProviderMetadata = sameModel && message.error === undefined const reasoningInterrupted = message.content.some( (item) => item.type === "reasoning" && item.time !== undefined && item.time.completed === undefined, @@ -292,8 +293,7 @@ function toLLMMessage(message: SessionMessage.Info, model: Model.Ref, providerMe case "compaction": { if (message.status !== "completed") return [] // History selection only keeps native windows the target model can replay. - if (message.native?.messages !== undefined) - return [...Schema.decodeUnknownSync(SessionHistory.NativeWindow)(message.native.messages)] + if (message.native?.messages !== undefined) return [...decodeNativeWindow(message.native.messages)] const kept = SessionRecentText.kept(message) return [ Message.make({ diff --git a/packages/core/src/session/session.ts b/packages/core/src/session/session.ts index 0e44a456b4f2..06c2b241e6dd 100644 --- a/packages/core/src/session/session.ts +++ b/packages/core/src/session/session.ts @@ -303,7 +303,7 @@ export const make = Effect.fn("Session.make")(function* () { description: input.description, metadata: input.metadata, }), - delivery: SessionInbox.Delivery.make(input.delivery ?? "steer"), + delivery: input.delivery ?? "steer", } satisfies SessionInbox.Item const admitted = yield* admission .admit({ diff --git a/packages/core/src/session/sql.ts b/packages/core/src/session/sql.ts index 4465d885b33a..88e9f925600b 100644 --- a/packages/core/src/session/sql.ts +++ b/packages/core/src/session/sql.ts @@ -4,6 +4,7 @@ import { directoryColumn, pathColumn } from "../database/path.js" import { ProjectTable } from "../project/sql.js" import type { SessionMessage } from "./message.js" import type { SessionInbox } from "./inbox.js" +import type { Agent } from "@opencode/schema/agent" import type { FileDiff } from "@opencode/schema/file-diff" import type { Permission } from "@opencode/schema/permission" import type { Project } from "@opencode/schema/project" @@ -50,7 +51,7 @@ export const SessionTable = sqliteTable( tokens_cache_write: integer().notNull().default(0), revert: text({ mode: "json" }).$type(), permission: text({ mode: "json" }).$type(), - agent: text(), + agent: text().$type(), model: text({ mode: "json" }).$type<{ id: string providerID: string diff --git a/packages/core/src/session/store.ts b/packages/core/src/session/store.ts index d86cd6ab60c5..c9698231bbcf 100644 --- a/packages/core/src/session/store.ts +++ b/packages/core/src/session/store.ts @@ -189,7 +189,7 @@ const layer = Layer.effect( .pipe(Effect.orDie) return row ? { - sessionID: Session.ID.make(row.session_id), + sessionID: row.session_id, message: yield* SessionHistory.decodeMessageRow(row).pipe(Effect.orDie), } : undefined diff --git a/packages/core/src/session/title.ts b/packages/core/src/session/title.ts index 586937e211e3..aca28811bed8 100644 --- a/packages/core/src/session/title.ts +++ b/packages/core/src/session/title.ts @@ -30,6 +30,11 @@ export interface Interface { export class Service extends Context.Service()("@opencode/SessionTitle") {} const truncate = (value: string) => (value.length <= MAX_LENGTH ? value : `${value.slice(0, MAX_LENGTH - 3)}...`) +// Framing the input as text to name keeps models from answering or carrying out the request instead. +const messagePrompt = (text: string) => + `Here is the first message of the conversation:\n\n\n${text}\n\n\nReply with only the title.` +const conversationPrompt = (text: string) => + `Here is the conversation so far:\n\n\n${text}\n\n\nReply with only the title.` export const isUntitled = (session: SessionSchema.Info) => isExactRootFallback({ title: session.title, @@ -117,13 +122,14 @@ export const layer = Layer.effect( return text ? [`Assistant: ${text}`] : [] }) .join("\n\n") - if (!recent) return original + if (!recent) return conversationPrompt(original) const prefix = `${original}\n\nRecent conversation:\n` - return `${prefix}${recent.slice(-(MAX_CONTEXT_LENGTH - prefix.length))}` + const budget = MAX_CONTEXT_LENGTH - conversationPrompt("").length - prefix.length + return conversationPrompt(`${prefix}${recent.slice(-budget)}`) }), - Effect.orElseSucceed(() => firstUser.text), + Effect.orElseSucceed(() => messagePrompt(firstUser.text)), ) - : firstUser.text + : messagePrompt(firstUser.text) const selection = yield* context.selectTitle(session) if (!selection) return const title = diff --git a/packages/core/src/session/to-session-error.ts b/packages/core/src/session/to-session-error.ts index d8eb0f683afa..2fa9f70a1b34 100644 --- a/packages/core/src/session/to-session-error.ts +++ b/packages/core/src/session/to-session-error.ts @@ -25,6 +25,9 @@ const tokenSharingMessages = { "This ChatGPT connection isn't authorized for this request. Reconnect to ChatGPT or choose another connection.", } +const FREE_USAGE_LIMIT_MESSAGE = + "Free usage limit reached. Connect a provider or subscribe to OpenCode Go to keep going, or switch models." + export function toSessionError(cause: unknown): SessionError.Error { if (cause instanceof AIError) { switch (cause.reason._tag) { @@ -33,7 +36,11 @@ export function toSessionError(cause: unknown): SessionError.Error { case "Authentication": return providerError("provider.auth", cause.reason) case "QuotaExceeded": - return providerError("provider.quota", cause.reason) + if (!cause.reason.body?.includes("FreeUsageLimitError")) return providerError("provider.quota", cause.reason) + return { + ...providerError("provider.free-limit", cause.reason), + message: FREE_USAGE_LIMIT_MESSAGE, + } case "ContentPolicy": return providerError("provider.content-filter", cause.reason) case "Transport": @@ -96,10 +103,12 @@ export function contentFilterError(summary: string, reason: FinishReasonDetails) function providerError(type: string, reason: AIError["reason"]): SessionError.Error { const status = reason.http?.status + const retryAfterMs = "retryAfterMs" in reason ? reason.retryAfterMs : undefined return { type, message: Object.entries(tokenSharingMessages).find(([code]) => reason.body?.includes(code))?.[1] ?? reason.message, ...(status === undefined ? {} : { status }), ...(reason.body === undefined ? {} : { response: { body: reason.body } }), + ...(retryAfterMs === undefined ? {} : { retryAfterMs }), } } diff --git a/packages/core/src/skill.ts b/packages/core/src/skill.ts index 51cf0bc170ad..8f7909e0e2e8 100644 --- a/packages/core/src/skill.ts +++ b/packages/core/src/skill.ts @@ -104,13 +104,14 @@ const layer = Layer.effect( list: () => Array.from(editor.skills.values()), get: (id) => editor.skills.get(ID.make(id)), add: (skill) => { - editor.skills.set(skill.id, { ...skill } as Types.DeepMutable) + editor.skills.set(skill.id, { ...skill }) }, update: (id, update) => { - const current = editor.skills.get(ID.make(id)) + const key = ID.make(id) + const current = editor.skills.get(key) if (!current) return update(current) - current.id = ID.make(id) + current.id = key }, remove: (id) => { editor.skills.delete(ID.make(id)) diff --git a/packages/core/src/skill/instructions.ts b/packages/core/src/skill/instructions.ts index 14811dace4fa..e26c3f19e9d7 100644 --- a/packages/core/src/skill/instructions.ts +++ b/packages/core/src/skill/instructions.ts @@ -39,8 +39,8 @@ const update = (previous: ReadonlyArray, current: ReadonlyArray skill.id, (before, after) => before.name !== after.name || before.description !== after.description, ) - // Additions and removals render as small deltas; anything else restates the full list. - if (diff.changed.length > 0 || (diff.added.length === 0 && diff.removed.length === 0)) + // Added, changed, and removed skills render as deltas; the full list is restated only when no entry differs. + if (diff.added.length === 0 && diff.changed.length === 0 && diff.removed.length === 0) return [ "The available skills have changed. This list supersedes the previous available skills list.", render(current), @@ -49,6 +49,12 @@ const update = (previous: ReadonlyArray, current: ReadonlyArray change.current)), + ]), ...(diff.removed.length === 0 ? [] : [ diff --git a/packages/core/src/snapshot.ts b/packages/core/src/snapshot.ts index 3ac6df3e2414..e9d5785b5987 100644 --- a/packages/core/src/snapshot.ts +++ b/packages/core/src/snapshot.ts @@ -168,6 +168,7 @@ const layer = Layer.effect( const changed = yield* git.tree .files({ repository: compared.repository, from: compared.from, to: compared.to }) .pipe(Effect.mapError((cause) => failure("files", cause))) + if (changed.length === 0) return [] const skipped = yield* ignored("files", compared.source, changed) return changed.filter((file) => !skipped.has(file)) }) @@ -185,12 +186,10 @@ const layer = Layer.effect( paths: input.paths, }) .pipe(Effect.mapError((cause) => failure("diff", cause))) - const skipped = yield* ignored( - "diff", - compared.source, - diffs.map((file) => RelativePath.make(file.file)), - ) - return diffs.filter((file) => !skipped.has(RelativePath.make(file.file))) + if (diffs.length === 0) return [] + const paths = diffs.map((file) => RelativePath.make(file.file)) + const skipped = yield* ignored("diff", compared.source, paths) + return diffs.filter((_, index) => !skipped.has(paths[index]!)) }) const plan = Effect.fnUntraced(function* (worktree: AbsolutePath, input: RestoreInput) { diff --git a/packages/core/src/tool-output.ts b/packages/core/src/tool-output.ts index cc30f5976b64..f64e16880945 100644 --- a/packages/core/src/tool-output.ts +++ b/packages/core/src/tool-output.ts @@ -30,6 +30,8 @@ export type Editor = { export interface Interface extends State.Transformable { readonly truncate: (result: Result) => Effect.Effect + /** Writes bytes the model cannot receive inline to a retained file and returns its path. */ + readonly save: (bytes: Uint8Array, extension: string) => Effect.Effect readonly cleanup: () => Effect.Effect } @@ -117,10 +119,17 @@ const layer = Layer.effect( } }) + const save = Effect.fnUntraced(function* (bytes: Uint8Array, extension: string) { + const file = path.join(directory, fileName() + extension) + yield* fs.writeWithDirs(file, bytes).pipe(Effect.orDie) + return file + }) + return Service.of({ transform: state.transform, reload: state.reload, truncate, + save, cleanup: () => cleanup(fs, directory), }) }), diff --git a/packages/core/src/tool/mcp.ts b/packages/core/src/tool/mcp.ts index 4ba21e1bfe7f..6d1b8cc9753a 100644 --- a/packages/core/src/tool/mcp.ts +++ b/packages/core/src/tool/mcp.ts @@ -2,13 +2,19 @@ export * as McpTool from "./mcp.js" import { ToolFailure } from "@opencode/ai" import { McpEvent } from "@opencode/schema/mcp-event" -import { Context, Effect, Fiber, type JsonSchema, Layer, PubSub, Semaphore, Stream } from "effect" +import { Context, Effect, Fiber, type JsonSchema, Layer, Option, PubSub, Schema, Semaphore, Stream } from "effect" import { makeLocationNode } from "@opencode/util/effect/app-node" +import { extension } from "mime-types" import { Bus } from "../bus.js" import { Mcp } from "../mcp/index.js" +import { Mime } from "../mime.js" import { Permission } from "../permission.js" import { Tool } from "../tool.js" +import { ToolOutput } from "../tool-output.js" +import { ReadToolFileSystem } from "./read-filesystem.js" + +const decodeJson = Schema.decodeUnknownOption(Schema.fromJsonString(Schema.Unknown)) /** * Registry namespace and permission action names for MCP tools. @@ -30,6 +36,7 @@ export const layer = Layer.effect( const tools = yield* Tool.Service const bus = yield* Bus.Service const permission = yield* Permission.Service + const toolOutput = yield* ToolOutput.Service const lock = Semaphore.makeUnsafe(1) let discovered: Mcp.Tool[] = [] @@ -83,25 +90,14 @@ export const layer = Layer.effect( .join("\n") .trim() || "MCP tool returned an error", }) - const content = result.content.map((part) => - part.type === "text" - ? { type: "text" as const, text: part.text } - : { - type: "file" as const, - uri: `data:${part.mimeType};base64,${part.data}`, - mime: part.mimeType, - }, - ) + const content = yield* Effect.forEach(result.content, (part) => toContent(part, toolOutput)) const text = content.flatMap((part) => (part.type === "text" ? [part.text] : [])).join("\n") const output = () => { if (result.structured !== undefined) return result.structured if (text === "") return null // Agents assume JSON returned as text is already an object, so parse it when the server declares no schema. - if (tool.outputSchema === undefined && (text.startsWith("{") || text.startsWith("["))) { - try { - return JSON.parse(text) - } catch {} - } + if (tool.outputSchema === undefined && (text.startsWith("{") || text.startsWith("["))) + return Option.getOrElse(decodeJson(text), () => text) return text } return { @@ -145,8 +141,26 @@ export const layer = Layer.effect( }), ) +// Only media the read tool also accepts is sent inline; other types would fail request lowering on every later turn. +const toContent = Effect.fnUntraced(function* (part: Mcp.ToolResultContent, toolOutput: ToolOutput.Interface) { + if (part.type === "text") return { type: "text" as const, text: part.text } + const mime = part.mimeType.split(";", 1)[0].trim().toLowerCase() + if (ReadToolFileSystem.MEDIA_MIMES.has(mime)) + return { type: "file" as const, uri: `data:${part.mimeType};base64,${part.data}`, mime: part.mimeType } + const bytes = Buffer.from(part.data, "base64") + const label = part.uri ?? "content" + if (Mime.isTextual(mime)) + return { type: "text" as const, text: `[Resource ${label} (${part.mimeType})]\n${bytes.toString("utf8")}` } + const ext = extension(mime) + const file = yield* toolOutput.save(bytes, ext ? `.${ext}` : "") + return { + type: "text" as const, + text: `[Binary resource ${label} (${part.mimeType}, ${bytes.byteLength} bytes) saved to ${file}]`, + } +}) + export const node = makeLocationNode({ service: Service, layer, - deps: [Tool.node, Mcp.node, Bus.node, Permission.node], + deps: [Tool.node, Mcp.node, Bus.node, Permission.node, ToolOutput.node], }) diff --git a/packages/core/src/tool/plugin/mcp-resource.ts b/packages/core/src/tool/plugin/mcp-resource.ts index f7bf6e3db9da..845ab97ad802 100644 --- a/packages/core/src/tool/plugin/mcp-resource.ts +++ b/packages/core/src/tool/plugin/mcp-resource.ts @@ -5,6 +5,7 @@ import type { Context } from "@opencode/plugin/effect/plugin" import { Effect, Schema } from "effect" import { Mcp } from "../../mcp/index.js" import { Permission } from "../../permission.js" +import { ReadToolFileSystem } from "../read-filesystem.js" export const Plugin = { id: "opencode.tools.mcp-resources", @@ -79,7 +80,9 @@ export const Plugin = { return { output: resource, content: resource.contents.flatMap((part) => - part.type === "blob" && (part.mimeType?.startsWith("image/") || part.mimeType === "application/pdf") + part.type === "blob" && + part.mimeType !== undefined && + ReadToolFileSystem.MEDIA_MIMES.has(part.mimeType) ? [{ type: "file" as const, uri: `data:${part.mimeType};base64,${part.blob}`, mime: part.mimeType }] : [], ), diff --git a/packages/core/src/tool/plugin/question.ts b/packages/core/src/tool/plugin/question.ts index 60d115eae35f..67febf3a81b7 100644 --- a/packages/core/src/tool/plugin/question.ts +++ b/packages/core/src/tool/plugin/question.ts @@ -99,7 +99,7 @@ export const Plugin = { answers: input.questions.map((_, index): Question.Answer => { const value = state.answer[`q${index}`] if (value === undefined) return [] - if (typeof value === "object") return Array.from(value) + if (Array.isArray(value)) return Array.from(value) return [String(value)] }), } diff --git a/packages/core/src/tool/plugin/read.ts b/packages/core/src/tool/plugin/read.ts index 9150905fe4d6..61ff5639f0c6 100644 --- a/packages/core/src/tool/plugin/read.ts +++ b/packages/core/src/tool/plugin/read.ts @@ -105,12 +105,6 @@ export const Plugin = { Effect.catch(() => Effect.void), Effect.catchDefect(() => Effect.void), ) - if ( - result.content.type === "file" && - result.content.encoding === "base64" && - !ReadToolFileSystem.MEDIA_MIMES.has(result.content.mime) - ) - return yield* Effect.fail(new ReadToolFileSystem.BinaryFileError({ resource: result.target.resource })) return { output: result.content, path: result.path } }).pipe( Effect.map((result) => ({ diff --git a/packages/core/src/tool/plugin/subagent.ts b/packages/core/src/tool/plugin/subagent.ts index 87443f5e8a4b..0ff32a546c5e 100644 --- a/packages/core/src/tool/plugin/subagent.ts +++ b/packages/core/src/tool/plugin/subagent.ts @@ -188,7 +188,7 @@ export const Plugin = { .create({ parentID: context.sessionID, title: input.description, - agent: Agent.ID.make(input.agent), + agent: agent.id, model, }) .pipe( diff --git a/packages/core/src/tool/plugin/webfetch.ts b/packages/core/src/tool/plugin/webfetch.ts index 7c40d5b5106c..9e94d587e3b7 100644 --- a/packages/core/src/tool/plugin/webfetch.ts +++ b/packages/core/src/tool/plugin/webfetch.ts @@ -5,6 +5,7 @@ import { ToolFailure } from "@opencode/ai" import { Duration, Effect, Schema } from "effect" import { HttpClient, type HttpClientError, HttpClientRequest, HttpClientResponse } from "effect/http" import { Parser } from "htmlparser2" +import { Mime } from "../../mime.js" import { Permission } from "../../permission.js" import { convertHTMLToMarkdown, MAX_MARKDOWN_BYTES } from "../html-markdown.js" import { collectBoundedResponseBody } from "../http-body.js" @@ -84,15 +85,6 @@ const collectBody = (response: HttpClientResponse.HttpClientResponse) => const mimeFrom = (contentType: string) => contentType.split(";", 1)[0]?.trim().toLowerCase() ?? "" const isImageAttachment = (mime: string) => mime.startsWith("image/") && mime !== "image/svg+xml" && mime !== "image/vnd.fastbidsheet" -const isTextualMime = (mime: string) => - !mime || - mime.startsWith("text/") || - mime === "application/json" || - mime.endsWith("+json") || - mime === "application/xml" || - mime.endsWith("+xml") || - mime === "application/javascript" || - mime === "application/x-javascript" const convert = (content: string, contentType: string, format: Format) => { if (!contentType.includes("text/html")) return content if (format === "markdown") return convertHTMLToMarkdown(content) @@ -139,7 +131,7 @@ export const Plugin = { const mime = mimeFrom(contentType) if (isImageAttachment(mime)) return yield* Effect.fail(new Error(`Unsupported fetched image content type: ${mime}`)) - if (!isTextualMime(mime)) + if (!Mime.isTextual(mime)) return yield* Effect.fail(new Error(`Unsupported fetched file content type: ${mime}`)) return { body: yield* collectBody(response), contentType } }).pipe( diff --git a/packages/core/src/tool/runtime.ts b/packages/core/src/tool/runtime.ts index d1c5e29caf15..da38823c44c7 100644 --- a/packages/core/src/tool/runtime.ts +++ b/packages/core/src/tool/runtime.ts @@ -1,5 +1,6 @@ import type { ToolDefinition } from "@opencode/ai" import { emptyInputJsonSchema } from "@opencode/ai/tool" +import { isRecord } from "@opencode/ai/utils/record" import { Tool } from "@opencode/schema/tool" import type { StandardJSONSchemaV1, StandardSchemaV1 } from "@standard-schema/spec" import { Cache, Effect, JsonPointer, JsonSchema, Schema, SchemaIssue, SchemaRepresentation } from "effect" @@ -198,9 +199,6 @@ const flattenAllOf = (value: unknown): unknown => { return flattenAllOf({ ...Object.assign({}, ...allOf), ...rest }) } -const isRecord = (value: unknown): value is Record => - typeof value === "object" && value !== null && !Array.isArray(value) - const canFlattenAllOf = (allOf: ReadonlyArray>, parent: Record) => { const keys = new Set(Object.keys(parent).filter((key) => key !== "allOf")) return allOf.every((item) => @@ -264,7 +262,7 @@ export const normalizeContent = (value: string | ReadonlyArray | u return [{ type: "text" as const, text: stringify(output) }] } -const stringify = (value: unknown) => { +export const stringify = (value: unknown) => { if (typeof value === "string") return value try { return JSON.stringify(value) ?? String(value) diff --git a/packages/core/src/util/error.ts b/packages/core/src/util/error.ts deleted file mode 100644 index 574bc82953f0..000000000000 --- a/packages/core/src/util/error.ts +++ /dev/null @@ -1,68 +0,0 @@ -import { Schema } from "effect" - -export abstract class NamedError extends Error { - abstract schema(): Schema.Top - abstract toObject(): { name: string; data: unknown } - - static hasName(error: unknown, name: string): boolean { - return typeof error === "object" && error !== null && "name" in error && error.name === name - } - - static create( - name: Name, - fields: Fields, - ): ReturnType>> - static create( - name: Name, - data: DataSchema, - ): ReturnType> - static create(name: Name, data: Schema.Top | Schema.Struct.Fields) { - return NamedError.createSchemaClass(name, Schema.isSchema(data) ? data : Schema.Struct(data)) - } - - public static createSchemaClass(name: Name, data: DataSchema) { - const schema = Schema.Struct({ - name: Schema.Literal(name), - data, - }).annotate({ identifier: name }) - type Data = Schema.Schema.Type - - const result = class extends NamedError { - public static readonly Schema = schema - public static readonly EffectSchema = schema - public static readonly tag = name - - public override readonly name = name - - constructor( - public readonly data: Data, - options?: ErrorOptions, - ) { - super(name, options) - this.name = name - } - - static isInstance(input: unknown): input is InstanceType { - return NamedError.hasName(input, name) - } - - schema() { - return schema - } - - toObject() { - return { - name: name, - data: this.data, - } - } - } - Object.defineProperty(result, "name", { value: name }) - return result - } - - public static readonly Unknown = NamedError.create("UnknownError", { - message: Schema.String, - ref: Schema.optional(Schema.String), - }) -} diff --git a/packages/core/src/util/iife.ts b/packages/core/src/util/iife.ts deleted file mode 100644 index ca9ae6c10b44..000000000000 --- a/packages/core/src/util/iife.ts +++ /dev/null @@ -1,3 +0,0 @@ -export function iife(fn: () => T) { - return fn() -} diff --git a/packages/core/src/v1/config/error.ts b/packages/core/src/v1/config/error.ts deleted file mode 100644 index 65c7bd3ae906..000000000000 --- a/packages/core/src/v1/config/error.ts +++ /dev/null @@ -1,18 +0,0 @@ -export * as ConfigErrorV1 from "./error.js" - -import { Schema } from "effect" -import { NamedError } from "../../util/error.js" - -const Issue = Schema.StructWithRest( - Schema.Struct({ - message: Schema.String, - path: Schema.Array(Schema.String), - }), - [Schema.Record(Schema.String, Schema.Unknown)], -) - -export const InvalidError = NamedError.create("ConfigInvalidError", { - path: Schema.String, - issues: Schema.optional(Schema.Array(Issue)), - message: Schema.optional(Schema.String), -}) diff --git a/packages/core/src/v1/config/migrate.ts b/packages/core/src/v1/config/migrate.ts index 6c9eb2938d2e..2e59eb62667c 100644 --- a/packages/core/src/v1/config/migrate.ts +++ b/packages/core/src/v1/config/migrate.ts @@ -13,7 +13,6 @@ import { Model } from "../../model.js" const decodeOptions = { errors: "all", onExcessProperty: "ignore" } as const const decodeAgent = Schema.decodeUnknownSync(Schema.fromJsonString(ConfigAgent.Info), decodeOptions) -const encodeAgent = Schema.encodeSync(ConfigAgent.Info) function permissions(info?: ConfigPermissionV1.Info) { const rules = (info ?? []).flatMap(([key, rule]) => { @@ -26,7 +25,7 @@ function permissions(info?: ConfigPermissionV1.Info) { // Map v1 permission/tool keys onto their renamed v2 tool actions so migrated rules keep matching. export function normalizeAction(action: string) { - if (action === "write" || action === "patch") return "edit" + if (action === "write" || action === "patch" || action === "apply_patch") return "edit" if (action === "task") return "subagent" if (action === "bash") return "shell" return action @@ -38,21 +37,19 @@ export function migrateAgent(info: ConfigAgentV1.Info) { ...(info.temperature === undefined ? {} : { temperature: info.temperature }), ...(info.top_p === undefined ? {} : { top_p: info.top_p }), } - return encodeAgent( - decodeAgent( - JSON.stringify({ - model: modelSelection(info.model, info.variant), - request: Object.keys(body).length ? { body } : undefined, - system: info.prompt, - description: info.description, - mode: info.mode, - hidden: info.hidden, - color: info.color === undefined ? undefined : info.color.startsWith("#") ? info.color : "#aaaaaa", - steps: info.steps, - disabled: info.disable, - permissions: permissions(info.permission), - }), - ), + return decodeAgent( + JSON.stringify({ + model: modelSelection(info.model, info.variant), + request: Object.keys(body).length ? { body } : undefined, + system: info.prompt, + description: info.description, + mode: info.mode, + hidden: info.hidden, + color: info.color === undefined ? undefined : info.color.startsWith("#") ? info.color : "#aaaaaa", + steps: info.steps, + disabled: info.disable, + permissions: permissions(info.permission), + }), ) } diff --git a/packages/core/src/v1/config/provider-options.ts b/packages/core/src/v1/config/provider-options.ts index 8602acadfb03..16d2097d346e 100644 --- a/packages/core/src/v1/config/provider-options.ts +++ b/packages/core/src/v1/config/provider-options.ts @@ -1,5 +1,7 @@ export * as ConfigProviderOptionsV1 from "./provider-options.js" +import { isRecord } from "@opencode/ai/utils/record" + type Options = Readonly> export interface ProviderResult { @@ -12,13 +14,12 @@ export function provider(options: Options): ProviderResult { const headers = options.headers const body = options.body const settings = Object.fromEntries(Object.entries(options).filter(([key]) => key !== "headers" && key !== "body")) - const headerOverlay = - typeof headers === "object" && headers !== null && !Array.isArray(headers) - ? Object.fromEntries( - Object.entries(headers).filter((entry): entry is [string, string] => typeof entry[1] === "string"), - ) - : undefined - const bodyOverlay = typeof body === "object" && body !== null && !Array.isArray(body) ? { ...body } : undefined + const headerOverlay = isRecord(headers) + ? Object.fromEntries( + Object.entries(headers).filter((entry): entry is [string, string] => typeof entry[1] === "string"), + ) + : undefined + const bodyOverlay = isRecord(body) ? { ...body } : undefined return { settings, headers: headerOverlay, diff --git a/packages/core/src/v1/config/provider.ts b/packages/core/src/v1/config/provider.ts index eb93dbc9e7d3..4b00c1dfe5f9 100644 --- a/packages/core/src/v1/config/provider.ts +++ b/packages/core/src/v1/config/provider.ts @@ -93,19 +93,10 @@ export const Info = Schema.Struct({ setCacheKey: Schema.optional(Schema.Boolean).annotate({ description: "Enable promptCacheKey for this provider (default false)", }), - timeout: Schema.optional( - Schema.Union([PositiveInt, Schema.Literal(false)]).annotate({ - description: "Timeout in milliseconds for full requests to this provider. Set to false to disable timeout.", - }), - ).annotate({ + timeout: Schema.optional(Schema.Union([PositiveInt, Schema.Literal(false)])).annotate({ description: "Timeout in milliseconds for full requests to this provider. Set to false to disable timeout.", }), - headerTimeout: Schema.optional( - Schema.Union([PositiveInt, Schema.Literal(false)]).annotate({ - description: - "Timeout in milliseconds to wait for response headers. Provider integrations may set defaults. Set to false to disable timeout.", - }), - ).annotate({ + headerTimeout: Schema.optional(Schema.Union([PositiveInt, Schema.Literal(false)])).annotate({ description: "Timeout in milliseconds to wait for response headers. Provider integrations may set defaults. Set to false to disable timeout.", }), diff --git a/packages/core/src/voice.ts b/packages/core/src/voice.ts index fdfd664bf658..dea2d1e017f8 100644 --- a/packages/core/src/voice.ts +++ b/packages/core/src/voice.ts @@ -6,6 +6,7 @@ import { type RealtimeSession, type RealtimeTransport, type RealtimeTransportInput, + type RealtimeTransportType, } from "@opencode/ai/experimental" import { RequestExecutor } from "@opencode/ai/route" import { makeLocationNode } from "@opencode/util/effect/app-node" @@ -13,12 +14,52 @@ import { Context, Effect, Layer, Redacted, Schema } from "effect" import { requestExecutor } from "./effect/app-node-platform.js" import { Integration } from "./integration.js" +export interface ModelInfo { + readonly model: string + readonly provider: string + readonly transports: readonly RealtimeTransportType[] + readonly available: boolean + /** The integration whose connection starts its sessions: the provider's for a stored key, or a host's. */ + readonly integration?: string + /** The workspace its sessions bill to, for display: set only when a host that names one runs it. */ + readonly workspace?: string +} + +export interface Models { + readonly models: readonly ModelInfo[] + readonly default?: string +} + export interface SessionInput { readonly model: string readonly session?: RealtimeSession readonly transport: RealtimeTransportInput } +export interface Session { + readonly transport: RealtimeTransport + /** A hosted session's ID, which `instruct` names it by. */ + readonly id?: string +} + +export interface InstructInput { + readonly id: string + readonly instructions: string +} + +/** + * An integration that runs sessions for some models on its own account instead of a stored provider key, such as + * the OpenCode Console's hosted GPT-Live. Its sessions' backend instructions change through it. + */ +export interface Host { + readonly integration: string + readonly models: readonly string[] + /** The workspace its sessions bill to, for display. */ + readonly workspace?: string + readonly session: (input: SessionInput) => Effect.Effect + readonly instruct: (input: InstructInput) => Effect.Effect +} + export class InvalidRequestError extends Schema.TaggedError()("Voice.InvalidRequestError", { message: Schema.String, }) {} @@ -31,7 +72,12 @@ export class UnavailableError extends Schema.TaggedError()("Vo export type Error = InvalidRequestError | UnavailableError export interface Interface { - readonly session: (input: SessionInput) => Effect.Effect + readonly models: () => Effect.Effect + readonly session: (input: SessionInput) => Effect.Effect + /** Replaces the backend instructions of a running session the host started. */ + readonly instruct: (input: InstructInput) => Effect.Effect + /** Sets the host whose models start on it instead of a stored key, or removes it. */ + readonly host: (host: Host | undefined) => Effect.Effect } export class Service extends Context.Service()("@opencode/Voice") {} @@ -41,15 +87,47 @@ export const layer = Layer.effect( Effect.gen(function* () { const executor = yield* RequestExecutor.Service const integrations = yield* Integration.Service + let hosted: Host | undefined + + // Integration lists the active credential first, then newer stored credentials, then non-empty env vars. + const findConnection = (provider: string) => + integrations + .get(Integration.ID.make(provider)) + .pipe(Effect.map((info) => info?.connections.find((item) => item.type === "env" || item.method === "key"))) + + const models = Effect.fn("Voice.models")(function* () { + const host = hosted + const list = yield* Effect.forEach(Realtime.engines, (engine): Effect.Effect => + host?.models.includes(engine.model) + ? Effect.succeed({ + model: engine.model, + provider: engine.provider, + transports: engine.transports, + available: true, + integration: host.integration, + ...(host.workspace === undefined ? {} : { workspace: host.workspace }), + }) + : findConnection(engine.provider).pipe( + Effect.map((connection) => ({ + model: engine.model, + provider: engine.provider, + transports: engine.transports, + available: Boolean(connection), + ...(connection ? { integration: engine.provider } : {}), + })), + ), + ) + const first = list.find((item) => item.available)?.model + return first ? { models: list, default: first } : { models: list } + }) const session = Effect.fn("Voice.session")(function* (input: SessionInput) { const engine = Realtime.engine(input.model) if (!engine) return yield* new InvalidRequestError({ message: `Unsupported voice model: ${input.model}` }) + // A host offers a model on purpose, so it runs the model's sessions even where a stored key could. + if (hosted?.models.includes(input.model)) return yield* hosted.session(input) - // Integration lists the active credential first, then newer stored credentials, then its env vars. - const connection = (yield* integrations.get(Integration.ID.make(engine.provider)))?.connections.find( - (item) => item.type === "env" || item.method === "key", - ) + const connection = yield* findConnection(engine.provider) const credential = connection ? yield* integrations.connection .resolve(connection) @@ -63,7 +141,7 @@ export const layer = Layer.effect( message: missingKeyMessage[engine.provider], }) - return yield* Realtime.connect({ ...input, apiKey: Redacted.make(credential.key) }).pipe( + const transport = yield* Realtime.connect({ ...input, apiKey: Redacted.make(credential.key) }).pipe( Effect.provideService(RequestExecutor.Service, executor), Effect.mapError( (error: AIError): Error => @@ -72,9 +150,23 @@ export const layer = Layer.effect( : new UnavailableError({ message: error.message, service: engine.provider }), ), ) + return { transport } + }) + + const instruct = Effect.fn("Voice.instruct")(function* (input: InstructInput) { + if (!hosted) return yield* new InvalidRequestError({ message: "Only hosted voice sessions take instructions" }) + return yield* hosted.instruct(input) }) - return Service.of({ session }) + return Service.of({ + models, + session, + instruct, + host: (host) => + Effect.sync(() => { + hosted = host + }), + }) }), ) diff --git a/packages/core/src/worktree.ts b/packages/core/src/worktree.ts index 12a390adcd08..a7edc050a035 100644 --- a/packages/core/src/worktree.ts +++ b/packages/core/src/worktree.ts @@ -83,7 +83,7 @@ export const Event = Worktree.Event interface StoredInput { readonly directory: AbsolutePath - readonly strategy?: string + readonly strategy?: StrategyID readonly replace?: boolean } @@ -260,7 +260,7 @@ const layer = Layer.effect( const stored = yield* ops.find(input.projectID, worktreeDirectory) if (!stored?.strategy) return yield* new InvalidDirectoryError({ directory: worktreeDirectory }) const settings = yield* load(row.worktree, current) - const strategy = yield* getStrategy(StrategyID.make(stored.strategy), settings.strategies) + const strategy = yield* getStrategy(stored.strategy, settings.strategies) yield* strategy .remove({ directory: worktreeDirectory, diff --git a/packages/core/src/worktree/sql.ts b/packages/core/src/worktree/sql.ts index 86e1904fc2dd..63d01350d470 100644 --- a/packages/core/src/worktree/sql.ts +++ b/packages/core/src/worktree/sql.ts @@ -1,4 +1,5 @@ import { integer, primaryKey, sqliteTable, text } from "drizzle-orm/sqlite-core" +import type { Worktree } from "@opencode/schema/worktree" import { absoluteColumn } from "../database/path.js" import { ProjectSchema } from "../project/schema.js" import { ProjectTable } from "../project/sql.js" @@ -11,7 +12,7 @@ export const WorktreeTable = sqliteTable( .notNull() .references(() => ProjectTable.id, { onDelete: "cascade" }), directory: absoluteColumn().notNull(), - strategy: text(), + strategy: text().$type(), time_created: integer() .notNull() .$default(() => Date.now()), diff --git a/packages/core/test/aisdk.test.ts b/packages/core/test/aisdk.test.ts index c422716a8201..dab9790a044d 100644 --- a/packages/core/test/aisdk.test.ts +++ b/packages/core/test/aisdk.test.ts @@ -204,7 +204,9 @@ it.effect("uses only the provider timeout signal when the request signal is null let wrappedFetch: typeof fetch | undefined let requestSignal: AbortSignal | null | undefined yield* aisdk.hook.sdk((event) => { - wrappedFetch = event.options.fetch + if (typeof event.options.fetch === "function") { + wrappedFetch = event.options.fetch as typeof fetch + } event.sdk = { languageModel: () => ({ provider: event.model.providerID }) } }) diff --git a/packages/core/test/browser-idle.test.ts b/packages/core/test/browser-idle.test.ts index c0f6fc5ef22b..634b05f9721c 100644 --- a/packages/core/test/browser-idle.test.ts +++ b/packages/core/test/browser-idle.test.ts @@ -8,6 +8,7 @@ import { Global } from "@opencode/util/global" import { Bus } from "@opencode/core/bus" import { Database } from "@opencode/core/database/database" import { Location } from "@opencode/core/location" +import { Job } from "@opencode/core/job" import { LocationActivity } from "@opencode/core/location-activity" import { LocationServiceMap } from "@opencode/core/location-services" import { Plugin } from "@opencode/core/plugin" @@ -49,7 +50,7 @@ const it = testEffect( layer: LocationActivity.layer({ timeToLive, sweepInterval: "100 millis" }).pipe( Layer.provide(Layer.succeed(Clock.Clock, activityClock)), ), - deps: [Bus.node, LocationServiceMap.node, SessionExecution.node, SessionStore.node], + deps: [Bus.node, Job.node, LocationServiceMap.node, SessionExecution.node, SessionStore.node], }), ), ], diff --git a/packages/core/test/config/agent.test.ts b/packages/core/test/config/agent.test.ts index 7cd1758b3ecf..5b2f729a0ab0 100644 --- a/packages/core/test/config/agent.test.ts +++ b/packages/core/test/config/agent.test.ts @@ -32,8 +32,10 @@ function migrateV1(input: unknown) { } const defaultPermissions = Agent.Info.default(Agent.ID.make("test")).permissions -test("rejects named agent color tokens", () => { - expect(() => decode({ agents: { reviewer: { color: "warning" } } })).toThrow() +test("ignores named agent color tokens", () => { + expect(decode({ agents: { reviewer: { description: "Reviews", color: "warning" } } }).agents?.reviewer).toEqual({ + description: "Reviews", + }) }) test("keeps schema fields and name out of legacy agent options", () => { @@ -468,6 +470,10 @@ permissions: --- Use native v2 fields.`, ) + await fs.writeFile( + path.join(tmp.path, "agents", "themed.md"), + "---\ndescription: Themed\nmode: subagent\ncolor: accent\n---\nUse a theme color.", + ) await fs.writeFile(path.join(tmp.path, "agents", "disabled.md"), "---\ndisabled: true\n---\nDisabled") await fs.writeFile(path.join(tmp.path, "agents", "empty.md"), "") await fs.writeFile(path.join(tmp.path, "modes", "plan.md"), "Make a plan.") @@ -498,6 +504,9 @@ Use native v2 fields.`, request: { headers: { "x-agent": "native" }, body: { effort: "high" } }, permissions: [...defaultPermissions, { action: "edit", resource: "*", effect: "deny" }], }) + const themed = yield* agents.get(Agent.ID.make("themed")) + expect(themed).toMatchObject({ system: "Use a theme color.", description: "Themed" }) + expect(themed?.color).toBeUndefined() expect(yield* agents.get(Agent.ID.make("disabled"))).toBeUndefined() expect(yield* agents.get(Agent.ID.make("empty"))).toBeUndefined() expect(yield* agents.get(Agent.ID.make("plan"))).toMatchObject({ system: "Make a plan.", mode: "primary" }) diff --git a/packages/core/test/config/config.test.ts b/packages/core/test/config/config.test.ts index c7ccbb1b2cd7..61b6d5a7a606 100644 --- a/packages/core/test/config/config.test.ts +++ b/packages/core/test/config/config.test.ts @@ -1237,6 +1237,7 @@ describe("Config", () => { task: "ask", bash: { "git status": "allow", "*": "deny" }, write: "deny", + apply_patch: "deny", read: "allow", }, }).permissions, @@ -1245,6 +1246,7 @@ describe("Config", () => { { action: "shell", resource: "git status", effect: "allow" }, { action: "shell", resource: "*", effect: "deny" }, { action: "edit", resource: "*", effect: "deny" }, + { action: "edit", resource: "*", effect: "deny" }, { action: "read", resource: "*", effect: "allow" }, ]) }) diff --git a/packages/core/test/config/normalization.test.ts b/packages/core/test/config/normalization.test.ts index 09b8cfb93cd5..7725b7ff43e1 100644 --- a/packages/core/test/config/normalization.test.ts +++ b/packages/core/test/config/normalization.test.ts @@ -526,11 +526,17 @@ describe("ConfigNormalize", () => { portable_shell_scanner: true, todo: true, subagent_depth: 0, + location_idle_timeout: "2 hours", policies: [{ action: "provider.use", resource: "custom", effect: "allow" }], }, }).encoded, ).toMatchObject({ - experimental: { portable_shell_scanner: true, todo: true, subagent_depth: 0 }, + experimental: { + portable_shell_scanner: true, + todo: true, + subagent_depth: 0, + location_idle_timeout: "7200000 millis", + }, policies: [ { action: "provider.use", resource: "*", effect: "deny" }, { action: "provider.use", resource: "anthropic", effect: "allow" }, @@ -701,7 +707,7 @@ describe("ConfigNormalize", () => { agents: {}, providers: {}, instructions: [], - experimental: { subagent_depth: 0 }, + experimental: { subagent_depth: 0, location_idle_timeout: false }, }) expect(result.encoded).toMatchObject({ snapshots: false, @@ -710,7 +716,7 @@ describe("ConfigNormalize", () => { agents: {}, providers: {}, instructions: [], - experimental: { subagent_depth: 0 }, + experimental: { subagent_depth: 0, location_idle_timeout: false }, }) expect(result.encoded.share).toBeUndefined() }) diff --git a/packages/core/test/config/provider.test.ts b/packages/core/test/config/provider.test.ts index a4bab665fc20..9337ab2a0c3c 100644 --- a/packages/core/test/config/provider.test.ts +++ b/packages/core/test/config/provider.test.ts @@ -690,6 +690,38 @@ describe("ConfigProviderPlugin.Plugin", () => { }), ) + it.effect("lets a configured provider baseURL override catalog model baseURLs", () => + Effect.gen(function* () { + const providers = yield* Provider.Service + const models = yield* Model.Service + const providerID = Provider.ID.make("gateway") + yield* providers.transform((editor) => { + for (const id of ["chat", "pinned"]) + editor.models.update(providerID, Model.ID.make(id), (model) => { + model.settings = { baseURL: "https://catalog.example/v1" } + }) + }) + yield* addPlugin([ + new Document({ + type: "document", + info: decode({ + providers: { + gateway: { + settings: { baseURL: "http://127.0.0.1:17322" }, + models: { pinned: { settings: { baseURL: "http://127.0.0.1:9000" } } }, + }, + }, + }), + }), + ]) + + const baseURL = (id: string) => + models.get(providerID, Model.ID.make(id)).pipe(Effect.map((model) => required(model).settings?.baseURL)) + expect(yield* baseURL("chat")).toBe("http://127.0.0.1:17322") + expect(yield* baseURL("pinned")).toBe("http://127.0.0.1:9000") + }), + ) + it.effect("generates variants after rewriting a configured model package", () => Effect.gen(function* () { const models = yield* Model.Service diff --git a/packages/core/test/location-activity.test.ts b/packages/core/test/location-activity.test.ts index a71b0a82f459..e7a10ed15179 100644 --- a/packages/core/test/location-activity.test.ts +++ b/packages/core/test/location-activity.test.ts @@ -6,8 +6,10 @@ import { AppNodeBuilder } from "@opencode/core/effect/app-node-builder" import { LayerNode } from "@opencode/util/effect/layer-node" import { makeGlobalNode } from "@opencode/util/effect/app-node" import { Bus } from "@opencode/core/bus" +import { Config } from "@opencode/core/config" import { Database } from "@opencode/core/database/database" import { Form } from "@opencode/core/form" +import { Job } from "@opencode/core/job" import { Location } from "@opencode/core/location" import { LocationActivity } from "@opencode/core/location-activity" import { LocationServiceMap, type LocationServices } from "@opencode/core/location-services" @@ -21,8 +23,13 @@ import { SessionRunner } from "@opencode/core/session/runner/index" import { SessionTable } from "@opencode/core/session/sql" import { SessionStore } from "@opencode/core/session/store" import { Workspace } from "@opencode/core/workspace" +import { Document, Info } from "@opencode/schema/config" +import { ConfigExperimental } from "@opencode/schema/config/experimental" import { testEffect } from "./lib/effect" +// Per-workspace experimental.location_idle_timeout, read by each location's Config. +const idleTimeouts = new Map() + // Keep real execution ownership, location caching, forms, and eviction. The fixture // runner waits on a form instead of making a model request before asking a question. const locations = Layer.effect( @@ -31,9 +38,23 @@ const locations = Layer.effect( const bus = yield* Bus.Service const map = yield* LayerMap.make( (ref: Location.Ref) => - // The fixture only exercises these three Location services. + // The fixture only exercises these four Location services. // oxlint-disable-next-line typescript-eslint/no-unsafe-type-assertion - Layer.merge( + Layer.mergeAll( + Config.testLayer( + idleTimeouts.has(ref.workspaceID ?? "") + ? [ + new Document({ + type: "document", + info: new Info({ + experimental: new ConfigExperimental.Info({ + location_idle_timeout: idleTimeouts.get(ref.workspaceID ?? ""), + }), + }), + }), + ] + : [], + ), Layer.succeed( Location.Service, Location.Service.of({ @@ -84,6 +105,7 @@ const it = testEffect( LayerNode.group([ Database.node, Bus.node, + Job.node, SessionStore.node, LocationServiceMap.node, SessionExecution.node, @@ -102,121 +124,166 @@ const it = testEffect( ) describe("LocationActivity eviction", () => { - for (const [count, admission] of [ - [1, "none"], - [2, "none"], - [1, "other"], - [1, "same"], - ] as const) { - const newWork = admission !== "none" - it.effect( - `interrupts ${count} waiting executions before eviction (${admission} session admitted during cleanup)`, - () => - Effect.gen(function* () { - const db = (yield* Database.Service).db - const bus = yield* Bus.Service - const map = yield* LocationServiceMap.Service - const execution = yield* SessionExecution.Service - const store = yield* SessionStore.Service - const sessionIDs = Array.from({ length: count }, (_, index) => - Session.ID.make(`ses_waiting_question_${index}`), + for (const count of [1, 2]) { + it.effect(`keeps a location with ${count} executions waiting on a question until they go idle`, () => + Effect.gen(function* () { + const db = (yield* Database.Service).db + const bus = yield* Bus.Service + const map = yield* LocationServiceMap.Service + const execution = yield* SessionExecution.Service + const store = yield* SessionStore.Service + const sessionIDs = Array.from({ length: count }, (_, index) => Session.ID.make(`ses_waiting_question_${index}`)) + const ref = LocationServiceMap.canonical({ directory: AbsolutePath.make("/project") }) + const idle = Location.Ref.make({ directory: ref.directory, workspaceID: Workspace.ID.make("wrk_idle") }) + yield* db + .insert(ProjectTable) + .values({ id: Project.ID.global, worktree: ref.directory, sandboxes: [] }) + .run() + .pipe(Effect.orDie) + yield* db + .insert(SessionTable) + .values( + sessionIDs.map((sessionID) => ({ + id: sessionID, + project_id: Project.ID.global, + slug: "question", + directory: ref.directory, + title: "Waiting question", + version: "test", + })), ) - const newcomer = admission === "same" ? sessionIDs[0] : Session.ID.make("ses_new_question") - const ref = LocationServiceMap.canonical({ directory: AbsolutePath.make("/project") }) - const idle = Location.Ref.make({ directory: ref.directory, workspaceID: Workspace.ID.make("wrk_idle") }) - yield* db - .insert(ProjectTable) - .values({ id: Project.ID.global, worktree: ref.directory, sandboxes: [] }) - .run() - .pipe(Effect.orDie) - yield* db - .insert(SessionTable) - .values( - Array.from(new Set([...sessionIDs, newcomer]), (sessionID) => ({ - id: sessionID, - project_id: Project.ID.global, - slug: "question", - directory: ref.directory, - title: "Waiting question", - version: "test", - })), - ) - .run() - .pipe(Effect.orDie) + .run() + .pipe(Effect.orDie) - const created = yield* Deferred.make() - const newCreated = yield* Deferred.make() - const pending: Form.Info[] = [] - const interrupted: SessionEvent.Execution.Interrupted["data"][] = [] - const unsubscribe = yield* bus.listen((event) => - Effect.gen(function* () { - if (event.type === SessionEvent.Execution.Interrupted.type) { - interrupted.push(Schema.decodeUnknownSync(SessionEvent.Execution.Interrupted.data)(event.data)) - } - if (event.type !== Form.Event.Created.type) return - pending.push(Schema.decodeUnknownSync(Form.Event.Created.data)(event.data).form) - if (pending.length === count) yield* Deferred.succeed(created, undefined) - if (pending.length > count) yield* Deferred.succeed(newCreated, undefined) - }), - ) - yield* Effect.addFinalizer(() => unsubscribe) - const running = yield* Effect.forEach(sessionIDs, (sessionID) => - execution.resume(sessionID).pipe(Effect.exit, Effect.forkScoped), - ) - yield* Effect.addFinalizer(() => - Effect.forEach([...sessionIDs, newcomer], (sessionID) => execution.interrupt(sessionID)).pipe( - Effect.andThen(TestClock.adjust("5 minutes")), - ), - ) - yield* Deferred.await(created) - const context = yield* map.contextEffect(ref).pipe(Effect.scoped) - const forms = Context.get(context, Form.Service) - expect((yield* store.listSuspended()).toSorted()).toEqual(sessionIDs.toSorted()) - yield* Location.Service.pipe(Effect.provide(map.get(idle)), Effect.scoped) + const created = yield* Deferred.make() + const pending: Form.Info[] = [] + const interrupted: SessionEvent.Execution.Interrupted["data"][] = [] + const unsubscribe = yield* bus.listen((event) => + Effect.gen(function* () { + if (event.type === SessionEvent.Execution.Interrupted.type) { + interrupted.push(Schema.decodeUnknownSync(SessionEvent.Execution.Interrupted.data)(event.data)) + } + if (event.type !== Form.Event.Created.type) return + pending.push(Schema.decodeUnknownSync(Form.Event.Created.data)(event.data).form) + if (pending.length === count) yield* Deferred.succeed(created, undefined) + }), + ) + yield* Effect.addFinalizer(() => unsubscribe) + const running = yield* Effect.forEach(sessionIDs, (sessionID) => + execution.resume(sessionID).pipe(Effect.exit, Effect.forkScoped), + ) + yield* Deferred.await(created) + const context = yield* map.contextEffect(ref).pipe(Effect.scoped) + const forms = Context.get(context, Form.Service) + yield* Location.Service.pipe(Effect.provide(map.get(idle)), Effect.scoped) - // Human input produces no durable activity while the question is pending. - yield* TestClock.adjust("1 minute") - yield* TestClock.adjust("62 minutes") - // Interruption has cancelled each question, but slow cleanup still owns the graph. - expect(Array.from(yield* execution.active).toSorted()).toEqual(sessionIDs.toSorted()) - expect(Array.from(yield* RcMap.keys(map.rcMap))).toEqual([ref]) - expect(yield* forms.list()).toEqual([]) - for (const form of pending) expect(yield* forms.state(form.id)).toEqual({ status: "cancelled" }) + // Human input produces no durable activity while the question is pending. + yield* TestClock.adjust("1 minute") + yield* TestClock.adjust("62 minutes") + // The idle sibling expired in the same sweep; the waiting executions keep theirs. + expect(Array.from(yield* RcMap.keys(map.rcMap))).toEqual([ref]) + expect(Array.from(yield* execution.active).toSorted()).toEqual(sessionIDs.toSorted()) + expect(interrupted).toEqual([]) + expect((yield* forms.list()).map((form) => form.id).toSorted()).toEqual( + pending.map((form) => form.id).toSorted(), + ) - if (newWork) { - yield* execution.wake(newcomer) - if (admission === "other") yield* Deferred.await(newCreated) - } - yield* TestClock.adjust("5 minutes") - if (newWork) yield* Deferred.await(newCreated) - const results = yield* Effect.forEach(running, Fiber.join) - expect(results.every((exit) => exit._tag === "Failure")).toBe(true) - expect(Array.from(yield* execution.active)).toEqual(newWork ? [newcomer] : []) - expect(yield* store.listSuspended()).toEqual(newWork ? [newcomer] : []) - expect(interrupted.toSorted((a, b) => a.sessionID.localeCompare(b.sessionID))).toEqual( - sessionIDs.map((sessionID) => ({ sessionID, reason: "inactivity" })), - ) - expect(Array.from(yield* RcMap.keys(map.rcMap))).toEqual(newWork ? [ref] : []) - if (newWork) { - expect(yield* forms.list({ sessionID: newcomer })).toEqual([pending[count]]) - if (admission === "same") { - const later = LocationServiceMap.canonical({ directory: AbsolutePath.make("/later") }) - yield* Location.Service.pipe(Effect.provide(map.get(later)), Effect.scoped) - yield* TestClock.adjust("30 minutes") - // Keep fresh work active while a different graph reaches its own deadline. - yield* bus.publish(SessionEvent.Execution.Started, { sessionID: newcomer }, { location: ref }) - yield* TestClock.adjust("32 minutes") - expect(Array.from(yield* execution.active)).toEqual([newcomer]) - expect(Array.from(yield* RcMap.keys(map.rcMap))).toEqual([ref]) - } - yield* execution.interrupt(newcomer) - yield* TestClock.adjust("5 minutes") - yield* execution.awaitIdle(newcomer) - yield* TestClock.adjust("62 minutes") - expect(yield* store.listSuspended()).toEqual([]) - expect(Array.from(yield* RcMap.keys(map.rcMap))).toEqual([]) - } - }), + yield* Effect.forEach(sessionIDs, (sessionID) => execution.interrupt(sessionID)) + yield* TestClock.adjust("5 minutes") + const results = yield* Effect.forEach(running, Fiber.join) + expect(results.every((exit) => exit._tag === "Failure")).toBe(true) + expect(yield* execution.active).toEqual(new Set()) + expect(yield* store.listSuspended()).toEqual([]) + yield* TestClock.adjust("62 minutes") + expect(Array.from(yield* RcMap.keys(map.rcMap))).toEqual([]) + }), ) } + + for (const owner of ["shell", "subagent child", "subagent parent"] as const) + it.effect(`keeps a location with a running background job owned by its ${owner} until the job finishes`, () => + Effect.gen(function* () { + const db = (yield* Database.Service).db + const map = yield* LocationServiceMap.Service + const jobs = yield* Job.Service + const sessionID = Session.ID.make("ses_background_job") + // The other side of the subagent is unknown, so only the owner's location can keep the graph. + const missing = Session.ID.make("ses_missing") + const ref = LocationServiceMap.canonical({ directory: AbsolutePath.make("/project") }) + const idle = Location.Ref.make({ directory: ref.directory, workspaceID: Workspace.ID.make("wrk_idle") }) + yield* db + .insert(ProjectTable) + .values({ id: Project.ID.global, worktree: ref.directory, sandboxes: [] }) + .run() + .pipe(Effect.orDie) + yield* db + .insert(SessionTable) + .values({ + id: sessionID, + project_id: Project.ID.global, + slug: "background", + directory: ref.directory, + title: "Background job", + version: "test", + }) + .run() + .pipe(Effect.orDie) + yield* Location.Service.pipe(Effect.provide(map.get(ref)), Effect.scoped) + yield* Location.Service.pipe(Effect.provide(map.get(idle)), Effect.scoped) + const finish = yield* Deferred.make() + const job = yield* jobs.start({ + id: "job_background", + type: owner === "shell" ? "shell" : "subagent", + recovery: + owner === "shell" + ? { kind: "shell", sessionID, shellID: "shell_background", command: "sleep 9999" } + : { + kind: "subagent", + parentSessionID: owner === "subagent parent" ? sessionID : missing, + childSessionID: owner === "subagent child" ? sessionID : missing, + agent: "general", + description: "Background subagent", + }, + run: Deferred.await(finish), + }) + yield* jobs.background(job.id) + + yield* TestClock.adjust("1 minute") + yield* TestClock.adjust("62 minutes") + // The idle sibling expired in the same sweep; the job keeps its location. + expect(Array.from(yield* RcMap.keys(map.rcMap))).toEqual([ref]) + expect((yield* jobs.get(job.id))?.status).toBe("running") + + yield* Deferred.succeed(finish, "done") + yield* jobs.wait({ id: job.id }) + yield* TestClock.adjust("62 minutes") + expect(Array.from(yield* RcMap.keys(map.rcMap))).toEqual([]) + }), + ) + + for (const [timeout, after, evicted] of [ + ["10 minutes", "11 minutes", true], + [false, "3 hours", false], + ] as const) + it.effect(`honors experimental.location_idle_timeout = ${JSON.stringify(timeout)}`, () => + Effect.gen(function* () { + const map = yield* LocationServiceMap.Service + const workspaceID = Workspace.ID.make(`wrk_timeout_${String(timeout).replace(" ", "_")}`) + idleTimeouts.set(workspaceID, timeout === false ? false : Duration.fromInputUnsafe(timeout)) + yield* Effect.addFinalizer(() => Effect.sync(() => idleTimeouts.delete(workspaceID))) + const configured = Location.Ref.make({ directory: AbsolutePath.make("/project"), workspaceID }) + const fallback = Location.Ref.make({ directory: AbsolutePath.make("/project") }) + yield* Location.Service.pipe(Effect.provide(map.get(configured)), Effect.scoped) + yield* Location.Service.pipe(Effect.provide(map.get(fallback)), Effect.scoped) + + yield* TestClock.adjust("1 minute") + yield* TestClock.adjust(after) + const cached = Array.from(yield* RcMap.keys(map.rcMap)) + expect(cached.some((ref) => ref.workspaceID === workspaceID)).toBe(!evicted) + // Locations without the setting keep the 60 minute default. + expect(cached.some((ref) => ref.workspaceID === undefined)).toBe( + Duration.toMinutes(Duration.fromInputUnsafe(after)) < 60, + ) + }), + ) }) diff --git a/packages/core/test/location-layer.test.ts b/packages/core/test/location-layer.test.ts index 1987338e0f24..065508523059 100644 --- a/packages/core/test/location-layer.test.ts +++ b/packages/core/test/location-layer.test.ts @@ -2,7 +2,7 @@ import type { FileSystem } from "@opencode/core/filesystem" import fs from "fs/promises" import path from "path" import { describe, expect } from "bun:test" -import { Config } from "@opencode/schema/config" +import { Event } from "@opencode/schema/config" import { Money } from "@opencode/schema/money" import { DateTime, @@ -29,6 +29,7 @@ import { Global } from "@opencode/util/global" import { LocationServiceMap, type LocationServices } from "@opencode/core/location-services" import { LocationActivity } from "@opencode/core/location-activity" import { Location } from "@opencode/core/location" +import { Config } from "@opencode/core/config" import { LocationWatcher } from "@opencode/core/filesystem/location-watcher" import { Plugin } from "@opencode/core/plugin" import { Model } from "@opencode/core/model" @@ -59,14 +60,18 @@ const activityLocations = Layer.effect( LocationServiceMap.Service, LayerMap.make( (ref) => + // LocationActivity reads each location's config during its sweep. // oxlint-disable-next-line typescript-eslint/no-unsafe-type-assertion - Layer.succeed( - Location.Service, - Location.Service.of({ - directory: ref.directory, - workspaceID: ref.workspaceID, - project: { id: Project.ID.global, directory: ref.directory, canonical: ref.directory }, - }), + Layer.merge( + Layer.succeed( + Location.Service, + Location.Service.of({ + directory: ref.directory, + workspaceID: ref.workspaceID, + project: { id: Project.ID.global, directory: ref.directory, canonical: ref.directory }, + }), + ), + Config.testLayer(), ) as unknown as Layer.Layer, { idleTimeToLive: Duration.infinity }, ), @@ -338,19 +343,19 @@ describe("LocationServiceMap", () => { const firstContext = yield* locations.contextEffect(firstRef) const secondContext = yield* locations.contextEffect(secondRef) const received = { first: 0, second: 0 } - yield* bus.subscribe(Config.Event.Updated).pipe( + yield* bus.subscribe(Event.Updated).pipe( Stream.runForEach(() => Effect.sync(() => received.first++)), Effect.provideContext(firstContext), Effect.forkScoped({ startImmediately: true }), ) - yield* bus.subscribe(Config.Event.Updated).pipe( + yield* bus.subscribe(Event.Updated).pipe( Stream.runForEach(() => Effect.sync(() => received.second++)), Effect.provideContext(secondContext), Effect.forkScoped({ startImmediately: true }), ) yield* Effect.sleep("10 millis") - yield* bus.publish(Config.Event.Updated, {}, { location: firstRef }) + yield* bus.publish(Event.Updated, {}, { location: firstRef }) yield* Effect.sleep("10 millis") expect(received).toEqual({ first: 1, second: 0 }) diff --git a/packages/core/test/mcp-instructions.test.ts b/packages/core/test/mcp-instructions.test.ts index 906a6d78ca85..8635a80d0e46 100644 --- a/packages/core/test/mcp-instructions.test.ts +++ b/packages/core/test/mcp-instructions.test.ts @@ -116,7 +116,7 @@ describe("McpInstructions", () => { ), ) - it.effect("restates guidance when Code Mode is disabled for a server", () => { + it.effect("replaces a server entry when Code Mode is disabled for it", () => { let tools: Mcp.Tool[] = [tool("alpha")] return Effect.gen(function* () { const service = yield* McpInstructions.Service @@ -126,12 +126,10 @@ describe("McpInstructions", () => { const changed = yield* readUpdate(yield* service.load([]), initialized) expect(changed.text).toBe( [ - "The available MCP server instructions have changed. This list supersedes the previous one.", - "", + "The following MCP server instructions have changed. Each entry replaces the previously listed server with the same name:", ' ', " Alpha instructions", " ", - "", ].join("\n"), ) }).pipe( @@ -167,17 +165,11 @@ describe("McpInstructions", () => { const changed = yield* readUpdate(yield* service.load([]), added) expect(changed.text).toBe( [ - "The available MCP server instructions have changed. This list supersedes the previous one.", - "", + "The following MCP server instructions have changed. Each entry replaces the previously listed server with the same name:", ' ', ' Use tools from this server through `execute` under `tools["alpha"]`.', " Updated alpha", " ", - ' ', - ' Use tools from this server through `execute` under `tools["beta"]`.', - " Beta instructions", - " ", - "", ].join("\n"), ) diff --git a/packages/core/test/mcp.test.ts b/packages/core/test/mcp.test.ts index 7edd7c4e953a..bee57c0a0e0a 100644 --- a/packages/core/test/mcp.test.ts +++ b/packages/core/test/mcp.test.ts @@ -1,5 +1,6 @@ import path from "node:path" import fs from "node:fs/promises" +import os from "node:os" import { describe, expect, test } from "bun:test" import { Client, InMemoryTransport, StreamableHTTPClientTransport } from "@modelcontextprotocol/client" import { @@ -56,6 +57,7 @@ import { TestClock } from "effect/testing" import { ChildProcess, ChildProcessSpawner } from "effect/process" import { ExitCode, makeHandle, ProcessId } from "effect/process/ChildProcessSpawner" import { Image } from "@opencode/core/image" +import { Global } from "@opencode/util/global" import { advance, drain } from "./lib/clock" import { testEffect } from "./lib/effect" import { registerIntegrationPolicy } from "./fixture/policy" @@ -446,6 +448,13 @@ const mcp = Layer.mock(Mcp.Service, { description: "Returns text and an image", inputSchema: { type: "object", properties: {} }, } satisfies Mcp.Tool, + { + server: Mcp.ServerName.make("direct"), + name: "attachments", + codemode: false, + description: "Returns resource blobs that cannot be sent inline", + inputSchema: { type: "object", properties: {} }, + } satisfies Mcp.Tool, ]), callTool: (input) => Effect.sync(() => { @@ -468,6 +477,26 @@ const mcp = Layer.mock(Mcp.Service, { { type: "media", data: "aGVsbG8=", mimeType: "image/png" }, ], } satisfies Mcp.ToolResult + if (input.name === "attachments") + return { + server: Mcp.ServerName.make(input.server), + tool: input.name, + isError: false, + content: [ + { + type: "media", + data: Buffer.from("# Notes").toString("base64"), + mimeType: "text/markdown", + uri: "file:///README.md", + }, + { + type: "media", + data: "UEFSMQ==", + mimeType: "application/vnd.apache.parquet", + uri: "file:///data.parquet", + }, + ], + } satisfies Mcp.ToolResult if (input.name === "status") return { server: Mcp.ServerName.make(input.server), @@ -507,12 +536,14 @@ const permissions = Layer.mock(Permission.Service, { }), }) const events = Layer.mock(Bus.Service, { subscribe: () => Stream.never }) +const outputData = await fs.mkdtemp(path.join(os.tmpdir(), "opencode-mcp-output-")) const it = testEffect( AppNodeBuilder.build(LayerNode.group([Tool.node, McpTool.node]), [ Mcp.node.replace(mcp), Permission.node.replace(permissions), Bus.node.replace(events), Image.node.replace(imagePassthrough), + Global.node.replace(Global.layerWith({ data: outputData })), ]), ) @@ -2245,6 +2276,7 @@ it.effect("advertises MCP output schemas to Code Mode", () => const execute = toolSet.definitions.find((tool) => tool.name === "execute") expect(toolSet.definitions.map((tool) => tool.name)).toEqual([ + "direct_attachments", "direct_fail", "direct_issues", "direct_lookup", @@ -2426,6 +2458,31 @@ it.effect("preserves MCP text and media content for the model", () => }), ) +it.effect("decodes text blobs and saves other MCP blobs instead of attaching them", () => + Effect.gen(function* () { + assertion = yield* Deferred.make() + decision = Effect.void + const registry = yield* Tool.Service + const registration = yield* McpTool.Service + yield* registration.flush + + const execution = yield* executeTool(registry, { + sessionID: Session.ID.make("ses_mcp_attachments"), + ...toolIdentity, + call: { type: "tool-call", id: "call_mcp_attachments", name: "direct_attachments", input: {} }, + }) + + const text = execution.content?.map((part) => (part.type === "text" ? part.text : part.mime)) ?? [] + const saved = + /^\[Binary resource file:\/\/\/data\.parquet \(application\/vnd\.apache\.parquet, 4 bytes\) saved to (.+)\]$/.exec( + text[1] ?? "", + )?.[1] + expect(text[0]).toBe("[Resource file:///README.md (text/markdown)]\n# Notes") + expect(saved?.startsWith(path.join(outputData, ToolOutput.DIRECTORY, "tool_"))).toBe(true) + expect(yield* Effect.promise(() => fs.readFile(saved ?? "", "utf8"))).toBe("PAR1") + }), +) + it.effect("waits for permission before calling an MCP tool", () => Effect.gen(function* () { calls = 0 diff --git a/packages/core/test/model-free.test.ts b/packages/core/test/model-free.test.ts new file mode 100644 index 000000000000..cc61728c21ed --- /dev/null +++ b/packages/core/test/model-free.test.ts @@ -0,0 +1,90 @@ +import { expect } from "bun:test" +import { DateTime, Effect, Layer } from "effect" +import { AppNodeBuilder } from "@opencode/core/effect/app-node-builder" +import { LayerNode } from "@opencode/util/effect/layer-node" +import { Location } from "@opencode/core/location" +import { Model } from "@opencode/core/model" +import { ModelResolver } from "@opencode/core/model-resolver" +import { Project } from "@opencode/core/project" +import { Provider } from "@opencode/core/provider" +import { AbsolutePath } from "@opencode/core/schema" +import { Session } from "@opencode/core/session" +import { SessionRunnerModel } from "@opencode/core/session/runner/model" +import { Money } from "@opencode/schema/money" +import { withEnv } from "./fixture/env" +import { location } from "./fixture/location" +import { it } from "./lib/effect" + +const ref = location({ directory: AbsolutePath.make("test") }) +const layer = AppNodeBuilder.build( + LayerNode.group([Provider.node, Model.node, ModelResolver.node, SessionRunnerModel.node]), + [Location.node.replace(Layer.succeed(Location.Service, Location.Service.of(ref)))], +) +const free = Model.ID.make("big-pickle") +const paid = Model.ID.make("paid") +const local = Model.ID.make("local") +const other = Provider.ID.make("other") + +for (const flag of [undefined, "0", "1", "true"]) { + for (const providerID of [Provider.ID.opencode, Provider.ID.make("opencode-go")]) + it.effect(`${providerID} catalog with OPENCODE_DISABLE_FREE_MODELS=${flag}`, () => + withEnv({ OPENCODE_DISABLE_FREE_MODELS: flag }, () => + Effect.gen(function* () { + const providers = yield* Provider.Service + const models = yield* Model.Service + const runner = yield* SessionRunnerModel.Service + const resolver = yield* ModelResolver.Service + yield* providers.transform((editor) => { + editor.update(providerID, () => {}) + editor.models.update(providerID, free, (model) => { + model.time.released = 3000 + model.family = Model.Family.make("gpt-luna") + model.capabilities = { ...model.capabilities, input: ["text"], output: ["text"] } + }) + editor.models.update(providerID, paid, (model) => { + model.time.released = 2000 + model.cost = [ + { + input: Money.USDPerMillionTokens.make(1), + output: Money.USDPerMillionTokens.zero, + cache: { read: Money.USDPerMillionTokens.zero, write: Money.USDPerMillionTokens.zero }, + }, + ] + }) + editor.update(other, () => {}) + editor.models.update(other, local, () => {}) + }) + yield* models.transform((editor) => editor.default.set(providerID, free)) + const disabled = flag === "1" || flag === "true" + expect((yield* models.all()).map((model) => model.id)).toEqual(disabled ? [paid, local] : [free, paid, local]) + expect(yield* models.available()).toEqual(yield* models.all()) + expect((yield* models.default())?.id).toBe(disabled ? paid : free) + expect((yield* models.get(providerID, free))?.id).toBe(disabled ? undefined : free) + expect((yield* models.small(providerID))?.id).toBe(disabled ? undefined : free) + if (!disabled) return + const session = Session.Info.make({ + id: Session.ID.make("ses_free_model"), + projectID: Project.ID.global, + title: "test", + model: { id: free, providerID }, + cost: Money.USD.zero, + tokens: { input: 0, output: 0, reasoning: 0, cache: { read: 0, write: 0 } }, + time: { created: DateTime.makeUnsafe(0), updated: DateTime.makeUnsafe(0) }, + location: ref, + }) + const failure = yield* runner.resolve(session, models.available).pipe(Effect.flip) + expect(failure.message).toBe(`Model unavailable: ${providerID}/big-pickle`) + expect(yield* resolver.resolve(Model.Ref.make({ providerID, id: free }))).toBeUndefined() + // Later transforms cannot restore a free model, even when all paid models are removed. + yield* models.transform((editor) => { + editor.update(providerID, free, (model) => { + model.enabled = true + }) + editor.remove(providerID, paid) + }) + expect((yield* models.default())?.providerID).toBe(other) + expect(yield* models.get(providerID, free)).toBeUndefined() + }).pipe(Effect.provide(layer)), + ), + ) +} diff --git a/packages/core/test/permission.test.ts b/packages/core/test/permission.test.ts index d7a81110d435..d85e03247d85 100644 --- a/packages/core/test/permission.test.ts +++ b/packages/core/test/permission.test.ts @@ -49,7 +49,7 @@ function setup(rules: Permission.Ruleset = [], sessionID = Session.ID.make("ses_ directory: "/project", title: "test", version: "test", - agent: "test", + agent: Agent.ID.make("test"), }) .onConflictDoNothing() .run() diff --git a/packages/core/test/persistent-pty-daemon.test.ts b/packages/core/test/persistent-pty-daemon.test.ts index a9657376f014..8c0ffc9c7bff 100644 --- a/packages/core/test/persistent-pty-daemon.test.ts +++ b/packages/core/test/persistent-pty-daemon.test.ts @@ -345,7 +345,7 @@ function terminal(id: number) { pid: null, title: "test", foreground_process: null, - group_id: "test", + group_id: "ses_test", command: ["test"], cwd: "/tmp", cols: 80, diff --git a/packages/core/test/plugin/fixture.ts b/packages/core/test/plugin/fixture.ts index 3631e1f94191..ed1486f79d3e 100644 --- a/packages/core/test/plugin/fixture.ts +++ b/packages/core/test/plugin/fixture.ts @@ -35,6 +35,7 @@ import { SkillDiscovery } from "@opencode/core/skill/discovery" import { Watcher } from "@opencode/core/filesystem/watcher" import { Tool } from "@opencode/core/tool" import { Vcs } from "@opencode/core/vcs" +import { Voice } from "@opencode/core/voice" import { WebSearch } from "@opencode/core/websearch" import { Worktree } from "@opencode/core/worktree" import { WorktreeStrategies } from "@opencode/core/worktree/strategies" @@ -107,6 +108,7 @@ export const makePluginTestLayer = (permission = permissionLayer) => SkillDiscovery.node, Tool.node, Vcs.node, + Voice.node, Watcher.node, WebSearch.node, Worktree.node, diff --git a/packages/core/test/plugin/provider-opencode.test.ts b/packages/core/test/plugin/provider-opencode.test.ts index 832acaa71e13..258a4f959ada 100644 --- a/packages/core/test/plugin/provider-opencode.test.ts +++ b/packages/core/test/plugin/provider-opencode.test.ts @@ -22,6 +22,7 @@ import { PluginHost } from "@opencode/core/plugin/host" import { OpencodePlugin } from "@opencode/core/plugin/provider/opencode" import { Provider } from "@opencode/core/provider" import { Session } from "@opencode/core/session" +import { Voice } from "@opencode/core/voice" import { WebSearch } from "@opencode/core/websearch" import { withEnv } from "../fixture/env" import { emptyMcp } from "../fixture/mcp" @@ -1165,13 +1166,14 @@ describe("OpencodePlugin", () => { ), ) - it.live("explains rejected Console inference with the sign-in the connection needs", () => + it.effect("explains rejected Console inference from at most one config check per 30 seconds", () => Effect.acquireUseRelease( Effect.sync(() => { - const state = { sso: false } + const state = { sso: false, requests: 0 } const server = Bun.serve({ port: 0, fetch: () => { + state.requests++ if (state.sso) return Response.json( { _tag: "SsoRequired", orgId: "org_acme", connectionId: "ssoconn_1" }, @@ -1194,7 +1196,8 @@ describe("OpencodePlugin", () => { metadata: { server: server.url.origin, orgID: "org_acme", orgName: "Acme" }, }), }) - yield* addPlugin() + yield* addPlugin().pipe(inProcess(server)) + yield* drain const rejected = (providerID: Provider.ID) => hooks .trigger("session", "http.response", { @@ -1209,24 +1212,36 @@ describe("OpencodePlugin", () => { Effect.flatMap((event) => Effect.promise(async () => ({ status: event.response.status, body: await event.response.json() })), ), + inProcess(server), ) const denied = { status: 403, body: { error: { message: "Workspace access denied" } } } + const explained = { + status: 403, + body: { error: { type: "authentication_error", message: "Sign in with SSO again to use Acme." } }, + } + expect(state.requests).toBe(1) expect(yield* rejected(Provider.ID.make("remote"))).toEqual(denied) + expect(state.requests).toBe(1) state.sso = true - const explained = { - status: 403, - body: { - error: { - type: "authentication_error", - message: "Sign in with SSO again to use Acme.", - }, - }, - } - expect(yield* rejected(Provider.ID.make("remote"))).toEqual(explained) + yield* TestClock.adjust("30 seconds") + expect( + yield* Effect.all( + Array.from({ length: 10 }, () => rejected(Provider.ID.make("remote"))), + { concurrency: "unbounded" }, + ), + ).toEqual(Array.from({ length: 10 }, () => explained)) + expect(state.requests).toBe(2) + + expect( + yield* Effect.forEach(Array.from({ length: 20 }), () => + TestClock.adjust("1 second").pipe(Effect.andThen(rejected(Provider.ID.make("remote")))), + ), + ).toEqual(Array.from({ length: 20 }, () => explained)) expect(yield* rejected(Provider.ID.opencode)).toEqual(explained) expect(yield* rejected(Provider.ID.anthropic)).toEqual(denied) + expect(state.requests).toBe(2) }), ({ server }) => Effect.promise(() => server.stop(true)), ), @@ -1299,6 +1314,25 @@ describe("OpencodePlugin", () => { }), ) + it.effect("sends credentials saved with the retired console.opencode.ai host to the default Console", () => + Effect.gen(function* () { + const credentials = yield* Credential.Service + const requests: string[] = [] + const http = HttpClient.make((request) => { + requests.push(request.url) + return Effect.succeed(HttpClientResponse.fromWeb(request, new Response(null, { status: 404 }))) + }) + yield* credentials.create({ + integrationID: Integration.ID.make("opencode"), + value: Credential.Key.make({ type: "key", key: "secret", metadata: { server: "https://console.opencode.ai" } }), + }) + yield* addPlugin().pipe(Effect.provideService(HttpClient.HttpClient, http)) + yield* drain + + expect(requests).toEqual(["https://opencode.ai/console/api/v2/config"]) + }), + ) + it.effect("reports a rejected Console refresh token as signed out without removing the credential", () => Effect.acquireUseRelease( Effect.sync(() => @@ -1994,3 +2028,202 @@ describe("OpencodePlugin", () => { ), ) }) + +describe("OpencodePlugin hosted voice", () => { + // A Console whose config offers hosted GPT-Live while `offered`, answering sessions with `reply`. + const voiceConsole = (reply: () => Response) => + Effect.sync(() => { + const state = { offered: true } + const requests: { path: string; authorization: string | null; orgID: string | null; body: unknown }[] = [] + const server = Bun.serve({ + port: 0, + fetch: async (request) => { + const url = new URL(request.url) + if (url.pathname === "/api/v2/config") + return Response.json({ + providers: {}, + ...(state.offered + ? { + voice: { + baseURL: `${url.origin}/inference/openai/v1`, + models: ["gpt-live-1"], + workspace: "Anomaly", + }, + } + : {}), + }) + requests.push({ + path: url.pathname, + authorization: request.headers.get("authorization"), + orgID: request.headers.get("x-org-id"), + body: await request.json(), + }) + if (url.pathname.endsWith("/instructions")) return new Response(null, { status: 204 }) + return reply() + }, + }) + return { server, state, requests } + }) + + const connect = (server: ReturnType) => + Effect.gen(function* () { + const credentials = yield* Credential.Service + yield* credentials.create({ + integrationID: Integration.ID.make("opencode"), + value: Credential.OAuth.make({ + type: "oauth", + methodID: Integration.MethodID.make("device"), + access: "secret", + refresh: "refresh", + expires: Date.now() + 600_000, + metadata: { server: server.url.origin, orgID: "org_test" }, + }), + }) + yield* addPlugin().pipe(inProcess(server)) + }) + + const offer = { + model: "gpt-live-1", + session: { audio: { output: { voice: "marin" } } }, + transport: { type: "webrtc", sdp: "v=0" }, + } as const + + it.effect("starts GPT-Live sessions on the Console with the connection's credential while it offers them", () => + Effect.acquireUseRelease( + voiceConsole(() => Response.json({ session: { id: "sess_console" }, transport: { sdp: "answer:console" } })), + ({ server, state, requests }) => + Effect.gen(function* () { + const voice = yield* Voice.Service + yield* connect(server) + const live = Effect.map(voice.models(), (list) => list.models.find((item) => item.model === "gpt-live-1")) + + expect(yield* live).toMatchObject({ integration: "opencode", workspace: "Anomaly" }) + expect(yield* voice.session(offer)).toEqual({ + transport: { type: "webrtc", sdp: "answer:console" }, + id: "sess_console", + }) + yield* voice.instruct({ id: "sess_console", instructions: "The screen shows ses_1." }) + expect(requests).toEqual([ + { + path: "/inference/openai/v1/live/sessions", + authorization: "Bearer secret", + orgID: "org_test", + body: { + session: { audio: { output: { voice: "marin" } }, model: "gpt-live-1" }, + transport: { type: "webrtc", sdp: "v=0" }, + }, + }, + { + path: "/inference/openai/v1/live/sessions/sess_console/instructions", + authorization: "Bearer secret", + orgID: "org_test", + body: { instructions: "The screen shows ses_1." }, + }, + ]) + + state.offered = false + yield* TestClock.adjust("1 minute") + yield* drain + expect((yield* live)?.integration).not.toBe("opencode") + }), + ({ server }) => Effect.promise(() => server.stop(true)), + ), + ) + + for (const item of [ + { + name: "a refusal's reason", + reply: () => + Response.json( + { error: { type: "permission_error", message: "GPT-Live isn't enabled for this workspace" } }, + { status: 403 }, + ), + expected: { + _tag: "Voice.UnavailableError", + service: "opencode", + message: "GPT-Live isn't enabled for this workspace", + }, + }, + { + name: "an invalid session as the caller's", + reply: () => + Response.json( + { error: { type: "invalid_request_error", message: "Stored GPT-Live sessions aren't available" } }, + { status: 400 }, + ), + expected: { _tag: "Voice.InvalidRequestError", message: "Stored GPT-Live sessions aren't available" }, + }, + { + name: "a rejected credential as a reconnect", + reply: () => Response.json({ error: { message: "Invalid credential" } }, { status: 401 }), + expected: { _tag: "Voice.UnavailableError", message: "Reconnect OpenCode Console to continue." }, + }, + ]) + it.effect(`reports ${item.name}`, () => + Effect.acquireUseRelease( + voiceConsole(item.reply), + ({ server }) => + Effect.gen(function* () { + const voice = yield* Voice.Service + yield* connect(server) + expect(yield* voice.session(offer).pipe(Effect.flip)).toMatchObject(item.expected) + }), + ({ server }) => Effect.promise(() => server.stop(true)), + ), + ) +}) + +for (const key of [undefined, "secret"]) { + testEffect(Layer.empty).effect( + `disabled free models with ${key ? "an API key" : "a Console login after startup"}`, + () => + withEnv({ OPENCODE_DISABLE_FREE_MODELS: "true", OPENCODE_API_KEY: key }, () => + Effect.gen(function* () { + const catalog = yield* Provider.Service + const models = yield* Model.Service + const integrations = yield* Integration.Service + const credentials = yield* Credential.Service + yield* integrations.transform((editor) => + editor.method.update({ + integrationID: Integration.ID.make("opencode"), + method: { type: "env", names: ["OPENCODE_API_KEY"] }, + }), + ) + yield* catalog.transform((editor) => { + editor.update(Provider.ID.opencode, () => {}) + editor.models.update(Provider.ID.opencode, Model.ID.make("free"), (model) => { + model.cost = cost(0) + }) + editor.models.update(Provider.ID.opencode, Model.ID.make("paid"), (model) => { + model.cost = cost(1) + }) + }) + yield* addPlugin().pipe(Effect.provideService(HttpClient.HttpClient, noRemoteConfig)) + if (!key) { + expect(yield* catalog.available()).toEqual([]) + expect(yield* models.all()).toEqual([]) + expect(yield* models.default()).toBeUndefined() + expect((yield* integrations.get(Integration.ID.make("opencode")))?.methods.length).toBeGreaterThan(0) + yield* credentials.create({ + integrationID: Integration.ID.make("opencode"), + value: Credential.OAuth.make({ + type: "oauth", + methodID: Integration.MethodID.make("device"), + access: "access", + refresh: "refresh", + expires: Date.now() + 600_000, + }), + }) + } + const available = yield* eventually(models.available(), (models) => + models.some((model) => model.providerID === Provider.ID.opencode), + ) + expect((yield* catalog.available()).map((provider) => provider.id)).toContain(Provider.ID.opencode) + expect(available.map((model) => model.id)).toEqual([Model.ID.make("paid")]) + expect(yield* models.all()).toEqual(available) + expect((yield* models.default())?.id).toBe(Model.ID.make("paid")) + expect(yield* models.get(Provider.ID.opencode, Model.ID.make("free"))).toBeUndefined() + }).pipe(Effect.provide(PluginTestLayer)), + ), + ) +} diff --git a/packages/core/test/reference-instructions.test.ts b/packages/core/test/reference-instructions.test.ts index b1e8fce27776..ee1338437eef 100644 --- a/packages/core/test/reference-instructions.test.ts +++ b/packages/core/test/reference-instructions.test.ts @@ -108,4 +108,40 @@ describe("ReferenceInstructions", () => { Effect.provide(instructionsLayer(Layer.mock(Reference.Service, { list: () => Effect.succeed(references) }))), ) }) + + it.effect("announces added, changed, and removed references in one delta", () => { + const reference = (name: string, description: string) => + Reference.Info.make({ + name, + path: AbsolutePath.make(`/${name}`), + description, + source: Reference.LocalSource.make({ type: "local", path: AbsolutePath.make(`/${name}`), description }), + }) + let references = [reference("docs", "Use for product documentation"), reference("legacy", "Use for old code")] + return Effect.gen(function* () { + const instructions = yield* ReferenceInstructions.Service + const initialized = yield* readInitial(yield* instructions.load()) + + references = [reference("docs", "Use for API documentation"), reference("examples", "Use for examples")] + expect((yield* readUpdate(yield* instructions.load(), initialized)).text).toBe( + [ + "New project references are available in addition to those previously listed:", + " ", + " examples", + " /examples", + " Use for examples", + " ", + "The following project references have changed. Each entry replaces the previously listed reference with the same name:", + " ", + " docs", + " /docs", + " Use for API documentation", + " ", + "The following project references are no longer available and must not be used: legacy.", + ].join("\n"), + ) + }).pipe( + Effect.provide(instructionsLayer(Layer.mock(Reference.Service, { list: () => Effect.succeed(references) }))), + ) + }) }) diff --git a/packages/core/test/session-compaction.test.ts b/packages/core/test/session-compaction.test.ts index bd7bd3029f6f..85eb563db4cc 100644 --- a/packages/core/test/session-compaction.test.ts +++ b/packages/core/test/session-compaction.test.ts @@ -199,6 +199,43 @@ it.effect("forks rebuild the recent text of a summary they copied", () => }), ) +it.effect("ignores non-message keepID strings when loading history or forking", () => + Effect.gen(function* () { + const parentID = Session.ID.make("ses_invalid_keepid_parent") + const childID = Session.ID.make("ses_invalid_keepid_child") + yield* insertSession(parentID) + const compaction = SessionMessage.CompactionCompleted.make({ + id: SessionMessage.ID.create(), + type: "compaction", + status: "completed", + reason: "manual", + summary: "## Objective\n- plugin summary", + metadata: { keepID: "not-a-msg-id" }, + time: { created: DateTime.makeUnsafe(0) }, + }) + yield* storeMessages(parentID, [textUser("Earlier question"), compaction]) + + const store = yield* SessionStore.Service + const [parent] = yield* store.context(parentID) + expect(parent).toMatchObject({ + type: "compaction", + metadata: { keepID: "not-a-msg-id" }, + }) + + const bus = yield* Bus.Service + yield* bus.publish(SessionEvent.Forked, { + sessionID: childID, + parentID, + boundary: { type: "through", messageID: compaction.id }, + }) + const [child] = yield* store.context(childID) + expect(child).toMatchObject({ + type: "compaction", + metadata: { keepID: "not-a-msg-id" }, + }) + }), +) + const textUser = (text: string) => SessionMessage.User.make({ id: SessionMessage.ID.create(), diff --git a/packages/core/test/session-create.test.ts b/packages/core/test/session-create.test.ts index 78d51f411296..db72afb4e9bb 100644 --- a/packages/core/test/session-create.test.ts +++ b/packages/core/test/session-create.test.ts @@ -921,7 +921,7 @@ describe("Session.create", () => { const input = { id, location } const created = yield* session.create(input) - yield* db.update(SessionTable).set({ agent: "build" }).where(eq(SessionTable.id, id)).run().pipe(Effect.orDie) + yield* db.update(SessionTable).set({ agent: Agent.defaultID }).where(eq(SessionTable.id, id)).run().pipe(Effect.orDie) expect(yield* session.create(input)).toMatchObject({ id: created.id, agent: "build" }) }), diff --git a/packages/core/test/session-error.test.ts b/packages/core/test/session-error.test.ts index 19728224936c..fcf8fe210838 100644 --- a/packages/core/test/session-error.test.ts +++ b/packages/core/test/session-error.test.ts @@ -17,6 +17,7 @@ import { ToolFailure, HttpContext, } from "@opencode/ai" +import { RequestExecutor } from "@opencode/ai/route" import { Permission } from "@opencode/core/permission" import { ID } from "@opencode/core/model" import { ModelResolver } from "@opencode/core/model-resolver" @@ -32,14 +33,23 @@ describe("toSessionError", () => { expect(toSessionError(llm(new RateLimitError({ message: "rate", retryAfterMs: 123 })))).toEqual({ type: "provider.rate-limit", message: "rate", + retryAfterMs: 123, }) expect(toSessionError(llm(new AuthenticationError({ message: "auth" }))).type).toBe("provider.auth") - expect(toSessionError(llm(new QuotaExceededError({ message: "quota" }))).type).toBe("provider.quota") + expect(toSessionError(llm(new QuotaExceededError({ message: "quota", retryAfterMs: 456 })))).toEqual({ + type: "provider.quota", + message: "quota", + retryAfterMs: 456, + }) expect(toSessionError(llm(new ContentPolicyError({ message: "blocked" }))).type).toBe("provider.content-filter") expect( toSessionError(llm(new TransportError({ message: "transport", transport: "http", operation: "request" }))).type, ).toBe("provider.transport") - expect(toSessionError(llm(new ProviderInternalError({ message: "internal" }))).type).toBe("provider.internal") + expect(toSessionError(llm(new ProviderInternalError({ message: "internal", retryAfterMs: 789 })))).toEqual({ + type: "provider.internal", + message: "internal", + retryAfterMs: 789, + }) expect(toSessionError(llm(new InvalidProviderOutputError({ message: "output" }))).type).toBe( "provider.invalid-output", ) @@ -71,6 +81,55 @@ describe("toSessionError", () => { expect(toSessionError(llm(new UnknownProviderError({ message: "unknown" }))).type).toBe("provider.unknown") }) + test("separates the free usage limit from paid rate limits, Go quota, and other providers' 429s", () => { + const fail = (url: string, body: unknown, headers?: Record) => + RequestExecutor.httpFailure({ + message: "Rate limit exceeded. Please try again later.", + url, + status: 429, + responseHeaders: headers, + responseBody: JSON.stringify(body), + }) + + const freeBody = { + type: "error", + error: { type: "FreeUsageLimitError", message: "Rate limit exceeded. Please try again later." }, + } + const free = fail("https://opencode.ai/inference/openai/v1/chat/completions", freeBody, { "retry-after": "3600" }) + expect(SessionRunnerRetry.isRetryable(free)).toBe(false) + expect(toSessionError(free)).toEqual({ + type: "provider.free-limit", + message: "Free usage limit reached. Connect a provider or subscribe to OpenCode Go to keep going, or switch models.", + status: 429, + response: { body: JSON.stringify(freeBody) }, + retryAfterMs: 3_600_000, + }) + + const paid = fail( + "https://opencode.ai/inference/openai/v1/chat/completions", + { error: { type: "server_error", message: "Upstream request failed: Rate limit exceeded." } }, + { "retry-after": "45" }, + ) + expect(SessionRunnerRetry.isRetryable(paid)).toBe(true) + expect(toSessionError(paid)).toMatchObject({ type: "provider.rate-limit", retryAfterMs: 45_000 }) + + const go = fail( + "https://opencode.ai/inference/go/openai/v1/chat/completions", + { type: "error", error: { type: "GoUsageLimitError", message: "Subscription quota exceeded." } }, + { "retry-after": "120" }, + ) + expect(SessionRunnerRetry.isRetryable(go)).toBe(false) + expect(toSessionError(go)).toMatchObject({ type: "provider.quota", retryAfterMs: 120_000 }) + + const other = fail( + "https://api.openai.com/v1/chat/completions", + { error: { type: "rate_limit_error", message: "Too many requests" } }, + { "retry-after": "10" }, + ) + expect(SessionRunnerRetry.isRetryable(other)).toBe(true) + expect(toSessionError(other)).toMatchObject({ type: "provider.rate-limit", retryAfterMs: 10_000 }) + }) + test("preserves the permission rejection type without exposing internal fields", () => { const blocked = new Permission.BlockedError({ rules: [], permission: "external_directory", resources: [] }) expect(toSessionError(blocked)).toEqual({ diff --git a/packages/core/test/session-projector.test.ts b/packages/core/test/session-projector.test.ts index 5cc9a070a6ad..96f45aa7f51d 100644 --- a/packages/core/test/session-projector.test.ts +++ b/packages/core/test/session-projector.test.ts @@ -331,7 +331,7 @@ describe("SessionProjector", () => { it.effect("projects durable context messages supported by the updater", () => Effect.gen(function* () { - const db = yield* seedSession({ agent: "plan", model: previousModel }) + const db = yield* seedSession({ agent: Agent.ID.make("plan"), model: previousModel }) const bus = yield* Bus.Service yield* bus.publish(SessionEvent.AgentSelected, { diff --git a/packages/core/test/session-runner.test.ts b/packages/core/test/session-runner.test.ts index 6f6ad5c4a9df..7289ba02cb72 100644 --- a/packages/core/test/session-runner.test.ts +++ b/packages/core/test/session-runner.test.ts @@ -1067,8 +1067,10 @@ describe("SessionRunnerLLM", () => { yield* Fiber.join(renamed) expect(s.requests).toHaveLength(6) - expect(s.requests[2]?.messages).toContainEqual(Message.user("First prompt")) - expect(s.requests[4]?.messages).toContainEqual(Message.user("First prompt")) + expect(JSON.stringify(s.requests[2]?.messages)).toContain("First prompt") + expect(JSON.stringify(s.requests[2]?.messages)).not.toContain("Second prompt") + expect(JSON.stringify(s.requests[4]?.messages)).toContain("First prompt") + expect(JSON.stringify(s.requests[4]?.messages)).not.toContain("Third prompt") expect((yield* s.session.get(sessionID)).title).toBe("Generated title") }) @@ -1841,7 +1843,7 @@ describe("SessionRunnerLLM", () => { ) yield* s.db .update(SessionTable) - .set({ agent: "reviewer" }) + .set({ agent: Agent.ID.make("reviewer") }) .where(eq(SessionTable.id, sessionID)) .run() .pipe(Effect.orDie) @@ -1861,7 +1863,7 @@ describe("SessionRunnerLLM", () => { scenario("fails before the model request when the selected agent is unavailable", function* (s) { yield* s.db .update(SessionTable) - .set({ agent: "explore" }) + .set({ agent: Agent.ID.make("explore") }) .where(eq(SessionTable.id, sessionID)) .run() .pipe(Effect.orDie) diff --git a/packages/core/test/session-title.test.ts b/packages/core/test/session-title.test.ts index 82a1f4ee73a6..a508f97b4dce 100644 --- a/packages/core/test/session-title.test.ts +++ b/packages/core/test/session-title.test.ts @@ -39,6 +39,7 @@ import { testEffect } from "./lib/effect" let requests: LLMRequest[] = [] let selectedSmall: Model.Info | undefined +let selectedPrimary: Model.Info | undefined let selections: Array = [] const model = LanguageModel.make({ id: "title-model", @@ -102,6 +103,7 @@ const models = Layer.mock(SessionRunnerModel.Service)({ }) const smallModels = Layer.mock(Model.Service, { small: () => Effect.succeed(selectedSmall), + get: (_providerID, modelID) => Effect.succeed(modelID === selectedPrimary?.id ? selectedPrimary : undefined), }) const it = testEffect( AppNodeBuilder.build( @@ -183,6 +185,7 @@ const lowSmall = Model.Info.make({ beforeEach(() => { requests = [] selectedSmall = undefined + selectedPrimary = undefined selections = [] titleStream = successfulTitle }) @@ -343,6 +346,95 @@ it.effect("falls back to the primary model when the small model fails", () => }), ) +const primaryRef = (variant?: string) => + Model.Ref.make({ + providerID: Provider.ID.make("test"), + id: Model.ID.make("title-model"), + ...(variant ? { variant: Model.VariantID.make(variant) } : {}), + }) +const primaryInfo = (variants: string[]) => + Model.Info.make({ + ...Model.Info.default(Provider.ID.make("test"), Model.ID.make("title-model")), + variants: variants.map((id) => ({ id: Model.VariantID.make(id) })), + }) + +const titleAttempts = (sessionID: Session.ID, text: string, variant?: string) => + Effect.gen(function* () { + yield* enableTitleAgent + yield* insertSession(sessionID, undefined, undefined, primaryRef(variant)) + yield* prompt(sessionID, text) + const attempted: Model.Ref[] = [] + const hooks = yield* PluginHooks.Service + yield* hooks.register("session", "model.request", (event) => + Effect.sync(() => { + attempted.push(event.model) + }), + ) + const title = yield* SessionTitle.Service + yield* title.generate(sessionID) + const store = yield* SessionStore.Service + return { attempted: attempted.map((model) => String(model.variant)), title: (yield* store.get(sessionID))?.title } + }) + +it.effect("uses the primary model at its lowest variant without a small model", () => + Effect.gen(function* () { + selectedPrimary = primaryInfo(["none", "low", "high"]) + const result = yield* titleAttempts( + Session.ID.make("ses_title_primary_lowest"), + "Title with the lowest variant", + "high", + ) + expect(result.attempted).toEqual(["none"]) + expect(result.title).toBe("Generated Title") + }), +) + +it.effect("falls back to the selected primary variant when the lowest variant fails", () => + Effect.gen(function* () { + titleStream = () => + requests.length === 1 ? Stream.make(LLMEvent.providerError({ message: "Variant rejected" })) : successfulTitle() + selectedPrimary = primaryInfo(["low", "high"]) + const result = yield* titleAttempts( + Session.ID.make("ses_title_primary_fallback"), + "Fall back to the selection", + "high", + ) + expect(result.attempted).toEqual(["low", "high"]) + expect(result.title).toBe("Generated Title") + }), +) + +it.effect("uses the selected primary variant when no lower variant exists", () => + Effect.gen(function* () { + selectedPrimary = primaryInfo(["high", "max"]) + const result = yield* titleAttempts(Session.ID.make("ses_title_primary_selected"), "No lower variant here", "high") + expect(result.attempted).toEqual(["high"]) + expect(result.title).toBe("Generated Title") + }), +) + +it.effect("keeps the primary default variant instead of a low variant that may enable reasoning", () => + Effect.gen(function* () { + selectedPrimary = primaryInfo(["low", "high"]) + const result = yield* titleAttempts( + Session.ID.make("ses_title_primary_default"), + "Default may not reason", + undefined, + ) + expect(result.attempted).toEqual(["default"]) + expect(result.title).toBe("Generated Title") + }), +) + +it.effect("uses the primary none variant from its default variant", () => + Effect.gen(function* () { + selectedPrimary = primaryInfo(["none", "low", "high"]) + const result = yield* titleAttempts(Session.ID.make("ses_title_primary_none"), "Disable reasoning", undefined) + expect(result.attempted).toEqual(["none"]) + expect(result.title).toBe("Generated Title") + }), +) + it.effect("generates from the first user message after later messages exist", () => Effect.gen(function* () { yield* enableTitleAgent diff --git a/packages/core/test/skill/instructions.test.ts b/packages/core/test/skill/instructions.test.ts index 6af55cf3d0a5..5529955f273a 100644 --- a/packages/core/test/skill/instructions.test.ts +++ b/packages/core/test/skill/instructions.test.ts @@ -117,7 +117,7 @@ describe("SkillInstructions", () => { }).pipe(Effect.provide(layer(() => skills))) }) - it.effect("restates the full skill list when a description changes", () => { + it.effect("renders only the changed skill when a description changes", () => { const agent = Agent.Info.make(Agent.Info.default(build)) let skills = [effect] return Effect.gen(function* () { @@ -125,15 +125,19 @@ describe("SkillInstructions", () => { const initialized = yield* instructions.load(agent.permissions).pipe(Effect.flatMap(readInitial)) skills = [Skill.Info.make({ ...effect, description: "Build applications with Effect v4" })] - expect( - yield* instructions - .load(agent.permissions) - .pipe(Effect.flatMap((context) => readUpdate(context, initialized))), - ).toMatchObject({ - text: expect.stringContaining( - "The available skills have changed. This list supersedes the previous available skills list.", - ), - }) + const changed = yield* instructions + .load(agent.permissions) + .pipe(Effect.flatMap((context) => readUpdate(context, initialized))) + expect(changed.text).toBe( + [ + "The following skills have changed. Each entry replaces the previously listed skill with the same ID:", + " ", + " effect", + " Effect", + " Build applications with Effect v4", + " ", + ].join("\n"), + ) }).pipe(Effect.provide(layer(() => skills))) }) diff --git a/packages/core/test/tool-read.test.ts b/packages/core/test/tool-read.test.ts index 04ff8c522e86..4bf3af248d07 100644 --- a/packages/core/test/tool-read.test.ts +++ b/packages/core/test/tool-read.test.ts @@ -933,26 +933,4 @@ describe("ReadTool", () => { ]) }), ) - - it.effect("rejects unsupported binary discovered by a direct read", () => - Effect.gen(function* () { - readResult = { - type: "file", - uri: "file:///late-binary", - name: "late-binary", - content: "AAECAw==", - encoding: "base64", - mime: "application/octet-stream", - } - const registry = yield* Tool.Service - - expect( - yield* executeTool(registry, { - sessionID, - ...toolIdentity, - call: { type: "tool-call", id: "call-direct-binary", name: "read", input: { path: "late-binary" } }, - }), - ).toEqual({ status: "error", error: { type: "unknown", message: "Cannot read binary file: late-binary" } }) - }), - ) }) diff --git a/packages/core/test/v1-migration.test.ts b/packages/core/test/v1-migration.test.ts index 5ba6d5f95e54..e4e3f7d500a5 100644 --- a/packages/core/test/v1-migration.test.ts +++ b/packages/core/test/v1-migration.test.ts @@ -8,6 +8,7 @@ import { SessionMessage } from "@opencode/core/session/message" import { SessionSchema } from "@opencode/core/session/schema" import { SessionTable } from "@opencode/core/session/sql" import { Project } from "@opencode/core/project" +import { Agent } from "@opencode/schema/agent" import { ProjectTable } from "@opencode/core/project/sql" import { AbsolutePath } from "@opencode/core/schema" import { Global } from "@opencode/util/global" @@ -661,7 +662,7 @@ describe("V1Migration.transformSession", () => { const compact = user("msg_000000000027aaaaaaaaaaaaaa", {}, 3) const summary = assistant("msg_000000000028aaaaaaaaaaaaaa", compact.id, { summary: true }, 4) const existing = session({ - agent: "existing", + agent: Agent.ID.make("existing"), model: { id: "existing-model", providerID: "existing-provider", variant: "existing" }, }) const result = transform( @@ -680,7 +681,7 @@ describe("V1Migration.transformSession", () => { metadata: { recent: "[User]: question\n\n[Assistant]: answer" }, }) expect(result.messages[2].data).not.toHaveProperty("recent") - expect(result.session.agent).toBe("existing") + expect(result.session.agent).toBe(Agent.ID.make("existing")) expect(result.session.model).toEqual(existing.model) }) @@ -722,7 +723,7 @@ describe("V1Migration.transformSession", () => { }, ]) expect(result.session).toEqual({ - agent: "review", + agent: Agent.ID.make("review"), model: { id: "m2", providerID: "p2", variant: "default" }, cost: 7, tokens_input: 8, @@ -1120,13 +1121,46 @@ describe("V1Migration database workflow", () => { ) }) - test("reassigns V1 sessions whose projects are missing to the global project", async () => { + test.each([ + { worktree: "/repo", directory: ".", expected: "/repo" }, + { worktree: "/repo", directory: "src", expected: "/repo" }, + { worktree: "C:/repo", directory: ".", expected: "C:/repo" }, + { worktree: "//server/share/repo", directory: ".", expected: "//server/share/repo" }, + { worktree: "/repo", directory: "C:/other", expected: "C:/other" }, + { worktree: "/repo", directory: "//server/share/other", expected: "//server/share/other" }, + { worktree: "/repo", directory: "", expected: "" }, + { worktree: "relative", directory: ".", expected: "" }, + ])("migrates $directory in $worktree", async (input) => { + await database( + Effect.gen(function* () { + const { db } = yield* Database.Service + yield* db.run(sql` + INSERT INTO project (id, worktree, time_created, time_updated, sandboxes) + VALUES ('saved-project', ${input.worktree}, 1, 2, '[]') + `) + yield* db.run(sql` + INSERT INTO session (id, project_id, slug, directory, title, version, time_created, time_updated) + VALUES ('ses_test', 'saved-project', 'test', ${input.directory}, 'Test', '1', 1, 2) + `) + + expect(yield* V1Migration.run()).toEqual({ status: "completed" }) + expect(yield* db.get(sql`SELECT directory FROM session_v2 WHERE id = 'ses_test'`)).toEqual({ + directory: input.expected, + }) + expect(yield* db.get(sql`SELECT directory FROM session WHERE id = 'ses_test'`)).toEqual({ + directory: input.directory, + }) + }), + ) + }) + + test.each(["/tmp/orphan", "."])("reassigns V1 sessions with missing projects and directory %s", async (directory) => { await database( Effect.gen(function* () { const { db } = yield* Database.Service const global = yield* Global.Service yield* db.run( - sql`INSERT INTO session (id, project_id, slug, directory, title, version, time_created, time_updated) VALUES ('ses_orphan', 'missing-project', 'orphan', '/tmp/orphan', 'Orphan', '1', 1, 2)`, + sql`INSERT INTO session (id, project_id, slug, directory, title, version, time_created, time_updated) VALUES ('ses_orphan', 'missing-project', 'orphan', ${directory}, 'Orphan', '1', 1, 2)`, ) expect(yield* V1Migration.run()).toEqual({ status: "completed" }) @@ -1136,6 +1170,9 @@ describe("V1Migration database workflow", () => { expect(yield* db.get(sql`SELECT worktree FROM project WHERE id = 'global'`)).toEqual({ worktree: path.parse(global.data).root, }) + expect(yield* db.get(sql`SELECT directory FROM session_v2 WHERE id = 'ses_orphan'`)).toEqual({ + directory: directory === "." ? path.parse(global.data).root.replaceAll("\\", "/") : directory, + }) expect(yield* db.get(sql`SELECT value FROM kv WHERE key = 'migration.v1-v2'`)).toEqual({ value: '{"phase":"completed"}', }) diff --git a/packages/core/test/voice.test.ts b/packages/core/test/voice.test.ts index 14970f94bbbd..65c6f57a6e7b 100644 --- a/packages/core/test/voice.test.ts +++ b/packages/core/test/voice.test.ts @@ -24,6 +24,29 @@ const keys: Record = { GEMINI_API_KEY: "gemini-env-key", } +const unused = () => Effect.die("unused") + +const mockIntegration = ( + connections: (id: Integration.ID) => Integration.Info["connections"], + resolve: Integration.Interface["connection"]["resolve"] = () => Effect.die("Voice.models resolved a credential"), +) => + Layer.mock(Integration.Service, { + revision: () => 0, + get: (id) => Effect.succeed({ id, name: id, methods: [], connections: connections(id) }), + connection: { + active: unused, + resolve, + key: unused, + external: unused, + activate: unused, + update: unused, + remove: unused, + status: unused, + }, + oauth: { connect: unused, status: unused, complete: unused, cancel: unused }, + command: { connect: unused, status: unused, cancel: unused }, + }) + it.effect("uses the first key or env connection Integration lists and never a ChatGPT sign-in", () => Effect.gen(function* () { yield* TestClock.setTime(1760000000_000) @@ -42,39 +65,18 @@ it.effect("uses the first key or env connection Integration lists and never a Ch Voice.layer.pipe( Layer.provide( Layer.merge( - Layer.mock(Integration.Service, { - revision: () => 0, - get: (id) => Effect.succeed({ id, name: id, methods: [], connections }), - connection: { - active: () => Effect.die("unused"), - resolve: (connection) => - connection.type === "credential" && connection.method === "oauth" - ? Effect.die("a ChatGPT sign-in was resolved") - : Effect.succeed( - Credential.Key.make({ - type: "key", - key: keys[connection.type === "env" ? connection.name : connection.id], - }), - ), - key: () => Effect.die("unused"), - external: () => Effect.die("unused"), - activate: () => Effect.die("unused"), - update: () => Effect.die("unused"), - remove: () => Effect.die("unused"), - status: () => Effect.die("unused"), - }, - oauth: { - connect: () => Effect.die("unused"), - status: () => Effect.die("unused"), - complete: () => Effect.die("unused"), - cancel: () => Effect.die("unused"), - }, - command: { - connect: () => Effect.die("unused"), - status: () => Effect.die("unused"), - cancel: () => Effect.die("unused"), - }, - }), + mockIntegration( + () => connections, + (connection) => + connection.type === "credential" && connection.method === "oauth" + ? Effect.die("a ChatGPT sign-in was resolved") + : Effect.succeed( + Credential.Key.make({ + type: "key", + key: keys[connection.type === "env" ? connection.name : connection.id], + }), + ), + ), RequestExecutor.layer.pipe( Layer.provide( Layer.succeed( @@ -107,16 +109,20 @@ it.effect("uses the first key or env connection Integration lists and never a Ch const gemini = yield* call("gemini-3.8-live", [{ type: "env", name: "GEMINI_API_KEY" }]) expect(auths).toEqual(["Bearer sk-active", "Bearer sk-stored", "Bearer xai-env-key", "gemini-env-key"]) expect(grok).toEqual({ - type: "websocket", - url: "wss://api.x.ai/v1/realtime?model=grok-voice-3.0", - token: "ek_xai", - expiresAt: 1760000999, + transport: { + type: "websocket", + url: "wss://api.x.ai/v1/realtime?model=grok-voice-3.0", + token: "ek_xai", + expiresAt: 1760000999, + }, }) expect(gemini).toEqual({ - type: "websocket", - url: "wss://generativelanguage.googleapis.com/ws/google.ai.generativelanguage.v1beta.GenerativeService.BidiGenerateContentConstrained", - token: "auth_tokens/ek_gemini", - expiresAt: 1760000060, + transport: { + type: "websocket", + url: "wss://generativelanguage.googleapis.com/ws/google.ai.generativelanguage.v1beta.GenerativeService.BidiGenerateContentConstrained", + token: "auth_tokens/ek_gemini", + expiresAt: 1760000060, + }, }) const signInOnly = yield* call("gpt-live-1", [credential("chatgpt", "oauth")]).pipe(Effect.flip) @@ -128,3 +134,136 @@ it.effect("uses the first key or env connection Integration lists and never a Ch expect(missingGoogle.message).toContain("GEMINI_API_KEY") }), ) + +it.effect("lists voice models with availability and default without resolving credentials", () => + Effect.gen(function* () { + const list = (connections: Partial>) => + Effect.flatMap(Voice.Service, (voice) => voice.models()).pipe( + Effect.provide( + Voice.layer.pipe( + Layer.provide( + Layer.merge( + mockIntegration((id) => connections[id] ?? []), + Layer.mock(RequestExecutor.Service, {}), + ), + ), + ), + ), + ) + + expect( + yield* list({ + openai: [credential("chatgpt", "oauth")], + xai: [{ type: "env", name: "XAI_API_KEY" }], + google: [credential("gemini", "key")], + }), + ).toEqual({ + models: [ + { model: "gpt-live-1", provider: "openai", transports: ["webrtc"], available: false }, + { model: "gpt-realtime-2.1", provider: "openai", transports: ["webrtc", "websocket"], available: false }, + { model: "grok-voice-3.0", provider: "xai", transports: ["websocket"], available: true, integration: "xai" }, + { + model: "gemini-3.8-live", + provider: "google", + transports: ["websocket"], + available: true, + integration: "google", + }, + ], + default: "grok-voice-3.0", + }) + + const none = yield* list({ openai: [credential("chatgpt", "oauth")] }) + expect(none).toEqual({ + models: [ + { model: "gpt-live-1", provider: "openai", transports: ["webrtc"], available: false }, + { model: "gpt-realtime-2.1", provider: "openai", transports: ["webrtc", "websocket"], available: false }, + { model: "grok-voice-3.0", provider: "xai", transports: ["websocket"], available: false }, + { model: "gemini-3.8-live", provider: "google", transports: ["websocket"], available: false }, + ], + }) + expect("default" in none).toBe(false) + }), +) + +// An OpenAI key the server could use instead, and the requests that reach OpenAI with it. +const withStoredKey = (requests: string[]) => + Voice.layer.pipe( + Layer.provide( + Layer.merge( + mockIntegration( + (id) => (id === "openai" ? [credential("stored", "key")] : []), + () => Effect.succeed(Credential.Key.make({ type: "key", key: "sk-stored" })), + ), + RequestExecutor.layer.pipe( + Layer.provide( + Layer.succeed( + HttpClient.HttpClient, + HttpClient.make((req: HttpClientRequest.HttpClientRequest) => + Effect.sync(() => { + requests.push(req.headers.authorization ?? "") + return HttpClientResponse.fromWeb(req, Response.json({ transport: { sdp: "answer:openai" } })) + }), + ), + ), + ), + ), + ), + ), + ) + +const hostedLive = (instructed: Voice.InstructInput[]): Voice.Host => ({ + integration: "opencode", + models: ["gpt-live-1"], + workspace: "Anomaly", + session: (input) => + Effect.succeed({ + transport: { type: "webrtc", sdp: input.transport.type === "webrtc" ? `answer:hosted:${input.transport.sdp}` : "" }, + id: "sess_hosted", + }), + instruct: (input) => + Effect.sync(() => { + instructed.push(input) + }), +}) + +const offer = { model: "gpt-live-1", transport: { type: "webrtc", sdp: "v=0" } } as const + +it.effect("starts a host's models and relays their instructions on it, even beside a stored key", () => { + const requests: string[] = [] + return Effect.gen(function* () { + const voice = yield* Voice.Service + const instructed: Voice.InstructInput[] = [] + yield* voice.host(hostedLive(instructed)) + + expect(yield* voice.session(offer)).toEqual({ + transport: { type: "webrtc", sdp: "answer:hosted:v=0" }, + id: "sess_hosted", + }) + expect((yield* voice.models()).models.find((item) => item.model === "gpt-live-1")).toEqual({ + model: "gpt-live-1", + provider: "openai", + transports: ["webrtc"], + available: true, + integration: "opencode", + workspace: "Anomaly", + }) + yield* voice.instruct({ id: "sess_hosted", instructions: "The screen shows ses_1." }) + expect(instructed).toEqual([{ id: "sess_hosted", instructions: "The screen shows ses_1." }]) + expect(requests).toEqual([]) + }).pipe(Effect.provide(withStoredKey(requests))) +}) + +it.effect("goes back to the stored key once the host is removed, and refuses instructions", () => { + const requests: string[] = [] + return Effect.gen(function* () { + const voice = yield* Voice.Service + yield* voice.host(hostedLive([])) + yield* voice.host(undefined) + + expect(yield* voice.session(offer)).toEqual({ transport: { type: "webrtc", sdp: "answer:openai" } }) + expect(requests).toEqual(["Bearer sk-stored"]) + const refused = yield* voice.instruct({ id: "sess_hosted", instructions: "Hello" }).pipe(Effect.flip) + expect(refused._tag).toBe("Voice.InvalidRequestError") + }).pipe(Effect.provide(withStoredKey(requests))) +}) diff --git a/packages/core/test/worktree.test.ts b/packages/core/test/worktree.test.ts index 6d2869d442a2..ec033f151e25 100644 --- a/packages/core/test/worktree.test.ts +++ b/packages/core/test/worktree.test.ts @@ -136,7 +136,9 @@ function stored(projectID: Project.ID) { .all() .pipe( Effect.orDie, - Effect.map((rows) => rows.toSorted((a, b) => a.directory.localeCompare(b.directory))), + Effect.map((rows): Array<{ directory: AbsolutePath; strategy: string | null }> => + rows.toSorted((a, b) => a.directory.localeCompare(b.directory)), + ), ), ) } @@ -504,7 +506,11 @@ describe("Worktree", () => { yield* Effect.addFinalizer(() => Effect.promise(() => fs.rm(unavailable, { recursive: true, force: true }))) yield* input.db .insert(WorktreeTable) - .values({ project_id: input.projectID, directory: unavailable, strategy: "acme/missing" }) + .values({ + project_id: input.projectID, + directory: unavailable, + strategy: Worktree.StrategyID.make("acme/missing"), + }) .run() .pipe(Effect.orDie) @@ -612,7 +618,7 @@ describe("Worktree", () => { .insert(WorktreeTable) .values([ { project_id: input.projectID, directory: target }, - { project_id: input.projectID, directory: unchanged, strategy: "git" }, + { project_id: input.projectID, directory: unchanged, strategy: Worktree.StrategyID.make("git") }, ]) .run() .pipe(Effect.orDie) diff --git a/packages/http-recorder/package.json b/packages/http-recorder/package.json index 8c62779aa5c6..eb0390426b98 100644 --- a/packages/http-recorder/package.json +++ b/packages/http-recorder/package.json @@ -1,6 +1,6 @@ { "$schema": "https://json.schemastore.org/package.json", - "version": "2.1.0", + "version": "2.1.2", "name": "@opencode/http-recorder", "description": "Record and replay Effect HTTP and WebSocket traffic with deterministic cassettes", "type": "module", diff --git a/packages/httpapi-codegen/package.json b/packages/httpapi-codegen/package.json index 820c78d0cf6b..3b8976106c63 100644 --- a/packages/httpapi-codegen/package.json +++ b/packages/httpapi-codegen/package.json @@ -1,7 +1,7 @@ { "$schema": "https://json.schemastore.org/package.json", "name": "@opencode/httpapi-codegen", - "version": "2.1.0", + "version": "2.1.2", "private": true, "type": "module", "exports": { diff --git a/packages/latex/package.json b/packages/latex/package.json index c1d7de21adda..97a0b976139e 100644 --- a/packages/latex/package.json +++ b/packages/latex/package.json @@ -1,7 +1,7 @@ { "$schema": "https://json.schemastore.org/package.json", "name": "@opencode/latex", - "version": "2.1.0", + "version": "2.1.2", "private": true, "type": "module", "exports": { diff --git a/packages/merman/package.json b/packages/merman/package.json index 4419a24362da..948ede74c406 100644 --- a/packages/merman/package.json +++ b/packages/merman/package.json @@ -1,7 +1,7 @@ { "$schema": "https://json.schemastore.org/package.json", "name": "@opencode/merman", - "version": "2.1.0", + "version": "2.1.2", "private": true, "type": "module", "exports": { diff --git a/packages/plugin-browser/package.json b/packages/plugin-browser/package.json index 6d3302eda712..7780f6cf4c73 100644 --- a/packages/plugin-browser/package.json +++ b/packages/plugin-browser/package.json @@ -1,7 +1,7 @@ { "$schema": "https://json.schemastore.org/package.json", "name": "@opencode/plugin-browser", - "version": "2.1.0", + "version": "2.1.2", "description": "OpenCode's desktop browser plugin", "type": "module", "license": "MIT", diff --git a/packages/plugin/package.json b/packages/plugin/package.json index 1723f248b633..0cdb32cf24f8 100644 --- a/packages/plugin/package.json +++ b/packages/plugin/package.json @@ -1,7 +1,7 @@ { "$schema": "https://json.schemastore.org/package.json", "name": "@opencode/plugin", - "version": "2.1.0", + "version": "2.1.2", "type": "module", "license": "MIT", "scripts": { diff --git a/packages/plugins/figma/package.json b/packages/plugins/figma/package.json index b0fa0e943989..d5944206bff9 100644 --- a/packages/plugins/figma/package.json +++ b/packages/plugins/figma/package.json @@ -1,7 +1,7 @@ { "$schema": "https://json.schemastore.org/package.json", "name": "@opencode/plugin-figma", - "version": "2.1.0", + "version": "2.1.2", "description": "Figma MCP integration and official Figma skills for OpenCode", "type": "module", "license": "MIT", @@ -34,7 +34,7 @@ }, "devDependencies": { "@tsconfig/bun": "1.0.9", - "@types/bun": "1.4.0", + "@types/bun": "1.4.3", "@typescript/native": "npm:typescript@7.0.2" } } diff --git a/packages/protocol/openapi.json b/packages/protocol/openapi.json index 829c7e20c1f5..1c68daeaadd1 100644 --- a/packages/protocol/openapi.json +++ b/packages/protocol/openapi.json @@ -5382,6 +5382,68 @@ } } }, + "/api/experimental/voice/models": { + "get": { + "tags": ["voice"], + "operationId": "experimental.voice.models", + "parameters": [], + "security": [], + "responses": { + "200": { + "description": "VoiceModelsResponse", + "content": { + "application/json": { + "schema": { + "$ref": "#/components/schemas/VoiceModelsResponse" + } + } + } + }, + "400": { + "description": "InvalidRequestError", + "content": { + "application/json": { + "schema": { + "$ref": "#/components/schemas/InvalidRequestErrorEncoded" + } + } + } + }, + "401": { + "description": "UnauthorizedError", + "content": { + "application/json": { + "schema": { + "$ref": "#/components/schemas/UnauthorizedErrorEncoded" + } + } + } + }, + "404": { + "description": "LocationNotFoundError", + "content": { + "application/json": { + "schema": { + "$ref": "#/components/schemas/LocationNotFoundErrorEncoded" + } + } + } + }, + "503": { + "description": "ServiceUnavailableError", + "content": { + "application/json": { + "schema": { + "$ref": "#/components/schemas/ServiceUnavailableErrorEncoded" + } + } + } + } + }, + "description": "List supported voice models, their transports, whether the server has a usable API key for each provider or the connected OpenCode Console hosts the model, and the Console workspace a hosted model bills to.", + "summary": "List voice models" + } + }, "/api/experimental/voice/session": { "post": { "tags": ["voice"], @@ -5440,7 +5502,7 @@ } } }, - "description": "Create a server-brokered voice session using the server's configured provider credentials and return the WebRTC SDP answer or WebSocket ephemeral token.", + "description": "Create a server-brokered voice session using the server's configured provider credentials, or the connected OpenCode Console when it hosts the model, and return the WebRTC SDP answer or WebSocket ephemeral token.", "summary": "Create voice session", "requestBody": { "content": { @@ -5474,6 +5536,87 @@ } } }, + "/api/experimental/voice/session/{sessionID}/instructions": { + "post": { + "tags": ["voice"], + "operationId": "experimental.voice.instructions", + "parameters": [ + { + "name": "sessionID", + "in": "path", + "schema": { + "type": "string" + }, + "required": true + } + ], + "security": [], + "responses": { + "204": { + "description": "" + }, + "400": { + "description": "InvalidRequestError", + "content": { + "application/json": { + "schema": { + "$ref": "#/components/schemas/InvalidRequestErrorEncoded" + } + } + } + }, + "401": { + "description": "UnauthorizedError", + "content": { + "application/json": { + "schema": { + "$ref": "#/components/schemas/UnauthorizedErrorEncoded" + } + } + } + }, + "404": { + "description": "LocationNotFoundError", + "content": { + "application/json": { + "schema": { + "$ref": "#/components/schemas/LocationNotFoundErrorEncoded" + } + } + } + }, + "503": { + "description": "ServiceUnavailableError", + "content": { + "application/json": { + "schema": { + "$ref": "#/components/schemas/ServiceUnavailableErrorEncoded" + } + } + } + } + }, + "description": "Replace the backend instructions of a running session the connected OpenCode Console hosts, whose data channel can't change them.", + "summary": "Replace voice session instructions", + "requestBody": { + "content": { + "application/json": { + "schema": { + "type": "object", + "properties": { + "instructions": { + "type": "string" + } + }, + "required": ["instructions"], + "additionalProperties": false + } + } + }, + "required": true + } + } + }, "/api/provider": { "get": { "tags": ["provider"], @@ -14324,6 +14467,17 @@ "type": "integer", "minimum": 0 }, + "location_idle_timeout": { + "anyOf": [ + { + "type": "string" + }, + { + "type": "boolean", + "enum": [false] + } + ] + }, "policies": { "type": "array", "items": { @@ -19794,6 +19948,10 @@ }, "required": ["body"], "additionalProperties": false + }, + "retryAfterMs": { + "type": "integer", + "minimum": 0 } }, "required": ["type", "message"], @@ -20525,6 +20683,78 @@ "required": ["_tag", "providerID", "message"], "additionalProperties": false }, + "VoiceModel": { + "type": "object", + "properties": { + "model": { + "type": "string" + }, + "provider": { + "type": "string" + }, + "transports": { + "type": "array", + "items": { + "$ref": "#/components/schemas/VoiceTransportType" + } + }, + "available": { + "type": "boolean" + }, + "integration": { + "anyOf": [ + { + "type": "string" + }, + { + "type": "null" + } + ] + }, + "workspace": { + "anyOf": [ + { + "type": "string" + }, + { + "type": "null" + } + ] + } + }, + "required": ["model", "provider", "transports", "available"], + "additionalProperties": false + }, + "VoiceModelsResponse": { + "type": "object", + "properties": { + "data": { + "type": "object", + "properties": { + "models": { + "type": "array", + "items": { + "$ref": "#/components/schemas/VoiceModel" + } + }, + "default": { + "anyOf": [ + { + "type": "string" + }, + { + "type": "null" + } + ] + } + }, + "required": ["models"], + "additionalProperties": false + } + }, + "required": ["data"], + "additionalProperties": false + }, "VoiceSessionResponse": { "type": "object", "properties": { @@ -20533,6 +20763,23 @@ "properties": { "transport": { "$ref": "#/components/schemas/VoiceTransport" + }, + "session": { + "anyOf": [ + { + "type": "object", + "properties": { + "id": { + "type": "string" + } + }, + "required": ["id"], + "additionalProperties": false + }, + { + "type": "null" + } + ] } }, "required": ["transport"], @@ -20562,6 +20809,10 @@ } ] }, + "VoiceTransportType": { + "type": "string", + "enum": ["webrtc", "websocket"] + }, "VoiceWebRTCTransport": { "type": "object", "properties": { diff --git a/packages/protocol/package.json b/packages/protocol/package.json index 46e10ffa4503..84349b51fa3f 100644 --- a/packages/protocol/package.json +++ b/packages/protocol/package.json @@ -1,7 +1,7 @@ { "$schema": "https://json.schemastore.org/package.json", "name": "@opencode/protocol", - "version": "2.1.0", + "version": "2.1.2", "type": "module", "license": "MIT", "repository": { diff --git a/packages/protocol/src/groups/voice.ts b/packages/protocol/src/groups/voice.ts index 30e150bcd96a..ecebc8d9bebb 100644 --- a/packages/protocol/src/groups/voice.ts +++ b/packages/protocol/src/groups/voice.ts @@ -1,7 +1,22 @@ import { Schema } from "effect" -import { HttpApiEndpoint, HttpApiGroup, OpenApi } from "effect/http-api" +import { HttpApiEndpoint, HttpApiGroup, HttpApiSchema, OpenApi } from "effect/http-api" import { InvalidRequestError, LocationNotFoundError, ServiceUnavailableError } from "../errors.js" +export const VoiceTransportType = Schema.Literals(["webrtc", "websocket"]).annotate({ + identifier: "VoiceTransportType", +}) + +export const VoiceModel = Schema.Struct({ + model: Schema.String, + provider: Schema.String, + transports: Schema.Array(VoiceTransportType), + available: Schema.Boolean, + /** The integration whose connection starts its sessions: the provider's stored key, or `opencode` when the connected OpenCode Console hosts it. */ + integration: Schema.optional(Schema.String), + /** The OpenCode Console workspace a hosted model's sessions bill to. */ + workspace: Schema.optional(Schema.String), +}).annotate({ identifier: "VoiceModel" }) + export const VoiceWebRTCTransportInput = Schema.Struct({ type: Schema.Literal("webrtc"), sdp: Schema.String, @@ -35,6 +50,24 @@ export const VoiceTransport = Schema.Union([ ]).annotate({ identifier: "VoiceTransport" }) export const VoiceGroup = HttpApiGroup.make("server.voice") + .add( + HttpApiEndpoint.get("voice.models", "/api/experimental/voice/models", { + success: Schema.Struct({ + data: Schema.Struct({ + models: Schema.Array(VoiceModel), + default: Schema.optional(Schema.String), + }), + }).annotate({ identifier: "VoiceModelsResponse" }), + error: [ServiceUnavailableError, LocationNotFoundError], + }).annotateMerge( + OpenApi.annotations({ + identifier: "experimental.voice.models", + summary: "List voice models", + description: + "List supported voice models, their transports, whether the server has a usable API key for each provider or the connected OpenCode Console hosts the model, and the Console workspace a hosted model bills to.", + }), + ), + ) .add( HttpApiEndpoint.post("voice.session", "/api/experimental/voice/session", { payload: Schema.Struct({ @@ -45,6 +78,8 @@ export const VoiceGroup = HttpApiGroup.make("server.voice") success: Schema.Struct({ data: Schema.Struct({ transport: VoiceTransport, + /** A hosted session's ID, for `voice.instructions`. */ + session: Schema.optional(Schema.Struct({ id: Schema.String })), }), }).annotate({ identifier: "VoiceSessionResponse" }), error: [InvalidRequestError, ServiceUnavailableError, LocationNotFoundError], @@ -53,7 +88,22 @@ export const VoiceGroup = HttpApiGroup.make("server.voice") identifier: "experimental.voice.session", summary: "Create voice session", description: - "Create a server-brokered voice session using the server's configured provider credentials and return the WebRTC SDP answer or WebSocket ephemeral token.", + "Create a server-brokered voice session using the server's configured provider credentials, or the connected OpenCode Console when it hosts the model, and return the WebRTC SDP answer or WebSocket ephemeral token.", + }), + ), + ) + .add( + HttpApiEndpoint.post("voice.instructions", "/api/experimental/voice/session/:sessionID/instructions", { + params: { sessionID: Schema.String }, + payload: Schema.Struct({ instructions: Schema.String }), + success: HttpApiSchema.NoContent, + error: [InvalidRequestError, ServiceUnavailableError, LocationNotFoundError], + }).annotateMerge( + OpenApi.annotations({ + identifier: "experimental.voice.instructions", + summary: "Replace voice session instructions", + description: + "Replace the backend instructions of a running session the connected OpenCode Console hosts, whose data channel can't change them.", }), ), ) diff --git a/packages/schema/package.json b/packages/schema/package.json index c492c9540a6b..9fde8be0f04e 100644 --- a/packages/schema/package.json +++ b/packages/schema/package.json @@ -1,7 +1,7 @@ { "$schema": "https://json.schemastore.org/package.json", "name": "@opencode/schema", - "version": "2.1.0", + "version": "2.1.2", "type": "module", "license": "MIT", "repository": { diff --git a/packages/schema/src/config/agent.ts b/packages/schema/src/config/agent.ts index c28933251b58..ade684cd3b41 100644 --- a/packages/schema/src/config/agent.ts +++ b/packages/schema/src/config/agent.ts @@ -1,6 +1,6 @@ export * as ConfigAgent from "./agent.js" -import { Schema } from "effect" +import { Effect, Schema } from "effect" import { Permission } from "../permission.js" import { optional, PositiveInt } from "../schema.js" import { ConfigModel } from "./model.js" @@ -15,7 +15,11 @@ export class Info extends Schema.Class("Config.Agent")({ description: Schema.String.pipe(optional), mode: Schema.Literals(["subagent", "primary", "all"]).pipe(optional), hidden: Schema.Boolean.pipe(optional), - color: Color.pipe(optional), + // An unsupported color, such as a V1 theme name, should not drop the whole agent. + color: Color.pipe( + Schema.catchDecoding(() => Effect.succeedNone), + optional, + ), steps: PositiveInt.pipe(optional), disabled: Schema.Boolean.pipe(optional), permissions: Permission.Ruleset.pipe(optional), diff --git a/packages/schema/src/config/experimental.ts b/packages/schema/src/config/experimental.ts index 515cc77db074..4b378aeb1a60 100644 --- a/packages/schema/src/config/experimental.ts +++ b/packages/schema/src/config/experimental.ts @@ -14,6 +14,12 @@ export class Info extends Schema.Class("ConfigExperimental.Info")({ subagent_depth: NonNegativeInt.pipe(optional).annotate({ description: "Maximum subagent nesting depth. Defaults to 1.", }), + location_idle_timeout: Schema.Union([Schema.DurationFromString, Schema.Literal(false)]) + .pipe(optional) + .annotate({ + description: + 'How long a location can go without session activity before its services are unloaded. Running sessions and background jobs keep a location loaded. Set to false to never unload (default: "60 minutes")', + }), policies: ConfigPolicy.Info.pipe(Schema.Array, optional).annotate({ description: "Deprecated: use top-level policies. Ordered policies controlling access to configured resources", }), diff --git a/packages/schema/src/session-error.ts b/packages/schema/src/session-error.ts index e9bb70570fe3..b1a29226ac49 100644 --- a/packages/schema/src/session-error.ts +++ b/packages/schema/src/session-error.ts @@ -1,7 +1,7 @@ export * as SessionError from "./session-error.js" import { Schema } from "effect" -import { optional } from "./schema.js" +import { NonNegativeInt, optional } from "./schema.js" export interface Error extends Schema.Schema.Type {} export const Error = Schema.Struct({ @@ -9,4 +9,6 @@ export const Error = Schema.Struct({ message: Schema.String, status: Schema.Int.check(Schema.isBetween({ minimum: 100, maximum: 599 })).pipe(optional), response: Schema.Struct({ body: Schema.String }).pipe(optional), + /** How long the provider asked to wait before trying again, counted from when the failure was recorded. */ + retryAfterMs: NonNegativeInt.pipe(optional), }).annotate({ identifier: "Session.StructuredError" }) diff --git a/packages/schema/src/session-event.ts b/packages/schema/src/session-event.ts index cf26e0543aeb..c800a0235ff6 100644 --- a/packages/schema/src/session-event.ts +++ b/packages/schema/src/session-event.ts @@ -725,8 +725,8 @@ export const EphemeralDefinitions = Event.inventory( ) export const Durable = Schema.Union(DurableDefinitions, { mode: "oneOf" }) - .pipe(Schema.toTaggedUnion("type")) .annotate({ identifier: "Session.Event.Durable" }) + .pipe(Schema.toTaggedUnion("type")) export type DurableEvent = typeof Durable.Type export const All = Schema.Union([Durable, ...EphemeralDefinitions], { mode: "oneOf" }).pipe( diff --git a/packages/schema/test/session-error.test.ts b/packages/schema/test/session-error.test.ts index 9eeac44dab18..dfd1dfb55ef9 100644 --- a/packages/schema/test/session-error.test.ts +++ b/packages/schema/test/session-error.test.ts @@ -13,6 +13,7 @@ describe("SessionError", () => { { type: "provider.rate-limit", message: "Slow down" }, { type: "provider.rate-limit", message: "Slow down", status: 429, response: { body: '{"error":{}}' } }, { type: "provider.auth", message: "Authentication failed" }, + { type: "provider.free-limit", message: "Free usage limit reached", status: 429, retryAfterMs: 3_600_000 }, { type: "provider.future-condition", message: "A future provider failure" }, { type: "unknown", message: "Unexpected" }, ] @@ -29,7 +30,7 @@ describe("SessionError", () => { Schema.decodeUnknownSync(SessionError.Error)({ type: "provider.timeout", message: "Timeout", - retryAfterMs: 2_500, + retryable: true, }), ).toEqual({ type: "provider.timeout", message: "Timeout" }) }) diff --git a/packages/sdk/package.json b/packages/sdk/package.json index 7f58856913b7..16fa9d5a059f 100644 --- a/packages/sdk/package.json +++ b/packages/sdk/package.json @@ -1,6 +1,6 @@ { "$schema": "https://json.schemastore.org/package.json", - "version": "2.1.0", + "version": "2.1.2", "name": "@opencode/sdk", "type": "module", "license": "MIT", diff --git a/packages/server/package.json b/packages/server/package.json index 3bba41a6226b..b5669dc525a8 100644 --- a/packages/server/package.json +++ b/packages/server/package.json @@ -1,7 +1,7 @@ { "$schema": "https://json.schemastore.org/package.json", "name": "@opencode/server", - "version": "2.1.0", + "version": "2.1.2", "type": "module", "license": "MIT", "repository": { diff --git a/packages/server/src/handlers/voice.ts b/packages/server/src/handlers/voice.ts index afdcb21113dc..bb5e3f128545 100644 --- a/packages/server/src/handlers/voice.ts +++ b/packages/server/src/handlers/voice.ts @@ -9,31 +9,57 @@ import { HttpApiBuilder } from "effect/http-api" import { Api } from "../api" import { awaitPlugins, locationErrors } from "../location" +const voiceErrors = (error: Voice.Error) => + error._tag === "Voice.InvalidRequestError" + ? new InvalidRequestError({ message: error.message }) + : new ServiceUnavailableError({ message: error.message, service: error.service }) + export const VoiceHandler = HttpApiBuilder.group(Api, "server.voice", (handlers) => Effect.gen(function* () { const global = yield* Global.Service const locations = yield* LocationServiceMap.Service const services = locations.get(Location.Ref.make({ directory: AbsolutePath.make(global.config) })) - return handlers.handle( - "voice.session", - Effect.fn("server.voice.session")( - function* (request) { - yield* awaitPlugins - const voice = yield* Voice.Service - const transport = yield* voice - .session(request.payload) - .pipe( - Effect.mapError((error) => - error._tag === "Voice.InvalidRequestError" - ? new InvalidRequestError({ message: error.message }) - : new ServiceUnavailableError({ message: error.message, service: error.service }), - ), - ) - return { data: { transport } } - }, - Effect.provide(services), - locationErrors, - ), - ) + return handlers + .handle( + "voice.models", + Effect.fn("server.voice.models")( + function* () { + yield* awaitPlugins + const voice = yield* Voice.Service + return { data: yield* voice.models() } + }, + Effect.provide(services), + locationErrors, + ), + ) + .handle( + "voice.session", + Effect.fn("server.voice.session")( + function* (request) { + yield* awaitPlugins + const voice = yield* Voice.Service + const session = yield* voice.session(request.payload).pipe(Effect.mapError(voiceErrors)) + return { + data: { transport: session.transport, ...(session.id === undefined ? {} : { session: { id: session.id } }) }, + } + }, + Effect.provide(services), + locationErrors, + ), + ) + .handle( + "voice.instructions", + Effect.fn("server.voice.instructions")( + function* (request) { + yield* awaitPlugins + const voice = yield* Voice.Service + yield* voice + .instruct({ id: request.params.sessionID, instructions: request.payload.instructions }) + .pipe(Effect.mapError(voiceErrors)) + }, + Effect.provide(services), + locationErrors, + ), + ) }), ) diff --git a/packages/server/test/voice.test.ts b/packages/server/test/voice.test.ts index eda31544a104..bcb8196b2672 100644 --- a/packages/server/test/voice.test.ts +++ b/packages/server/test/voice.test.ts @@ -13,6 +13,23 @@ const voice = makeLocationNode({ layer: Layer.succeed( Voice.Service, Voice.Service.of({ + models: () => + Effect.succeed({ + models: [ + { + model: "gpt-live-1", + provider: "openai", + transports: ["webrtc"], + available: true, + integration: "opencode", + workspace: "Anomaly", + }, + { model: "gpt-realtime-2.1", provider: "openai", transports: ["webrtc", "websocket"], available: true }, + { model: "grok-voice-3.0", provider: "xai", transports: ["websocket"], available: false }, + { model: "gemini-3.8-live", provider: "google", transports: ["websocket"], available: false }, + ], + default: "gpt-live-1", + }), session: (input) => { if (input.model === "invalid") return Effect.fail(new Voice.InvalidRequestError({ message: "Unsupported voice model" })) @@ -20,21 +37,28 @@ const voice = makeLocationNode({ return Effect.fail(new Voice.UnavailableError({ message: "OpenAI API key is required", service: "openai" })) return Effect.succeed( input.transport.type === "webrtc" - ? { type: "webrtc" as const, sdp: `answer:${input.model}:${input.transport.sdp}` } + ? { transport: { type: "webrtc" as const, sdp: `answer:${input.model}:${input.transport.sdp}` }, id: "sess_1" } : { - type: "websocket" as const, - url: `wss://api.x.ai/v1/realtime?model=${input.model}`, - token: "ek_test", - expiresAt: 1760000000, + transport: { + type: "websocket" as const, + url: `wss://api.x.ai/v1/realtime?model=${input.model}`, + token: "ek_test", + expiresAt: 1760000000, + }, }, ) }, + instruct: (input) => + input.id === "sess_1" + ? Effect.void + : Effect.fail(new Voice.InvalidRequestError({ message: "Only hosted voice sessions take instructions" })), + host: () => Effect.void, }), ), deps: [], }) -it.live("routes POST /api/experimental/voice/session and maps errors", () => +it.live("routes experimental voice models and session endpoints", () => Effect.gen(function* () { const tmp = yield* Effect.acquireDisposable(Effect.promise(() => tmpdir("opencode-voice-endpoint-"))) const global = path.join(tmp.path, "global") @@ -48,6 +72,29 @@ it.live("routes POST /api/experimental/voice/session and maps errors", () => }, { overrides: [Voice.node.replace(voice)] }, ) + const modelsRes = yield* Effect.promise(() => + handler(new Request("http://opencode.local/api/experimental/voice/models")), + ) + expect(modelsRes.status).toBe(200) + expect(yield* Effect.promise(() => modelsRes.json())).toEqual({ + data: { + models: [ + { + model: "gpt-live-1", + provider: "openai", + transports: ["webrtc"], + available: true, + integration: "opencode", + workspace: "Anomaly", + }, + { model: "gpt-realtime-2.1", provider: "openai", transports: ["webrtc", "websocket"], available: true }, + { model: "grok-voice-3.0", provider: "xai", transports: ["websocket"], available: false }, + { model: "gemini-3.8-live", provider: "google", transports: ["websocket"], available: false }, + ], + default: "gpt-live-1", + }, + }) + const post = (body: unknown) => Effect.promise(() => handler( @@ -66,7 +113,7 @@ it.live("routes POST /api/experimental/voice/session and maps errors", () => }) expect(rtcRes.status).toBe(200) expect(yield* Effect.promise(() => rtcRes.json())).toEqual({ - data: { transport: { type: "webrtc", sdp: "answer:gpt-live-1:v=0" } }, + data: { transport: { type: "webrtc", sdp: "answer:gpt-live-1:v=0" }, session: { id: "sess_1" } }, }) const wsRes = yield* post({ model: "grok-voice-3.0", transport: { type: "websocket" } }) @@ -84,5 +131,18 @@ it.live("routes POST /api/experimental/voice/session and maps errors", () => expect((yield* post({ model: "invalid", transport: { type: "websocket" } })).status).toBe(400) expect((yield* post({ model: "unavailable", transport: { type: "websocket" } })).status).toBe(503) + + const instruct = (id: string) => + Effect.promise(() => + handler( + new Request(`http://opencode.local/api/experimental/voice/session/${id}/instructions`, { + method: "POST", + headers: { "content-type": "application/json" }, + body: JSON.stringify({ instructions: "The screen shows ses_1." }), + }), + ), + ) + expect((yield* instruct("sess_1")).status).toBe(204) + expect((yield* instruct("sess_unknown")).status).toBe(400) }), ) diff --git a/packages/simulation/package.json b/packages/simulation/package.json index 606425100285..7dbed546a634 100644 --- a/packages/simulation/package.json +++ b/packages/simulation/package.json @@ -1,7 +1,7 @@ { "$schema": "https://json.schemastore.org/package.json", "name": "@opencode/simulation", - "version": "2.1.0", + "version": "2.1.2", "type": "module", "license": "MIT", "repository": { diff --git a/packages/theme/package.json b/packages/theme/package.json index 0970dbc6b4e3..74dbbea7301f 100644 --- a/packages/theme/package.json +++ b/packages/theme/package.json @@ -1,7 +1,7 @@ { "$schema": "https://json.schemastore.org/package.json", "name": "@opencode/theme", - "version": "2.1.0", + "version": "2.1.2", "type": "module", "license": "MIT", "repository": { diff --git a/packages/tui/package.json b/packages/tui/package.json index 997f3f4ce284..de873e3474c8 100644 --- a/packages/tui/package.json +++ b/packages/tui/package.json @@ -1,7 +1,7 @@ { "$schema": "https://json.schemastore.org/package.json", "name": "@opencode/tui", - "version": "2.1.0", + "version": "2.1.2", "private": true, "type": "module", "license": "MIT", diff --git a/packages/tui/src/component/dialog-go-upsell.tsx b/packages/tui/src/component/dialog-go-upsell.tsx new file mode 100644 index 000000000000..025c7d6cb804 --- /dev/null +++ b/packages/tui/src/component/dialog-go-upsell.tsx @@ -0,0 +1,461 @@ +import { RGBA, TextAttributes, type Renderable } from "@opentui/core" +import { useRenderer, useTerminalDimensions } from "@opentui/solid" +import type { FormAnswer, IntegrationOauthConnectOutput, IntegrationOAuthMethod } from "@opencode/client" +import * as Open from "@opencode/util/open" +import { createEffect, createMemo, createSignal, For, onCleanup, Show } from "solid-js" +import { useConfig, useNewDesigns } from "../config" +import { useClient } from "../context/client" +import { useClipboard } from "../context/clipboard" +import { useData } from "../context/data" +import { Keymap } from "../context/keymap" +import { useLocal } from "../context/local" +import { useLocation } from "../context/location" +import { useTheme } from "../context/theme" +import { tint } from "../theme/color" +import { DialogFooter, DialogHeader, DialogLayout, useDialog } from "../ui/dialog" +import { Link } from "../ui/link" +import { useToast } from "../ui/toast" +import { errorMessage } from "../util/error" +import { credentialConnections } from "./dialog-integration" +import { DialogModel } from "./dialog-model" +import "./go-upsell-art" + +export const GO_URL = "https://opencode.ai/go" +const ACTIONS = ["don't show again", "dismiss", "subscribe"] as const +type Action = (typeof ACTIONS)[number] +type IntegrationAttempt = IntegrationOauthConnectOutput["data"] +const OAUTH_CODE = /[A-Z0-9]{4}-[A-Z0-9]{4,5}/ + +let inflightConnect: Promise = Promise.resolve() +function trackConnect(task: Promise) { + const next = task.catch(() => undefined) + inflightConnect = inflightConnect.then(() => next) +} + +export function waitForGoUpsellConnect() { + return inflightConnect +} + +export function DialogGoUpsell(props: { + reset?: string + canContinue?: () => boolean + onSubscribe: () => void + onDontShowAgain: () => void +}) { + const dialog = useDialog() + const dimensions = useTerminalDimensions() + const theme = useTheme().surface("dialog") + const config = useConfig().data + const renderer = useRenderer() + const narrow = () => dimensions().width < 52 + const short = () => dimensions().height < 22 + const compact = () => short() || dimensions().height < (narrow() ? 26 : 22) + const animate = () => (config.animations ?? true) && !compact() + const [selected, setSelected] = createSignal("subscribe") + const [anchor, setAnchor] = createSignal() + // Text sits on a translucent panel so the rings stay legible behind it. + const textBg = () => { + const [r, g, b] = theme.background.base.toInts() + return RGBA.fromInts(r, g, b, 186) + } + const primary = () => theme.background.action.primary.focused + + createEffect(() => { + if (!short()) return + dialog.setCentered(true) + onCleanup(() => { + if (dialog.stack.length > 0) dialog.setCentered(false) + }) + }) + + // Every cell changes each frame, so hold the renderer to 30 fps while the art plays. + createEffect(() => { + if (!animate()) return + const fps = { target: renderer.targetFps, max: renderer.maxFps } + renderer.targetFps = 30 + renderer.maxFps = 30 + onCleanup(() => { + renderer.targetFps = fps.target + renderer.maxFps = fps.max + }) + }) + + const run = (action: Action) => { + if (action === "subscribe") { + Open.openUrl(GO_URL).catch(() => {}) + dialog.replace(() => ) + return + } + if (action === "don't show again") props.onDontShowAgain() + dialog.clear() + } + const step = (by: number) => + setSelected((value) => ACTIONS[(ACTIONS.indexOf(value) + by + ACTIONS.length) % ACTIONS.length]!) + + Keymap.createLayer(() => ({ + mode: "modal", + commands: [ + { bind: "return", title: "Confirm Go option", group: "Dialog", run: () => run(selected()) }, + { bind: "left", title: "Previous Go option", group: "Dialog", run: () => step(-1) }, + { bind: "up", title: "Previous Go option", group: "Dialog", run: () => step(-1) }, + { bind: "shift+tab", title: "Previous Go option", group: "Dialog", run: () => step(-1) }, + { bind: "right", title: "Next Go option", group: "Dialog", run: () => step(1) }, + { bind: "down", title: "Next Go option", group: "Dialog", run: () => step(1) }, + { bind: "tab", title: "Next Go option", group: "Dialog", run: () => step(1) }, + ], + })) + + const button = (action: Action) => ( + setSelected(action)} + onMouseUp={() => run(action)} + > + + {action} + + + ) + + const inset = () => (narrow() ? 2 : 3) + const bodyMaxHeight = () => { + const topInset = short() ? 1 : Math.floor(dimensions().height / 4) + 1 + const footerRows = narrow() ? ACTIONS.length : 1 + const gaps = short() ? 1 : 4 + return Math.max(2, dimensions().height - topInset - footerRows - gaps) + } + + return ( + + + + + + + + + + Free limit reached + + dialog.clear()}> + esc + + + + + You've used today's free messages.{props.reset ? ` Resets at ${props.reset}.` : ""} + + + OpenCode Go gives you reliable access to the best open-source models for $10/month. + + + + } + > + + + You've used today's free messages.{props.reset ? ` Resets at ${props.reset}.` : ""} + + + OpenCode Go gives you reliable access to the best open-source models for $10/month. + + + + + + + + {(action) => button(action)} + + } + > + + {button("don't show again")} + + {(action) => button(action)} + + + + + + ) +} + +function DialogGoUpsellConnect(props: { canContinue?: () => boolean; onSubscribed: () => void }) { + const data = useData() + const currentLocation = useLocation() + const location = currentLocation.ref ?? data.location.default() + const dialog = useDialog() + const client = useClient() + const toast = useToast() + const clipboard = useClipboard() + const local = useLocal() + const newDesigns = useNewDesigns() + const theme = useTheme().surface("dialog") + const [attempt, setAttempt] = createSignal() + const integration = createMemo(() => data.location.integration.list(location)?.find((item) => item.id === "opencode")) + const method = createMemo(() => + integration()?.methods.find((item): item is IntegrationOAuthMethod => item.type === "oauth" && item.id === "device"), + ) + + let started = false + let settled = false + let closed = false + let timer: ReturnType | undefined + let pendingAttemptID: string | undefined + + const valid = () => { + const current = data.location.integration.list(location)?.find((item) => item.id === "opencode") + const active = current ? credentialConnections(current)[0] : undefined + return Boolean(active && !active.status) + } + + const finish = () => { + if (props.canContinue && !props.canContinue()) { + dialog.clear() + return + } + const goModel = (data.location.model.list(location) ?? []).find( + (item) => item.providerID === "opencode-go" && item.status !== "deprecated", + ) + if (goModel) { + local.model.set({ providerID: goModel.providerID, modelID: goModel.id }, { recent: true }) + dialog.clear() + props.onSubscribed() + return + } + dialog.replace(() => ) + } + + const sync = async (announce: boolean) => { + data.location.integration.invalidate(location) + data.location.model.invalidate(location) + data.location.provider.invalidate(location) + try { + await Promise.all([ + data.location.integration.sync(location), + data.location.model.sync(location), + data.location.provider.sync(location), + ]) + } catch (cause) { + if (closed || dialog.stack.length === 0) return false + toast.show({ variant: "error", message: errorMessage(cause) }) + dialog.clear() + return false + } + if (closed || dialog.stack.length === 0) return false + if (!valid()) return null + if (announce) { + toast.show({ variant: "success", message: `Connected ${integration()?.name ?? "OpenCode Console"}` }) + } + finish() + return true + } + + const poll = (current: IntegrationAttempt) => { + trackConnect( + (async () => { + try { + const result = await client.api.integration.oauth.status({ + integrationID: "opencode", + attemptID: current.attemptID, + location: { directory: location.directory }, + }) + if (closed) return + const status = result.data + if (status.status === "pending") { + timer = setTimeout(() => poll(current), 500) + return + } + settled = true + if (status.status === "complete") { + const ok = await sync(true) + if (ok === null && !closed) dialog.clear() + return + } + toast.show({ + variant: "error", + message: status.status === "failed" ? status.message : "Authorization expired", + }) + dialog.clear() + } catch (cause) { + if (closed) return + settled = true + toast.show({ variant: "error", message: errorMessage(cause) }) + dialog.clear() + } + })(), + ) + } + + const startOAuth = (oauthMethod: IntegrationOAuthMethod) => { + const answer: FormAnswer = {} + for (const field of oauthMethod.form ?? []) { + if (field.type !== "external" && field.default !== undefined) { + answer[field.key] = field.default + } + } + trackConnect( + (async () => { + try { + const base = { + integrationID: "opencode", + location: { directory: location.directory }, + methodID: oauthMethod.id, + } + const result = await client.api.integration.oauth.connect( + Object.keys(answer).length > 0 ? { ...base, answer } : base, + ) + if (closed) { + void client.api.integration.oauth.cancel({ + integrationID: "opencode", + attemptID: result.data.attemptID, + location: { directory: location.directory }, + }) + return + } + pendingAttemptID = result.data.attemptID + setAttempt(result.data) + poll(result.data) + } catch (cause) { + if (closed) return + toast.show({ variant: "error", message: errorMessage(cause) }) + dialog.clear() + } + })(), + ) + } + + createEffect(() => { + if (started || closed) return + const current = integration() + const oauthMethod = method() + if (!current || !oauthMethod) return + started = true + if (valid()) { + trackConnect( + (async () => { + const ok = await sync(false) + if (ok === null && !closed) startOAuth(oauthMethod) + })(), + ) + return + } + startOAuth(oauthMethod) + }) + + onCleanup(() => { + closed = true + if (timer) clearTimeout(timer) + if (settled || !pendingAttemptID) return + void client.api.integration.oauth.cancel({ + integrationID: "opencode", + attemptID: pendingAttemptID, + location: { directory: location.directory }, + }) + }) + + const open = () => { + const current = attempt() + if (!current) return + Open.openUrl(current.url).catch(() => + toast.show({ + message: "Could not open the browser. Copy the URL and continue manually.", + variant: "error", + }), + ) + } + + const copy = () => { + const current = attempt() + if (!current) return + const value = current.instructions.match(OAUTH_CODE)?.[0] ?? current.url + clipboard + .write(value) + .then(() => toast.show({ message: "Copied to clipboard", variant: "info" })) + .catch(toast.error) + } + + Keymap.createLayer(() => ({ + mode: "modal", + enabled: () => Boolean(attempt()), + commands: [ + { bind: newDesigns() ? "return" : "o", title: "Open authorization URL", group: "Dialog", run: open }, + { bind: "c", title: "Copy authorization details", group: "Dialog", run: copy }, + ], + })) + + const title = () => method()?.label ?? "Connect an integration" + + return ( + } + footer={{ + actions: { + primary: attempt() + ? [ + { + title: attempt()?.instructions.match(OAUTH_CODE) ? "copy code" : "copy link", + key: "c", + onTrigger: copy, + }, + { title: "open link", key: "enter", onTrigger: open }, + ] + : [], + }, + }} + legacyFooter={ + + } + > + + + {(current) => ( + + + + {(instructions) => {instructions()}} + + + )} + + {attempt() ? "Waiting for authorization…" : "Starting authorization…"} + + + ) +} diff --git a/packages/tui/src/component/go-upsell-art.ts b/packages/tui/src/component/go-upsell-art.ts new file mode 100644 index 000000000000..79c482647e4e --- /dev/null +++ b/packages/tui/src/component/go-upsell-art.ts @@ -0,0 +1,336 @@ +import { + Renderable, + RGBA, + TextAttributes, + type OptimizedBuffer, + type RenderableOptions, + type RenderContext, +} from "@opentui/core" +import { extend } from "@opentui/solid" +import { go } from "../logo" + +// Ported from the v1 TUI's Go upsell (bg-pulse-render.ts): rings radiate from the GO logo while a shimmer crosses it. +const PERIOD = 4600 +const RINGS = 3 +const WIDTH = 3.8 +const TAIL = 9.5 +const AMP = 0.55 +const TAIL_AMP = 0.16 +const BREATH_AMP = 0.05 +const BREATH_SPEED = 0.0008 +// Offset so the background ring emits from the GO center when the logo shimmer peaks. +const PHASE_OFFSET = 0.29 +// The frame shown when animations are off. +const STILL = PERIOD * 0.5 +const LOGO_LEFT_WIDTH = go.left[0]?.length ?? 0 +const LOGO_LINES = go.left.map((line, index) => line + " " + go.right[index]) +const LOGO_WIDTH = LOGO_LINES[0]?.length ?? 0 +const LOGO_HEIGHT = LOGO_LINES.length +const LOGO_REACH = Math.hypot(LOGO_WIDTH, LOGO_HEIGHT * 2) + 3 +const SPACE = 32 +const UPPER_HALF = "▀".codePointAt(0)! + +const enum Glyph { + None, + Shadow, + ShadowTop, + ShadowBottom, + Block, + Other, +} +const GLYPHS: Record = { + " ": Glyph.None, + _: Glyph.Shadow, + "^": Glyph.ShadowTop, + "~": Glyph.ShadowBottom, +} + +// The logo never changes, so its glyphs and the distance of each half-cell from its centre are fixed. +const LOGO_GLYPH = new Uint8Array(LOGO_WIDTH * LOGO_HEIGHT) +const LOGO_CHAR = new Uint32Array(LOGO_WIDTH * LOGO_HEIGHT) +const LOGO_ATTRIBUTES = new Uint32Array(LOGO_WIDTH * LOGO_HEIGHT) +const LOGO_TOP_DIST = new Float64Array(LOGO_WIDTH * LOGO_HEIGHT) +const LOGO_BOTTOM_DIST = new Float64Array(LOGO_WIDTH * LOGO_HEIGHT) +LOGO_LINES.forEach((line, ly) => { + Array.from(line).forEach((char, lx) => { + const index = ly * LOGO_WIDTH + lx + LOGO_GLYPH[index] = GLYPHS[char] ?? (char === "█" ? Glyph.Block : Glyph.Other) + LOGO_CHAR[index] = char.codePointAt(0) ?? SPACE + LOGO_ATTRIBUTES[index] = lx > LOGO_LEFT_WIDTH ? TextAttributes.BOLD : 0 + LOGO_TOP_DIST[index] = Math.hypot(lx + 0.5 - LOGO_WIDTH / 2, ly * 2 - LOGO_HEIGHT) + LOGO_BOTTOM_DIST[index] = Math.hypot(lx + 0.5 - LOGO_WIDTH / 2, ly * 2 + 1 - LOGO_HEIGHT) + }) +}) + +type Options = RenderableOptions & { + animate?: boolean + background?: RGBA + primary?: RGBA + logoBase?: RGBA + /** The logo is centred on this renderable's rows. */ + logoAnchor?: Renderable +} + +const clamp = (value: number) => Math.max(0, Math.min(1, value)) +const ease = (phase: number) => { + const envelope = Math.sin(phase * Math.PI) + return envelope * envelope * (3 - 2 * envelope) +} +const mix = (out: RGBA, a: RGBA, b: RGBA, amount: number) => { + const t = clamp(amount) + out.r = a.r + (b.r - a.r) * t + out.g = a.g + (b.g - a.g) * t + out.b = a.b + (b.b - a.b) * t + out.a = 1 + return out +} +const put = (plane: Uint16Array, offset: number, color: RGBA) => { + const source = color.buffer + plane[offset] = source[0]! + plane[offset + 1] = source[1]! + plane[offset + 2] = source[2]! + plane[offset + 3] = source[3]! +} + +/** + * Draws into its own frame buffer (`buffered`) through `withBuffers`, so a frame is plain typed-array + * writes plus one blit instead of a recorded draw call per cell. Live (and so the renderer's frame loop) + * only while animating and visible; a still frame is drawn once and reused until something changes. + */ +export class GoUpsellArtRenderable extends Renderable { + private elapsed: number + private _animate: boolean + private _background = RGBA.fromInts(0, 0, 0) + private _primary = RGBA.fromInts(255, 255, 255) + private _logoBase = RGBA.fromInts(180, 180, 180) + private anchor: Renderable | undefined + private readonly white = RGBA.fromInts(255, 255, 255) + private readonly shadow = RGBA.fromInts(0, 0, 0) + private readonly pulse = RGBA.fromInts(0, 0, 0) + private readonly fg = RGBA.fromInts(0, 0, 0) + private readonly bg = RGBA.fromInts(0, 0, 0) + private readonly ringHeads = new Float64Array(RINGS) + private readonly ringEased = new Float64Array(RINGS) + private readonly logoHeads = new Float64Array(2) + private readonly logoEased = new Float64Array(2) + private peak = 0 + private glow = 0 + // Distance and falloff for every cell, rebuilt only when the size or logo position changes. + private dist = new Float64Array(0) + private falloff = new Float64Array(0) + private fieldKey = "" + private reach = 0 + private drawnKey = "" + + constructor(ctx: RenderContext, options: Options = {}) { + const animate = options.animate ?? true + super(ctx, { ...options, live: animate, buffered: true }) + this._animate = animate + this.elapsed = animate ? 0 : STILL + if (options.background) this._background = options.background + if (options.primary) this._primary = options.primary + if (options.logoBase) this._logoBase = options.logoBase + this.anchor = options.logoAnchor + } + + set animate(value: boolean) { + if (value === this._animate) return + this._animate = value + this.live = value + if (!value) this.elapsed = STILL + this.redraw() + } + + set background(value: RGBA) { + this._background = value + this.redraw() + } + + set primary(value: RGBA) { + this._primary = value + this.redraw() + } + + set logoBase(value: RGBA) { + this._logoBase = value + this.redraw() + } + + set logoAnchor(value: Renderable | undefined) { + this.anchor = value + this.redraw() + } + + private redraw() { + this.drawnKey = "" + this.requestRender() + } + + protected override onUpdate(deltaTime: number) { + if (this._animate) this.elapsed = (this.elapsed + deltaTime) % PERIOD + } + + protected override destroySelf() { + this.live = false + this.anchor = undefined + this.dist = new Float64Array(0) + this.falloff = new Float64Array(0) + super.destroySelf() + } + + private field(width: number, height: number, logoX: number, logoY: number) { + const key = `${width}x${height}@${logoX},${logoY}` + if (key === this.fieldKey) return + this.fieldKey = key + const centerX = logoX + LOGO_WIDTH / 2 + const centerY = logoY + LOGO_HEIGHT / 2 + const reach = Math.hypot(Math.max(centerX, width - centerX), Math.max(centerY, height - centerY) * 2) + TAIL + this.reach = reach + if (this.dist.length !== width * height) { + this.dist = new Float64Array(width * height) + this.falloff = new Float64Array(width * height) + } + for (let y = 0; y < height; y++) { + for (let x = 0; x < width; x++) { + const dist = Math.hypot(x + 0.5 - centerX, (y + 0.5 - centerY) * 2) + this.dist[y * width + x] = dist + this.falloff[y * width + x] = Math.max(0, 1 - (dist / (reach * 0.85)) ** 2) + } + } + } + + /** Sets peak (white core) and glow (primary tint) for a logo half-cell at dist from the logo centre. */ + private logoPulse(dist: number) { + let peak = 0.04 + let glow = 0 + for (let i = 0; i < 2; i++) { + const delta = dist - this.logoHeads[i]! + const core = Math.exp(-(Math.abs(delta / 1.2) ** 1.8)) + const soft = Math.exp(-(Math.abs(delta / 7) ** 1.6)) + const tail = delta < 0 && delta > -7 ? (1 + delta / 7) ** 2.6 : 0 + peak += core * 0.65 * this.logoEased[i]! + glow += (soft * 0.16 + tail * 0.22) * this.logoEased[i]! + } + this.peak = Math.min(1, peak) + this.glow = Math.min(1, glow) + } + + private logoTint(out: RGBA, base: RGBA, glow: number, peak: number) { + return mix(out, mix(out, base, this._primary, glow), this.white, peak) + } + + protected override renderSelf(buffer: OptimizedBuffer) { + const width = this.width + const height = this.height + if (!this.visible || this.isDestroyed || width <= 0 || height <= 0) return + // Without its own frame buffer (not yet sized) there is nothing safe to write into directly. + if (buffer !== this.frameBuffer) return + const t = this.elapsed + const rgb = this._ctx.capabilities?.rgb === true + const logoX = Math.max(0, Math.floor((width - LOGO_WIDTH) / 2)) + const rawLogoY = this.anchor ? this.anchor.screenY - this.screenY : Math.round((height - LOGO_HEIGHT) / 2) - 1 + const logoY = Math.max(0, Math.min(Math.max(0, height - LOGO_HEIGHT), rawLogoY)) + this.field(width, height, logoX, logoY) + // The frame buffer keeps the last frame, so a still frame is drawn once. + const drawnKey = this._animate ? "" : `${this.fieldKey}:${rgb}` + if (drawnKey && drawnKey === this.drawnKey) return + const breath = (0.5 + 0.5 * Math.sin(t * BREATH_SPEED)) * BREATH_AMP + for (let ring = 0; ring < RINGS; ring++) { + const phase = (t / PERIOD + ring / RINGS - PHASE_OFFSET + 1) % 1 + this.ringHeads[ring] = phase * this.reach + this.ringEased[ring] = ease(phase) + } + const logoPhase0 = (t / PERIOD) % 1 + const logoPhase1 = (t / PERIOD + 0.5) % 1 + this.logoHeads[0] = logoPhase0 * LOGO_REACH + this.logoHeads[1] = logoPhase1 * LOGO_REACH + this.logoEased[0] = ease(logoPhase0) + this.logoEased[1] = ease(logoPhase1) + mix(this.shadow, this._background, this._logoBase, 0.25) + + buffer.withBuffers((cells) => { + const stride = cells.width + const rows = Math.min(height, cells.height) + const cols = Math.min(width, stride) + const { char, fg, bg, attributes } = cells + for (let y = 0; y < rows; y++) { + const ly = y - logoY + const inLogoRow = ly >= 0 && ly < LOGO_HEIGHT + for (let x = 0; x < cols; x++) { + const field = y * width + x + const dist = this.dist[field]! + let level = 0 + for (let ring = 0; ring < RINGS; ring++) { + const delta = dist - this.ringHeads[ring]! + const crest = Math.abs(delta) < WIDTH ? 0.5 + 0.5 * Math.cos((delta / WIDTH) * Math.PI) : 0 + const tail = delta < 0 && delta > -TAIL ? (1 + delta / TAIL) ** 2.3 : 0 + level += (crest * AMP + tail * TAIL_AMP) * this.ringEased[ring]! + } + mix( + this.pulse, + this._background, + this._primary, + Math.min(1, (level / RINGS + breath) * this.falloff[field]!) * 0.7, + ) + + const cell = y * stride + x + const color = cell * 4 + const lx = x - logoX + const logo = inLogoRow && lx >= 0 && lx < LOGO_WIDTH ? ly * LOGO_WIDTH + lx : -1 + const glyph = logo < 0 ? Glyph.None : (LOGO_GLYPH[logo]! as Glyph) + if (glyph === Glyph.None) { + char[cell] = SPACE + put(fg, color, this.pulse) + put(bg, color, this.pulse) + attributes[cell] = 0 + continue + } + attributes[cell] = LOGO_ATTRIBUTES[logo]! + this.logoPulse(LOGO_TOP_DIST[logo]!) + const topPeak = this.peak + const topGlow = this.glow + this.logoPulse(LOGO_BOTTOM_DIST[logo]!) + const bottomPeak = this.peak + const bottomGlow = this.glow + if (glyph === Glyph.Shadow) { + mix(this.bg, this.shadow, this.white, Math.max(topPeak, bottomPeak) * 0.18) + char[cell] = SPACE + put(fg, color, this.bg) + put(bg, color, this.bg) + } else if (glyph === Glyph.ShadowTop) { + this.logoTint(this.fg, this._logoBase, topGlow, topPeak) + mix(this.bg, this.shadow, this.white, bottomPeak * 0.18) + char[cell] = UPPER_HALF + put(fg, color, this.fg) + put(bg, color, this.bg) + } else if (glyph === Glyph.ShadowBottom) { + mix(this.fg, this.shadow, this.white, topPeak * 0.18) + char[cell] = UPPER_HALF + put(fg, color, this.fg) + put(bg, color, this.pulse) + } else if (glyph === Glyph.Block && rgb) { + this.logoTint(this.fg, this._logoBase, topGlow, topPeak) + this.logoTint(this.bg, this._logoBase, bottomGlow, bottomPeak) + char[cell] = UPPER_HALF + put(fg, color, this.fg) + put(bg, color, this.bg) + } else { + this.logoTint(this.fg, this._logoBase, (topGlow + bottomGlow) / 2, (topPeak + bottomPeak) / 2) + char[cell] = LOGO_CHAR[logo]! + put(fg, color, this.fg) + put(bg, color, this.pulse) + } + } + } + }) + this.drawnKey = drawnKey + } +} + +declare module "@opentui/solid" { + interface OpenTUIComponents { + go_upsell_art: typeof GoUpsellArtRenderable + } +} + +extend({ go_upsell_art: GoUpsellArtRenderable }) diff --git a/packages/tui/src/config/index.tsx b/packages/tui/src/config/index.tsx index 953d5977d00f..87c9a1bb373d 100644 --- a/packages/tui/src/config/index.tsx +++ b/packages/tui/src/config/index.tsx @@ -5,7 +5,7 @@ import { Vcs } from "@opencode/schema/vcs" import { Schema } from "effect" import { createContext, onCleanup, type JSX, useContext } from "solid-js" import { createStore, reconcile } from "solid-js/store" -import { watch } from "fs" +import { safeWatch } from "../util/watch" import path from "path" import { TuiKeybind } from "./keybind" @@ -361,7 +361,7 @@ export function ConfigProvider(props: { } let reload = Promise.resolve() const watcher = host?.path - ? watch(path.dirname(host.path), () => { + ? safeWatch(path.dirname(host.path), () => { reload = reload .then(() => host.get()) .then(apply) diff --git a/packages/tui/src/context/location.tsx b/packages/tui/src/context/location.tsx index 8c668757e75d..84fd6ae87b54 100644 --- a/packages/tui/src/context/location.tsx +++ b/packages/tui/src/context/location.tsx @@ -1,4 +1,5 @@ -import type { LocationGetOutput, LocationRef } from "@opencode/client" +import { isLocationNotFoundError, type LocationGetOutput, type LocationRef } from "@opencode/client" +import { retry } from "@opencode/util/retry" import { createContext, createMemo, createSignal, onCleanup, useContext, type ParentProps } from "solid-js" import { useClient } from "./client" import { useData } from "./data" @@ -28,9 +29,14 @@ export function LocationProvider(props: ParentProps) { ? undefined : location setError(undefined) - void data.location.sync(target).catch((cause) => { + // Only the server's typed LocationNotFoundError proves the folder is gone. Other failures, such as + // plugins still activating after a restart, are retried and never block the composer. + void retry(() => (generation === attempt ? data.location.sync(target) : Promise.resolve()), { + retryIf: (cause) => generation === attempt && !isLocationNotFoundError(cause), + }).catch((cause) => { const current = ref() if ( + !isLocationNotFoundError(cause) || generation !== attempt || current?.directory !== location.directory || current.workspaceID !== location.workspaceID diff --git a/packages/tui/src/context/storage.tsx b/packages/tui/src/context/storage.tsx index 57f45a68f86d..95264a66044f 100644 --- a/packages/tui/src/context/storage.tsx +++ b/packages/tui/src/context/storage.tsx @@ -1,9 +1,10 @@ import { batch, createContext, onCleanup, useContext, type ParentProps } from "solid-js" import { createStore, produce, reconcile, type Store } from "solid-js/store" import path from "path" -import { mkdirSync, readFileSync, watch } from "fs" +import { mkdirSync, readFileSync } from "fs" import { Flock } from "@opencode/util/flock" import { writeJsonAtomic } from "../util/persistence" +import { safeWatch } from "../util/watch" import { useTuiApp, useTuiPaths } from "./runtime" type Options = { @@ -111,26 +112,16 @@ export function createStorage(root: string, channel: string) { } let reloadTimer: ReturnType | undefined - let watcher: ReturnType | undefined - try { - watcher = watch(directory, () => { - clearTimeout(reloadTimer) - // Atomic writes notify for the temporary file before its final rename, and some - // platforms coalesce the rename event. Reload after the event burst has settled. - reloadTimer = setTimeout(() => entries.forEach((entry) => entry.reload()), 50) - }) - watcher.on("error", (error) => { - clearTimeout(reloadTimer) - watcher?.close() - watcher = undefined - console.error("Storage directory watcher failed, live-reload disabled", { directory, error }) - }) - } catch (error) { - // fs.watch throws synchronously (e.g. ENOSPC when the inotify watch limit is - // exhausted). Losing cross-process live-reload is recoverable; crashing the - // whole TUI over it is not, so degrade instead of propagating. - console.error("Failed to watch storage directory, live-reload disabled", { directory, error }) - } + let watcher = safeWatch(directory, () => { + clearTimeout(reloadTimer) + // Atomic writes notify for the temporary file before its final rename, and some + // platforms coalesce the rename event. Reload after the event burst has settled. + reloadTimer = setTimeout(() => entries.forEach((entry) => entry.reload()), 50) + }) + watcher?.on("error", () => { + clearTimeout(reloadTimer) + watcher = undefined + }) return { storage, close: () => { diff --git a/packages/tui/src/feature-plugins/system/notifications.ts b/packages/tui/src/feature-plugins/system/notifications.ts index 5435f49689fc..fe21e258a000 100644 --- a/packages/tui/src/feature-plugins/system/notifications.ts +++ b/packages/tui/src/feature-plugins/system/notifications.ts @@ -66,8 +66,15 @@ export default Plugin.define({ return } errored.add(sessionID) - notify(context, sessionID, event.data.error.message, "error") - context.ui.toast.show({ sessionID, title: "Session failed", message: event.data.error.message, variant: "error" }) + if (event.data.error.type !== "provider.free-limit") { + notify(context, sessionID, event.data.error.message, "error") + context.ui.toast.show({ + sessionID, + title: "Session failed", + message: event.data.error.message, + variant: "error", + }) + } ended(sessionID) }), ] diff --git a/packages/tui/src/model-preference.ts b/packages/tui/src/model-preference.ts index 0da826f882aa..2793130a99af 100644 --- a/packages/tui/src/model-preference.ts +++ b/packages/tui/src/model-preference.ts @@ -1,7 +1,7 @@ import { readJson, writeJsonAtomic } from "./util/persistence" import { isRecord } from "./util/record" +import { safeWatch } from "./util/watch" import { Flock } from "@opencode/util/flock" -import { watch } from "node:fs" import path from "node:path" export type ModelPreferenceModel = { @@ -103,7 +103,7 @@ function patch(value: Partial) { export function createModelPreferenceRepository(filePath: string) { let pending = Promise.resolve() let revision = 0 - let watcher: ReturnType | undefined + let watcher: ReturnType let reload: ReturnType | undefined const listeners = new Set<(value: ModelPreference) => void>() const read = () => @@ -152,15 +152,14 @@ export function createModelPreferenceRepository(filePath: string) { listeners.add(listener) void refresh() if (!watcher) { - watcher = watch(path.dirname(filePath), (_event, filename) => { + watcher = safeWatch(path.dirname(filePath), (_event, filename) => { const changed = filename?.toString() const name = path.basename(filePath) if (changed !== undefined && changed !== name && !changed.startsWith(name + ".")) return clearTimeout(reload) reload = setTimeout(() => void refresh(), 50) }) - watcher.on("error", () => { - watcher?.close() + watcher?.on("error", () => { watcher = undefined }) } diff --git a/packages/tui/src/plugin/watch.ts b/packages/tui/src/plugin/watch.ts index 9a33a4d36aef..863646660732 100644 --- a/packages/tui/src/plugin/watch.ts +++ b/packages/tui/src/plugin/watch.ts @@ -1,6 +1,7 @@ import path from "path" -import { existsSync, watch } from "fs" +import { existsSync, type FSWatcher } from "fs" import { lstat, realpath, stat } from "fs/promises" +import { safeWatch } from "../util/watch" // Watch plugin sources for changes. Files are watched through their parent // directory (editors that save by rename replace the inode, which silently @@ -14,7 +15,7 @@ import { lstat, realpath, stat } from "fs/promises" // polled until they can be armed without relying on a racy chain of ancestor // watches. export function createSourceWatcher(onChange: () => void) { - const watchers = new Map>() + const watchers = new Map() const watched = new Map | null>() const missing = new Set() const arming = new Map>() @@ -44,7 +45,7 @@ export function createSourceWatcher(onChange: () => void) { if (appeared) notify() return } - const watcher = watch(dir, (_event, filename) => { + const watcher = safeWatch(dir, (_event, filename) => { // A replaced directory keeps this watcher on the dead inode (Linux // emits rename, not error); forget it so a later add() re-arms on // the recreated path, and still schedule so reconcile runs now. @@ -58,6 +59,9 @@ export function createSourceWatcher(onChange: () => void) { if (filename && accept && !accept.has(filename.toString())) return notify() }) + // Watching is unavailable (opted out or the host refused the watch); + // leave this target unwatched instead of claiming it in the maps. + if (!watcher) return watched.set(dir, name === null ? null : new Set([name])) // Reconcile after watcher errors so every source is re-added and any // temporarily unavailable target moves into the polling set. diff --git a/packages/tui/src/routes/session/index.tsx b/packages/tui/src/routes/session/index.tsx index 607c5a6d6523..f791a867059b 100644 --- a/packages/tui/src/routes/session/index.tsx +++ b/packages/tui/src/routes/session/index.tsx @@ -70,6 +70,8 @@ import { statusLabel } from "../../component/dialog-workspace-file-changes" import { DialogMessage } from "./dialog-message" import { DialogFork } from "./dialog-fork" import { DialogTimeline } from "./dialog-timeline" +import { DialogGoUpsell } from "../../component/dialog-go-upsell" +import { useStorage } from "../../context/storage" import { Composer } from "./composer" import { filetype } from "../../util/filetype" import parsers from "../../parsers-config" @@ -626,6 +628,123 @@ export function Session(props: { current.submit() }) const dialog = useDialog() + const storage = useStorage() + const [goUpsell, updateGoUpsell] = storage.store<{ lastSeenAt?: number; dontShow?: boolean }>("go-upsell", { + initial: {}, + }) + const [goUpsellMemory, updateGoUpsellMemory] = storage.memory<{ lastSeenAt?: number; dontShow?: boolean }>( + "go-upsell", + { initial: {} }, + ) + const handledFreeLimits = new Set() + let sessionDisposed = false + onCleanup(() => { + sessionDisposed = true + }) + // Restore the prompt when the free allowance stops it, and offer OpenCode Go on the first hit in a day. + onCleanup( + data.on("session.step.failed", (event) => { + if (event.data.sessionID !== route.sessionID || event.data.error.type !== "provider.free-limit") return + if (handledFreeLimits.has(event.data.assistantMessageID)) return + handledFreeLimits.add(event.data.assistantMessageID) + const current = prompt() + const wasEmpty = Boolean( + current && + !current.current.text.trim() && + !current.current.files?.length && + !current.current.agents?.length && + !current.current.skills?.length && + !current.current.pasted.length, + ) + if (!wasEmpty || !current) return + const list = messages() + const end = list.findIndex((item) => item.id === event.data.assistantMessageID) + for (let i = end - 1; i >= 0; i--) { + const item = list[i] + if (!item || item.type === "shell" || item.type === "idle") break + if (item.type === "user" && !data.session.input.has(route.sessionID, item.id)) { + current.set({ ...projectedPromptInput(item), pasted: [] }) + break + } + } + if (dialog.stack.length > 0) return + if (goUpsell.dontShow || goUpsellMemory.dontShow) return + const now = Date.now() + const seen = Math.max(goUpsell.lastSeenAt ?? 0, goUpsellMemory.lastSeenAt ?? 0) + if (seen > 0 && now - seen < 86_400_000) return + const expectedText = current.current.text + const expectedFiles = current.current.files?.length ?? 0 + const expectedAgents = current.current.agents?.length ?? 0 + const expectedSkills = current.current.skills?.length ?? 0 + const expectedPasted = current.current.pasted.length + const hasRecalledPrompt = () => { + const active = prompt() + return Boolean( + !sessionDisposed && + route.sessionID === event.data.sessionID && + active && + active.current.text === expectedText && + (active.current.files?.length ?? 0) === expectedFiles && + (active.current.agents?.length ?? 0) === expectedAgents && + (active.current.skills?.length ?? 0) === expectedSkills && + active.current.pasted.length === expectedPasted, + ) + } + const canOpen = () => + Boolean(hasRecalledPrompt() && dialog.stack.length === 0 && !goUpsell.dontShow && !goUpsellMemory.dontShow) + const previousSeen = goUpsellMemory.lastSeenAt + updateGoUpsellMemory((draft) => { + draft.lastSeenAt = now + }) + let claimed = false + void updateGoUpsell((draft) => { + if (draft.dontShow) { + updateGoUpsellMemory((mem) => { + mem.dontShow = true + mem.lastSeenAt = draft.lastSeenAt ?? previousSeen + }) + return + } + if ((draft.lastSeenAt ?? 0) > 0 && now - (draft.lastSeenAt ?? 0) < 86_400_000) { + updateGoUpsellMemory((mem) => { + mem.lastSeenAt = draft.lastSeenAt + }) + return + } + if (!canOpen()) { + updateGoUpsellMemory((mem) => { + mem.lastSeenAt = previousSeen + }) + return + } + draft.lastSeenAt = now + claimed = true + }) + .catch(() => { + claimed = true + }) + .then(() => { + if (!claimed || !canOpen()) return + dialog.replace(() => ( + { + updateGoUpsellMemory((draft) => { + draft.dontShow = true + }) + void updateGoUpsell((draft) => { + draft.dontShow = true + }).catch(() => undefined) + }} + onSubscribe={() => { + prompt()?.submit() + }} + /> + )) + }) + }), + ) const runPendingAction = createSingleFlight() const mutatePending = async (action: PendingAction, inboxID: string, failureLabel?: string) => { const result = await runPendingAction(inboxID, async () => { @@ -1916,12 +2035,20 @@ function AssistantFooter(props: { message: SessionMessageAssistant }) { turnTokensPerSecond(props.message, messages(), ctx.messageIndex(props.message.id), ctx.legacyTurns()), ) const interrupted = createMemo(() => props.message.error?.message === "Step interrupted") + const freeLimit = createMemo(() => props.message.error?.type === "provider.free-limit") return ( <> - - Error: {errorMessage(props.message.error)} - + + Error: {errorMessage(props.message.error)} + + } + > + + @@ -2393,6 +2520,33 @@ function QueuedPromptDock(props: { prompts: { id: string; text: string }[]; onOp ) } +function freeLimitReset(delay: number | undefined, base: number) { + if (delay === undefined || !Number.isFinite(delay) || delay < 0 || delay > 7 * 86_400_000 || !Number.isFinite(base)) + return + const at = base + delay + if (Number.isNaN(new Date(at).getTime())) return + const now = Date.now() + return new Date(at).toDateString() === new Date(now).toDateString() && at - now < 86_400_000 + ? Locale.time(at) + : Locale.datetime(at) +} + +function FreeLimitNotice(props: { message: Pick }) { + const theme = useTheme() + const reset = createMemo(() => + freeLimitReset(props.message.error?.retryAfterMs, props.message.time.completed ?? props.message.time.created), + ) + return ( + + You've used today's free messages.{reset() ? ` They reset at ${reset()}.` : ""} + + Upgrade to OpenCode Go → /connect · Switch model →{" "} + /models + + + ) +} + function AssistantRetry(props: { retry: SessionMessageAssistant["retry"] }) { const theme = useTheme() const [seconds, setSeconds] = createSignal(0) diff --git a/packages/tui/src/util/watch.ts b/packages/tui/src/util/watch.ts new file mode 100644 index 000000000000..0d4dc5ff536d --- /dev/null +++ b/packages/tui/src/util/watch.ts @@ -0,0 +1,24 @@ +import { watch, type WatchListener } from "node:fs" + +/** + * Watch a path without letting a failed watch take down the TUI. fs.watch throws + * synchronously when the host cannot allocate one (e.g. ENOSPC/EMFILE when the + * inotify budget is exhausted) and emits "error" later if the watch breaks; an + * unhandled "error" event crashes the process. Losing live-reload is recoverable, + * so both paths close the watch and log. Honors the server's disable switch. + */ +export function safeWatch(target: string, listener: WatchListener) { + const disabled = process.env.OPENCODE_FILEWATCHER_DISABLE ?? process.env.OPENCODE_DISABLE_FILEWATCHER + if (disabled === "1" || disabled?.toLowerCase() === "true") return + try { + const watcher = watch(target, listener) + watcher.on("error", (error) => { + watcher.close() + console.error("Watch failed, live-reload disabled", { target, error }) + }) + return watcher + } catch (error) { + console.error("Failed to watch path, live-reload disabled", { target, error }) + return + } +} diff --git a/packages/tui/test/app-lifecycle.test.tsx b/packages/tui/test/app-lifecycle.test.tsx index d3394318874e..d3050c43cefb 100644 --- a/packages/tui/test/app-lifecycle.test.tsx +++ b/packages/tui/test/app-lifecycle.test.tsx @@ -1,8 +1,11 @@ -import { expect, test } from "bun:test" -import { EmbeddedTerminalRenderable } from "@opentui/core" +import { expect, spyOn, test } from "bun:test" +import { EmbeddedTerminalRenderable, type Renderable } from "@opentui/core" import { createTestRenderer } from "@opentui/core/testing" import { Effect, FileSystem } from "effect" import { Global } from "@opencode/util/global" +import * as Open from "@opencode/util/open" +import { waitForGoUpsellConnect } from "../src/component/dialog-go-upsell" +import { GoUpsellArtRenderable } from "../src/component/go-upsell-art" import { SimulationRenderer } from "@opencode/simulation/frontend/renderer" import path from "node:path" import { Writable } from "node:stream" @@ -1701,6 +1704,982 @@ test("server plugin failures share one notice and use source names before an ID expect(setup.captureCharFrame()).toContain("Open plugins") }) +test("free usage limit offers OpenCode Go once per 24h, transitions Subscribe to Connect Go, and restores focus", async () => { + await using state = await tmpdir() + const baseNow = 1_700_000_000_000 + let currentNow = baseNow + using _now = spyOn(Date, "now").mockImplementation(() => currentNow) + const session = { + id: "ses_free_limit", + projectID: "proj_test", + location: { directory }, + title: "Free limit", + agent: "build", + cost: 0, + tokens: { input: 0, output: 0, reasoning: 0, cache: { read: 0, write: 0 } }, + time: { created: 1, updated: 1 }, + } + const model = { id: "big-pickle", providerID: "opencode" } + const freeModel = { + id: "big-pickle", + providerID: "opencode", + name: "Big Pickle", + status: "active", + cost: [{ input: 0, output: 0, cache: { read: 0, write: 0 } }], + capabilities: { input: ["text"], output: ["text"], reasoning: false, toolCall: true, attachment: false }, + limit: { context: 200_000, output: 8_192 }, + variants: [], + time: { released: 1 }, + } + const goModel = { + id: "glm-5-go", + providerID: "opencode-go", + name: "GLM-5 Go", + status: "active", + cost: [{ input: 1, output: 2, cache: { read: 0, write: 0 } }], + capabilities: { input: ["text"], output: ["text"], reasoning: false, toolCall: true, attachment: false }, + limit: { context: 200_000, output: 8_192 }, + variants: [], + time: { released: 2 }, + } + const error = { + type: "provider.free-limit", + message: "Free usage limit reached.", + status: 429, + retryAfterMs: 3_600_000, + } + let authorized = false + let releaseOAuth!: () => void + const oauthGate = new Promise((resolve) => { + releaseOAuth = resolve + }) + const oauthRequests: unknown[] = [] + const switchedModels: unknown[] = [] + const resentPrompts: unknown[] = [] + const promptResent = Promise.withResolvers() + await using setup = await createAppFixture({ + width: 100, + state: state.path, + args: { sessionID: session.id }, + config: { animations: false, tabs: { mode: "off" } }, + fetch: async (url, request) => { + const location = { directory, project: { id: "proj_test", directory, canonical: directory } } + if (url.pathname === "/api/session") return json({ data: [session], cursor: {} }) + if (url.pathname === `/api/session/${session.id}`) return json({ data: session }) + if (url.pathname === "/api/agent") + return json({ location, data: [{ id: "build", mode: "primary", hidden: false, permissions: [] }] }) + if (url.pathname === "/api/provider") + return json({ + location, + data: [ + { id: "opencode", name: "OpenCode Zen" }, + ...(authorized ? [{ id: "opencode-go", integrationID: "opencode", name: "OpenCode Go" }] : []), + ], + }) + if (url.pathname === "/api/model") + return json({ + location, + data: authorized ? [goModel, freeModel] : [freeModel], + }) + if (url.pathname === "/api/integration") + return json({ + location, + data: [ + { + id: "opencode", + name: "OpenCode Console", + methods: [ + { + id: "device", + type: "oauth", + label: "OpenCode Console account", + form: [ + { + key: "server", + type: "string", + format: "uri", + hidden: true, + default: "https://opencode.ai/console", + }, + ], + }, + { type: "key", label: "API key (service account)" }, + ], + connections: authorized + ? [{ type: "credential", method: "oauth", id: "cred_console", label: "Anomaly" }] + : [], + }, + { + id: "opencode-go", + name: "OpenCode Go", + methods: [{ type: "key", label: "Enter Go API key" }], + connections: [], + }, + ], + }) + if (url.pathname === "/api/integration/opencode/connect/oauth" && request.method === "POST") { + oauthRequests.push(await request.json()) + return json({ + location, + data: { + attemptID: "attempt_go", + mode: "auto", + url: "https://opencode.ai/console/device?user_code=GOGO-1234", + instructions: "Enter code: GOGO-1234", + }, + }) + } + if (url.pathname === "/api/integration/opencode/connect/oauth/attempt_go") { + if (request.method === "DELETE") return new Response(null, { status: 204 }) + await oauthGate + authorized = true + return json({ location, data: { status: "complete" } }) + } + if (url.pathname === `/api/session/${session.id}/model` && request.method === "POST") { + switchedModels.push(await request.json()) + return new Response(null, { status: 204 }) + } + if (url.pathname === `/api/session/${session.id}/prompt` && request.method === "POST") { + resentPrompts.push(await request.json()) + promptResent.resolve() + return json({ data: {} }) + } + if (url.pathname === `/api/session/${session.id}/message`) + return json({ + data: [ + { id: "msg_queued_user", type: "user", text: "Queued follow-up", time: { created: 3 } }, + { + id: "msg_free_assistant", + type: "assistant", + agent: "build", + model, + content: [], + time: { created: 2 }, + }, + { id: "msg_free_user", type: "user", text: "Refactor the parser", time: { created: 1 } }, + ], + cursor: {}, + }) + if (url.pathname === `/api/session/${session.id}/inbox`) + return json({ + data: [ + { + id: "msg_queued_user", + sessionID: session.id, + type: "user", + delivery: "queue", + payload: { text: "Queued follow-up" }, + time: { created: 3 }, + }, + ], + }) + if (url.pathname === `/api/session/${session.id}/permission`) return json({ data: [] }) + return undefined + }, + }) + await setup.ready + await setup.waitForFrame((frame) => frame.includes("Refactor the parser")) + setup.events.emit({ + id: "evt_free_limit_1", + created: baseNow, + type: "session.step.failed", + durable: { aggregateID: session.id, seq: 1, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant", error }, + }) + setup.events.emit({ + id: "evt_free_limit_exec", + created: baseNow, + type: "session.execution.failed", + durable: { aggregateID: session.id, seq: 2, version: 1 }, + data: { sessionID: session.id, error }, + }) + using openSpy = spyOn(Open, "openUrl").mockResolvedValue(undefined as never) + await setup.waitForFrame((frame) => frame.includes("Free limit reached")) + const upsell = setup.captureCharFrame() + expect(upsell).toContain("OpenCode Go gives you reliable access") + expect(upsell).toContain("▀▀▀▀ ▀▀▀▀") + expect(upsell).toContain("opencode.ai/go") + expect(upsell).toContain("don't show again") + expect(upsell).toContain("dismiss") + expect(upsell).toContain("subscribe") + expect(upsell).not.toContain("Connect OpenCode Go") + expect(upsell).not.toContain("Session failed") + // Dismiss closes the modal, leaves the muted notice, and restores focus to the composer. + setup.mockInput.pressArrow("left") + setup.mockInput.pressEnter() + await setup.waitForFrame((frame) => !frame.includes("Free limit reached")) + const frame = setup.captureCharFrame() + expect(frame).toContain("You've used today's free messages. They reset at") + expect(frame).toContain("Upgrade to OpenCode Go → /connect · Switch model → /models") + expect(frame).not.toContain("Error:") + expect(frame).not.toContain("Session failed") + // The delivered prompt (not the queued one) shows once in the transcript and once more back in the composer. + await setup.waitForFrame((frame) => frame.split("Refactor the parser").length === 3) + expect(setup.captureCharFrame().split("Queued follow-up").length).toBe(2) + + // Clear the composer (which also verifies composer focus was restored after Dismiss): a duplicate event does not re-restore. + await Bun.sleep(10) + setup.mockInput.pressKey("c", { ctrl: true }) + await setup.waitForFrame((frame) => frame.split("Refactor the parser").length === 2) + setup.events.emit({ + id: "evt_free_limit_dup", + created: baseNow, + type: "session.step.failed", + durable: { aggregateID: session.id, seq: 3, version: 1 }, + data: { + sessionID: session.id, + assistantMessageID: "msg_free_assistant", + error: { ...error, retryAfterMs: undefined }, + }, + }) + await setup.waitForFrame((frame) => !frame.includes("They reset at")) + expect(setup.captureCharFrame().split("Refactor the parser").length).toBe(2) + + // Just before 24h (86_400_000 - 1 ms) with an empty composer: recalls the next turn's prompt without reopening Free limit reached. + currentNow = baseNow + 86_400_000 - 1 + setup.events.emit({ + id: "evt_free_limit_delivered", + created: 4, + type: "session.inbox.delivered", + durable: { aggregateID: session.id, seq: 4, version: 1 }, + data: { sessionID: session.id, inboxID: "msg_queued_user" }, + }) + setup.events.emit({ + id: "evt_free_limit_2", + created: 5, + type: "session.step.started", + durable: { aggregateID: session.id, seq: 5, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_2", agent: "build", model, started: 5 }, + }) + setup.events.emit({ + id: "evt_free_limit_3", + created: currentNow, + type: "session.step.failed", + durable: { aggregateID: session.id, seq: 6, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_2", error }, + }) + await setup.waitForFrame( + (frame) => + frame.split("You've used today's free messages").length === 3 && frame.split("Queued follow-up").length === 3, + ) + expect(setup.captureCharFrame()).not.toContain("Free limit reached") + expect(setup.captureCharFrame()).not.toContain("Connect OpenCode Go") + + // Exactly at 24h (86_400_000 ms) with an empty composer: reopens Free limit reached, and Subscribe starts OpenCode device OAuth directly, selects Go, and resends the saved prompt. + currentNow = baseNow + 86_400_000 + setup.mockInput.pressKey("c", { ctrl: true }) + await setup.waitForFrame((frame) => frame.split("Queued follow-up").length === 2) + setup.events.emit({ + id: "evt_free_limit_4", + created: 7, + type: "session.step.started", + durable: { aggregateID: session.id, seq: 7, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_3", agent: "build", model, started: 7 }, + }) + setup.events.emit({ + id: "evt_free_limit_5", + created: currentNow, + type: "session.step.failed", + durable: { aggregateID: session.id, seq: 8, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_3", error }, + }) + await setup.waitForFrame((frame) => frame.includes("Free limit reached")) + setup.mockInput.pressEnter() + await setup.waitForFrame((frame) => frame.includes("GOGO-1234")) + expect(openSpy).toHaveBeenCalledWith("https://opencode.ai/go") + expect(oauthRequests).toEqual([ + { methodID: "device", answer: { server: "https://opencode.ai/console" } }, + ]) + const connectFrame = setup.captureCharFrame() + expect(connectFrame).toContain("OpenCode Console account") + expect(connectFrame).not.toContain("Enter Go API key") + expect(connectFrame).not.toContain("API key (service account)") + releaseOAuth() + await promptResent.promise + await setup.waitForFrame((frame) => !frame.includes("GOGO-1234") && frame.includes("GLM-5 Go")) + expect(switchedModels).toEqual([{ model: { providerID: "opencode-go", id: "glm-5-go" } }]) + expect(resentPrompts).toHaveLength(1) + expect(resentPrompts[0]).toMatchObject({ + text: "Queued follow-up", + }) +}) + +test("free usage limit Subscribe handles mid-auth/mid-sync Escape, sync failure, already-connected accounts, and Don't show again across restart", async () => { + await using state = await tmpdir() + const baseNow = 1_700_000_000_000 + let currentNow = baseNow + using _now = spyOn(Date, "now").mockImplementation(() => currentNow) + using openSpy = spyOn(Open, "openUrl").mockResolvedValue(undefined as never) + const session = { + id: "ses_free_limit_key", + projectID: "proj_test", + location: { directory }, + title: "Free limit no Go subscription", + agent: "build", + cost: 0, + tokens: { input: 0, output: 0, reasoning: 0, cache: { read: 0, write: 0 } }, + time: { created: 1, updated: 1 }, + } + const model = { id: "big-pickle", providerID: "opencode" } + const freeModel = { + id: "big-pickle", + providerID: "opencode", + name: "Big Pickle", + status: "active", + cost: [{ input: 0, output: 0, cache: { read: 0, write: 0 } }], + capabilities: { input: ["text"], output: ["text"], reasoning: false, toolCall: true, attachment: false }, + limit: { context: 200_000, output: 8_192 }, + variants: [], + time: { released: 1 }, + } + const goModel = { + id: "glm-5-go", + providerID: "opencode-go", + name: "GLM-5 Go", + status: "active", + cost: [{ input: 1, output: 2, cache: { read: 0, write: 0 } }], + capabilities: { input: ["text"], output: ["text"], reasoning: false, toolCall: true, attachment: false }, + limit: { context: 200_000, output: 8_192 }, + variants: [], + time: { released: 2 }, + } + const error = { + type: "provider.free-limit", + message: "Free usage limit reached.", + status: 429, + retryAfterMs: 3_600_000, + } + let releaseCatalog!: () => void + const catalogGate = new Promise((resolve) => { + releaseCatalog = resolve + }) + let oauthGate = Promise.resolve() + let modelSyncGate = Promise.resolve() + let onModelSyncRequest: (() => void) | undefined + let failModelSync = false + let connected = false + let needsAuth = false + let goEntitled = false + let oauthConnectCalls = 0 + const switchedModels: unknown[] = [] + const resentPrompts: unknown[] = [] + const promptResent = Promise.withResolvers() + await using setup = await createAppFixture({ + width: 100, + state: state.path, + args: { sessionID: session.id }, + config: { animations: false, tabs: { mode: "off" } }, + fetch: async (url, request) => { + const location = { directory, project: { id: "proj_test", directory, canonical: directory } } + if (url.pathname === "/api/session") return json({ data: [session], cursor: {} }) + if (url.pathname === `/api/session/${session.id}`) return json({ data: session }) + if (url.pathname === "/api/agent") + return json({ location, data: [{ id: "build", mode: "primary", hidden: false, permissions: [] }] }) + if (url.pathname === "/api/provider") + return json({ + location, + data: [ + { id: "opencode", name: "OpenCode Zen" }, + ...(goEntitled ? [{ id: "opencode-go", integrationID: "opencode", name: "OpenCode Go" }] : []), + ], + }) + if (url.pathname === "/api/model") { + onModelSyncRequest?.() + await modelSyncGate + if (failModelSync) return json({ message: "Model catalog sync failed" }, { status: 500 }) + return json({ location, data: goEntitled ? [goModel, freeModel] : [freeModel] }) + } + if (url.pathname === "/api/integration") { + await catalogGate + return json({ + location, + data: [ + { + id: "opencode", + name: "OpenCode Console", + methods: [ + { id: "device", type: "oauth", label: "OpenCode Console account" }, + { type: "key", label: "API key (service account)" }, + ], + connections: connected + ? [ + { + type: "credential", + method: "oauth", + id: "cred_console", + label: "Personal", + ...(needsAuth ? { status: { type: "needs_auth", message: "Session expired" } } : {}), + }, + { type: "credential", method: "oauth", id: "cred_inactive", label: "Inactive backup" }, + ] + : [], + }, + ], + }) + } + if (url.pathname === "/api/integration/opencode/connect/oauth" && request.method === "POST") { + oauthConnectCalls += 1 + return json({ + location, + data: { + attemptID: "attempt_no_go", + mode: "auto", + url: "https://opencode.ai/console/device?user_code=NOGO-5678", + instructions: "Enter code: NOGO-5678", + }, + }) + } + if (url.pathname === "/api/integration/opencode/connect/oauth/attempt_no_go") { + if (request.method === "DELETE") return new Response(null, { status: 204 }) + await oauthGate + if (failModelSync) connected = true + return json({ location, data: { status: "complete" } }) + } + if (url.pathname === `/api/session/${session.id}/model` && request.method === "POST") { + switchedModels.push(await request.json()) + return new Response(null, { status: 204 }) + } + if (url.pathname === `/api/session/${session.id}/prompt` && request.method === "POST") { + resentPrompts.push(await request.json()) + promptResent.resolve() + return json({ data: {} }) + } + if (url.pathname === `/api/session/${session.id}/message`) + return json({ + data: [ + { + id: "msg_free_assistant", + type: "assistant", + agent: "build", + model, + content: [], + time: { created: 2 }, + }, + { id: "msg_free_user", type: "user", text: "Refactor the parser", time: { created: 1 } }, + ], + cursor: {}, + }) + if ([`/api/session/${session.id}/inbox`, `/api/session/${session.id}/permission`].includes(url.pathname)) + return json({ data: [] }) + return undefined + }, + }) + await setup.ready + await setup.waitForFrame((frame) => frame.includes("Refactor the parser")) + + // 1a. Escape mid-auth: cancelling while oauth.status is in flight must not select a model or resubmit the prompt. + let releaseMidAuth!: () => void + oauthGate = new Promise((resolve) => { + releaseMidAuth = resolve + }) + goEntitled = true + setup.events.emit({ + id: "evt_free_limit_mid_auth", + created: baseNow, + type: "session.step.failed", + durable: { aggregateID: session.id, seq: 1, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant", error }, + }) + await setup.waitForFrame((frame) => frame.includes("Free limit reached")) + setup.mockInput.pressEnter() + expect(openSpy).toHaveBeenCalledWith("https://opencode.ai/go") + await setup.waitForFrame((frame) => frame.includes("Connect an integration")) + releaseCatalog() + await setup.waitForFrame((frame) => frame.includes("Waiting for authorization…")) + setup.mockInput.pressEscape() + await setup.waitForFrame((frame) => !frame.includes("Waiting for authorization…")) + releaseMidAuth() + await waitForGoUpsellConnect() + expect(switchedModels).toHaveLength(0) + expect(resentPrompts).toHaveLength(0) + expect(setup.captureCharFrame().split("Refactor the parser").length).toBe(3) + + // 1b. Escape mid-sync: cancelling while model sync is awaited must not select a model or resubmit the prompt. + currentNow = baseNow + 86_400_000 + oauthGate = Promise.resolve() + let releaseMidSync!: () => void + const midSyncStarted = Promise.withResolvers() + onModelSyncRequest = () => midSyncStarted.resolve() + modelSyncGate = new Promise((resolve) => { + releaseMidSync = resolve + }) + await setup.mockInput.pressKey("c", { ctrl: true }) + await setup.waitForFrame((frame) => frame.split("Refactor the parser").length === 2) + setup.events.emit({ + id: "evt_free_limit_mid_sync_start", + created: 3, + type: "session.step.started", + durable: { aggregateID: session.id, seq: 2, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_2", agent: "build", model, started: 3 }, + }) + setup.events.emit({ + id: "evt_free_limit_mid_sync_fail", + created: currentNow, + type: "session.step.failed", + durable: { aggregateID: session.id, seq: 3, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_2", error }, + }) + await setup.waitForFrame((frame) => frame.includes("Free limit reached")) + setup.mockInput.pressEnter() + await setup.waitForFrame((frame) => frame.includes("Waiting for authorization…")) + await midSyncStarted.promise + setup.mockInput.pressEscape() + await setup.waitForFrame((frame) => !frame.includes("Waiting for authorization…")) + releaseMidSync() + await waitForGoUpsellConnect() + onModelSyncRequest = undefined + expect(switchedModels).toHaveLength(0) + expect(resentPrompts).toHaveLength(0) + expect(setup.captureCharFrame().split("Refactor the parser").length).toBe(3) + + // 2. Model-sync rejection: catches the failure, closes "Waiting for authorization…", shows the error, and keeps the recalled prompt. + currentNow = baseNow + 2 * 86_400_000 + modelSyncGate = Promise.resolve() + failModelSync = true + await setup.mockInput.pressKey("c", { ctrl: true }) + await setup.waitForFrame((frame) => frame.split("Refactor the parser").length === 2) + setup.events.emit({ + id: "evt_free_limit_sync_err_start", + created: 5, + type: "session.step.started", + durable: { aggregateID: session.id, seq: 4, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_3", agent: "build", model, started: 5 }, + }) + setup.events.emit({ + id: "evt_free_limit_sync_err_fail", + created: currentNow, + type: "session.step.failed", + durable: { aggregateID: session.id, seq: 5, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_3", error }, + }) + await setup.waitForFrame((frame) => frame.includes("Free limit reached")) + setup.mockInput.pressEnter() + await setup.waitForFrame( + (frame) => !frame.includes("Waiting for authorization…") && frame.includes("UnexpectedStatus: 500"), + ) + expect(switchedModels).toHaveLength(0) + expect(resentPrompts).toHaveLength(0) + expect(setup.captureCharFrame().split("Refactor the parser").length).toBe(3) + + // 3a. Expired/needs_auth active credential (even with a status-free inactive backup): refresh detects needs_auth and routes to device sign-in without submitting. + failModelSync = false + goEntitled = true + connected = true + needsAuth = true + let releaseNeedsAuthOAuth!: () => void + oauthGate = new Promise((resolve) => { + releaseNeedsAuthOAuth = resolve + }) + currentNow = baseNow + 3 * 86_400_000 + await setup.mockInput.pressKey("c", { ctrl: true }) + await setup.waitForFrame((frame) => frame.split("Refactor the parser").length === 2) + setup.events.emit({ + id: "evt_free_limit_expired_start", + created: 6, + type: "session.step.started", + durable: { aggregateID: session.id, seq: 6, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_expired", agent: "build", model, started: 6 }, + }) + setup.events.emit({ + id: "evt_free_limit_expired_fail", + created: currentNow, + type: "session.step.failed", + durable: { aggregateID: session.id, seq: 7, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_expired", error }, + }) + await setup.waitForFrame((frame) => frame.includes("Free limit reached")) + const oauthCallsBeforeExpired = oauthConnectCalls + setup.mockInput.pressEnter() + await setup.waitForFrame((frame) => frame.includes("Waiting for authorization…") && frame.includes("NOGO-5678")) + expect(oauthConnectCalls).toBe(oauthCallsBeforeExpired + 1) + expect(switchedModels).toHaveLength(0) + expect(resentPrompts).toHaveLength(0) + // Complete sign-in so the refreshed active credential becomes valid again, then cancel before testing already-connected. + setup.mockInput.pressEscape() + await setup.waitForFrame((frame) => !frame.includes("Waiting for authorization…")) + releaseNeedsAuthOAuth() + await waitForGoUpsellConnect() + + // 3b. Cached active credential with needs_auth (and status-free inactive backup): goes straight to device sign-in, then once re-authenticated without Go lands on /models. + oauthGate = Promise.resolve() + goEntitled = false + currentNow = baseNow + 4 * 86_400_000 + await setup.mockInput.pressKey("c", { ctrl: true }) + await setup.waitForFrame((frame) => frame.split("Refactor the parser").length === 2) + setup.events.emit({ + id: "evt_free_limit_no_go_start", + created: 7, + type: "session.step.started", + durable: { aggregateID: session.id, seq: 8, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_4", agent: "build", model, started: 7 }, + }) + setup.events.emit({ + id: "evt_free_limit_no_go_fail", + created: currentNow, + type: "session.step.failed", + durable: { aggregateID: session.id, seq: 9, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_4", error }, + }) + await setup.waitForFrame((frame) => frame.includes("Free limit reached")) + needsAuth = false + const oauthCallsBeforeCachedExpired = oauthConnectCalls + setup.mockInput.pressEnter() + await setup.waitForFrame((frame) => frame.includes("Select model") && frame.includes("Big Pickle")) + expect(oauthConnectCalls).toBe(oauthCallsBeforeCachedExpired + 1) + setup.mockInput.pressEscape() + await setup.waitForFrame((frame) => !frame.includes("Select model")) + + // 3c. Already-connected valid account with Go entitlement: skips device OAuth, refreshes catalog, selects Go, and resends the prompt. + const oauthCallsBeforeAlreadyConnected = oauthConnectCalls + goEntitled = true + currentNow = baseNow + 5 * 86_400_000 + await setup.mockInput.pressKey("c", { ctrl: true }) + await setup.waitForFrame((frame) => frame.split("Refactor the parser").length === 2) + setup.events.emit({ + id: "evt_free_limit_already_go_start", + created: 9, + type: "session.step.started", + durable: { aggregateID: session.id, seq: 10, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_5", agent: "build", model, started: 9 }, + }) + setup.events.emit({ + id: "evt_free_limit_already_go_fail", + created: currentNow, + type: "session.step.failed", + durable: { aggregateID: session.id, seq: 11, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_5", error }, + }) + await setup.waitForFrame((frame) => frame.includes("Free limit reached")) + setup.mockInput.pressEnter() + await promptResent.promise + await setup.waitForFrame((frame) => !frame.includes("Free limit reached") && frame.includes("GLM-5 Go")) + expect(oauthConnectCalls).toBe(oauthCallsBeforeAlreadyConnected) + expect(switchedModels).toEqual([{ model: { providerID: "opencode-go", id: "glm-5-go" } }]) + expect(resentPrompts).toHaveLength(1) + expect(resentPrompts[0]).toMatchObject({ text: "Refactor the parser" }) + + // Advance past 24h, trigger again, and choose "don't show again". + currentNow = baseNow + 6 * 86_400_000 + setup.events.emit({ + id: "evt_free_limit_key_2", + created: 11, + type: "session.step.started", + durable: { aggregateID: session.id, seq: 12, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_6", agent: "build", model, started: 11 }, + }) + setup.events.emit({ + id: "evt_free_limit_key_3", + created: currentNow, + type: "session.step.failed", + durable: { aggregateID: session.id, seq: 13, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_6", error }, + }) + await setup.waitForFrame((frame) => frame.includes("Free limit reached")) + setup.mockInput.pressArrow("right") + setup.mockInput.pressEnter() + await setup.waitForFrame((frame) => !frame.includes("Free limit reached")) + await Bun.sleep(30) + + // Restart against the same state directory past another 48h: fresh in-memory store still respects persisted dontShow. + currentNow = baseNow + 8 * 86_400_000 + await using restarted = await createAppFixture({ + width: 100, + state: state.path, + args: { sessionID: session.id }, + config: { animations: false, tabs: { mode: "off" } }, + fetch: (url) => { + if (url.pathname === "/api/session") return json({ data: [session], cursor: {} }) + if (url.pathname === `/api/session/${session.id}`) return json({ data: session }) + if (url.pathname === "/api/integration") + return json({ + location: { directory, project: { id: "proj_test", directory, canonical: directory } }, + data: [{ id: "opencode-go", name: "OpenCode Go", methods: [{ type: "key" }], connections: [] }], + }) + if (url.pathname === `/api/session/${session.id}/message`) + return json({ + data: [ + { + id: "msg_free_assistant_3", + type: "assistant", + agent: "build", + model, + content: [], + time: { created: 8 }, + }, + { id: "msg_free_user", type: "user", text: "Refactor the parser", time: { created: 1 } }, + ], + cursor: {}, + }) + if ([`/api/session/${session.id}/inbox`, `/api/session/${session.id}/permission`].includes(url.pathname)) + return json({ data: [] }) + return undefined + }, + }) + await restarted.ready + await restarted.waitForFrame((frame) => frame.includes("Refactor the parser")) + restarted.events.emit({ + id: "evt_free_limit_key_restart", + created: currentNow, + type: "session.step.failed", + durable: { aggregateID: session.id, seq: 1, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant_3", error }, + }) + await restarted.waitForFrame((frame) => frame.includes("You've used today's free messages. They reset at")) + await Bun.sleep(30) + await restarted.renderOnce() + expect(restarted.captureCharFrame()).not.toContain("Free limit reached") +}) + +test("free usage limit claims suppression atomically across two TUI instances and protects typing drafts", async () => { + await using state = await tmpdir() + const session = { + id: "ses_free_limit_race", + projectID: "proj_test", + location: { directory }, + title: "Free limit race", + cost: 0, + tokens: { input: 0, output: 0, reasoning: 0, cache: { read: 0, write: 0 } }, + time: { created: 1, updated: 1 }, + } + const model = { id: "big-pickle", providerID: "opencode" } + const error = { + type: "provider.free-limit", + message: "Free usage limit reached.", + status: 429, + retryAfterMs: 3_600_000, + } + const fetch = (url: URL) => { + if (url.pathname === "/api/session") return json({ data: [session], cursor: {} }) + if (url.pathname === `/api/session/${session.id}`) return json({ data: session }) + if (url.pathname === "/api/integration") + return json({ + location: { directory, project: { id: "proj_test", directory, canonical: directory } }, + data: [ + { + id: "opencode-go", + name: "OpenCode Go", + methods: [{ id: "device", type: "oauth", label: "Sign in with your OpenCode account" }, { type: "key" }], + connections: [], + }, + ], + }) + if (url.pathname === `/api/session/${session.id}/message`) + return json({ + data: [ + { + id: "msg_free_assistant", + type: "assistant", + agent: "build", + model, + content: [], + time: { created: 2 }, + }, + { id: "msg_free_user", type: "user", text: "Refactor the parser", time: { created: 1 } }, + ], + cursor: {}, + }) + if ([`/api/session/${session.id}/inbox`, `/api/session/${session.id}/permission`].includes(url.pathname)) + return json({ data: [] }) + return undefined + } + await using first = await createAppFixture({ + width: 100, + state: state.path, + args: { sessionID: session.id }, + config: { animations: false, tabs: { mode: "off" } }, + fetch, + }) + await using second = await createAppFixture({ + width: 100, + state: state.path, + args: { sessionID: session.id }, + config: { animations: false, tabs: { mode: "off" } }, + fetch, + }) + await Promise.all([first.ready, second.ready]) + await Promise.all([ + first.waitForFrame((frame) => frame.includes("Refactor the parser")), + second.waitForFrame((frame) => frame.includes("Refactor the parser")), + ]) + // Emit concurrently in both processes: only one claims the 24h window and opens the modal. + const event = { + id: "evt_free_limit_race_1", + created: 3, + type: "session.step.failed" as const, + durable: { aggregateID: session.id, seq: 1, version: 1 as const }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant", error }, + } + first.events.emit(event) + second.events.emit(event) + await Promise.all([ + first.waitForFrame((frame) => frame.includes("You've used today's free messages. They reset at")), + second.waitForFrame((frame) => frame.includes("You've used today's free messages. They reset at")), + ]) + await first.waitForFrame( + () => + first.captureCharFrame().includes("Free limit reached") || + second.captureCharFrame().includes("Free limit reached"), + ) + await Bun.sleep(30) + await Promise.all([first.renderOnce(), second.renderOnce()]) + const opened = [first.captureCharFrame(), second.captureCharFrame()].filter((frame) => + frame.includes("Free limit reached"), + ) + expect(opened).toHaveLength(1) + + // Typing draft protection: does not overwrite a draft or open Free limit reached while typing. + await using typingState = await tmpdir() + await using typing = await createAppFixture({ + width: 100, + state: typingState.path, + args: { sessionID: session.id }, + config: { animations: false, tabs: { mode: "off" } }, + fetch, + }) + await typing.ready + await typing.waitForFrame((frame) => frame.includes("Refactor the parser")) + await typing.mockInput.typeText("Draft in progress") + await typing.waitForFrame((frame) => frame.includes("Draft in progress")) + typing.events.emit(event) + await typing.waitForFrame((frame) => frame.includes("You've used today's free messages. They reset at")) + const frame = typing.captureCharFrame() + expect(frame).toContain("Draft in progress") + expect(frame.split("Refactor the parser").length).toBe(2) + expect(frame).not.toContain("Free limit reached") + expect(frame).not.toContain("Connect OpenCode Go") +}) + +test("free usage limit modal manages live animation FPS and adapts to 44x24 and 60x12 terminals", async () => { + const session = { + id: "ses_free_limit_anim", + projectID: "proj_test", + location: { directory }, + title: "Free limit anim", + cost: 0, + tokens: { input: 0, output: 0, reasoning: 0, cache: { read: 0, write: 0 } }, + time: { created: 1, updated: 1 }, + } + const model = { id: "big-pickle", providerID: "opencode" } + const error = { + type: "provider.free-limit", + message: "Free usage limit reached.", + status: 429, + retryAfterMs: 3_600_000, + } + const fetch = (url: URL) => { + if (url.pathname === "/api/session") return json({ data: [session], cursor: {} }) + if (url.pathname === `/api/session/${session.id}`) return json({ data: session }) + if (url.pathname === `/api/session/${session.id}/message`) + return json({ + data: [ + { + id: "msg_free_assistant", + type: "assistant", + agent: "build", + model, + content: [], + time: { created: 2 }, + }, + { id: "msg_free_user", type: "user", text: "Refactor the parser", time: { created: 1 } }, + ], + cursor: {}, + }) + if ([`/api/session/${session.id}/inbox`, `/api/session/${session.id}/permission`].includes(url.pathname)) + return json({ data: [] }) + return undefined + } + const findArt = (node: Renderable): GoUpsellArtRenderable | undefined => { + if (node instanceof GoUpsellArtRenderable) return node + for (const child of node.getChildren()) { + const found = findArt(child) + if (found) return found + } + return undefined + } + + await using animState = await tmpdir() + await using anim = await createAppFixture({ + width: 100, + height: 30, + state: animState.path, + args: { sessionID: session.id }, + config: { animations: true, tabs: { mode: "off" } }, + fetch, + }) + await anim.ready + await anim.waitForFrame((frame) => frame.includes("Refactor the parser")) + const initialTargetFps = anim.renderer.targetFps + const initialMaxFps = anim.renderer.maxFps + const initialLive = anim.renderer.liveRequestCount + anim.events.emit({ + id: "evt_free_limit_anim_1", + created: 3, + type: "session.step.failed", + durable: { aggregateID: session.id, seq: 1, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant", error }, + }) + await anim.waitForFrame((frame) => frame.includes("Free limit reached")) + const art = findArt(anim.renderer.root) + expect(art).toBeDefined() + expect(art!.live).toBe(true) + expect(art!.isDestroyed).toBe(false) + expect(anim.renderer.targetFps).toBe(30) + expect(anim.renderer.maxFps).toBe(30) + // Toggling animate off disables live updates; toggling back on re-enables them. + art!.animate = false + expect(art!.live).toBe(false) + art!.animate = true + expect(art!.live).toBe(true) + // Closing the modal destroys the art renderable and restores FPS. + anim.mockInput.pressEscape() + await anim.waitForFrame((frame) => !frame.includes("Free limit reached")) + expect(art!.isDestroyed).toBe(true) + expect(art!.live).toBe(false) + expect(findArt(anim.renderer.root)).toBeUndefined() + expect(anim.renderer.liveRequestCount).toBe(initialLive) + expect(anim.renderer.targetFps).toBe(initialTargetFps) + expect(anim.renderer.maxFps).toBe(initialMaxFps) + + // Compact sizes: 44x24 (narrow, stacked buttons) and 60x12 (short, scrollable body). + for (const [width, height] of [ + [44, 24], + [60, 12], + ] as const) { + await using compactState = await tmpdir() + await using compact = await createAppFixture({ + width, + height, + state: compactState.path, + args: { sessionID: session.id }, + config: { animations: true, tabs: { mode: "off" } }, + fetch, + }) + await compact.ready + await compact.waitForFrame((frame) => frame.includes("Refactor the parser")) + compact.events.emit({ + id: `evt_free_limit_${width}x${height}`, + created: 3, + type: "session.step.failed", + durable: { aggregateID: session.id, seq: 1, version: 1 }, + data: { sessionID: session.id, assistantMessageID: "msg_free_assistant", error }, + }) + await compact.waitForFrame((frame) => frame.includes("Free limit reached")) + const frame = compact.captureCharFrame() + expect(frame).toContain("Free limit reached") + expect(frame).toContain("don't show again") + expect(frame).toContain("dismiss") + expect(frame).toContain("subscribe") + expect(findArt(compact.renderer.root)).toBeUndefined() + // Locate "dismiss" in the rendered frame and click it with the mouse. + const lines = frame.split("\n") + const row = lines.findIndex((line) => line.includes("dismiss")) + expect(row).toBeGreaterThanOrEqual(0) + const col = lines[row]!.indexOf("dismiss") + 2 + await compact.mockMouse.click(col, row) + await compact.waitForFrame((next) => !next.includes("Free limit reached")) + } +}) + test.each([44, 100])( "retry countdown updates and clears with the retry lifecycle at width %s", async (width) => { diff --git a/packages/tui/test/cli/cmd/tui/notifications.test.ts b/packages/tui/test/cli/cmd/tui/notifications.test.ts index 75ef0d263fe5..6e2ed7e2cdb5 100644 --- a/packages/tui/test/cli/cmd/tui/notifications.test.ts +++ b/packages/tui/test/cli/cmd/tui/notifications.test.ts @@ -157,6 +157,19 @@ describe("internal notifications TUI plugin", () => { harness.emit(executionStarted("retry")) harness.emit(executionFailed("failed-again")) expect(harness.toasts).toHaveLength(2) + + harness.emit(executionStarted("free-limit")) + harness.emit({ + id: "free-limit-failed", + created: 0, + type: "session.execution.failed", + data: { + sessionID: "session", + error: { type: "provider.free-limit", message: "Free usage limit reached." }, + }, + } as OpenCodeEvent) + expect(harness.toasts).toHaveLength(2) + expect(harness.notifications).toHaveLength(2) }) test.each([{ type: "home" }, { type: "session", sessionID: "other" }])( diff --git a/packages/tui/test/component/go-upsell-art.test.tsx b/packages/tui/test/component/go-upsell-art.test.tsx new file mode 100644 index 000000000000..28f8c6e5e99e --- /dev/null +++ b/packages/tui/test/component/go-upsell-art.test.tsx @@ -0,0 +1,87 @@ +import { expect, spyOn, test } from "bun:test" +/** @jsxImportSource @opentui/solid */ +import { OptimizedBuffer, RGBA } from "@opentui/core" +import { testRender } from "@opentui/solid" +import { createSignal, Show } from "solid-js" +import "../../src/component/go-upsell-art" +import type { GoUpsellArtRenderable } from "../../src/component/go-upsell-art" + +const colors = { + background: RGBA.fromHex("#0a0a0a"), + primary: RGBA.fromHex("#fab283"), + logoBase: RGBA.fromHex("#9a9a9a"), +} + +test("the art stops every frame when hidden, closed, or still, and leaves nothing live after repeated opens", async () => { + const [open, setOpen] = createSignal(true) + const [visible, setVisible] = createSignal(true) + const [animate, setAnimate] = createSignal(true) + let art: GoUpsellArtRenderable | undefined + const app = await testRender( + () => ( + + + (art = value)} + width="100%" + height="100%" + visible={visible()} + animate={animate()} + {...colors} + /> + + + ), + { width: 60, height: 20 }, + ) + const draws = spyOn(OptimizedBuffer.prototype, "withBuffers") + const setCell = spyOn(OptimizedBuffer.prototype, "setCell") + + try { + await app.renderOnce() + expect(app.renderer.liveRequestCount).toBe(1) + // Each frame is one direct write into the art's own buffer, never a draw call per cell. + expect(draws).toHaveBeenCalledTimes(1) + expect(setCell).not.toHaveBeenCalled() + expect(app.captureCharFrame()).toContain("█▀▀▀ █▀▀█") + + setVisible(false) + expect(app.renderer.liveRequestCount).toBe(0) + setVisible(true) + expect(app.renderer.liveRequestCount).toBe(1) + + // A still frame is drawn once and then reused. + setAnimate(false) + expect(app.renderer.liveRequestCount).toBe(0) + draws.mockClear() + await app.renderOnce() + await app.renderOnce() + expect(draws).toHaveBeenCalledTimes(1) + setAnimate(true) + + for (let i = 0; i < 20; i++) { + const previous = art! + setOpen(false) + expect(app.renderer.liveRequestCount).toBe(0) + expect(app.renderer.root.liveCount).toBe(0) + await app.renderOnce() + expect(previous.isDestroyed).toBe(true) + expect(previous.live).toBe(false) + setOpen(true) + expect(art).not.toBe(previous) + expect(app.renderer.liveRequestCount).toBe(1) + } + + setOpen(false) + await app.renderOnce() + expect(app.renderer.liveRequestCount).toBe(0) + expect(app.renderer.isRunning).toBe(false) + const frames = app.renderer.frameId + await Bun.sleep(120) + expect(app.renderer.frameId).toBe(frames) + } finally { + draws.mockRestore() + setCell.mockRestore() + app.renderer.destroy() + } +}) diff --git a/packages/ui/bunfig.toml b/packages/ui/bunfig.toml new file mode 100644 index 000000000000..ca0d68a75990 --- /dev/null +++ b/packages/ui/bunfig.toml @@ -0,0 +1,2 @@ +[test] +preload = ["./test/solid-jsx.ts"] diff --git a/packages/ui/package.json b/packages/ui/package.json index b110129324e5..19d7ff615f69 100644 --- a/packages/ui/package.json +++ b/packages/ui/package.json @@ -1,6 +1,6 @@ { "name": "@opencode/ui", - "version": "2.1.0", + "version": "2.1.2", "type": "module", "license": "MIT", "repository": { @@ -113,15 +113,14 @@ "@kitlangton/solid-motion": "0.2.2", "@kobalte/core": "catalog:", "@pierre/diffs": "catalog:", + "@pierre/highlights": "catalog:", "@solid-primitives/event-listener": "catalog:", "@solid-primitives/resize-observer": "catalog:", "fuzzysort": "catalog:", "katex": "0.16.47", "marked": "catalog:", - "marked-shiki": "catalog:", "motion": "12.34.5", "remeda": "catalog:", - "shiki": "catalog:", "solid-list": "catalog:", "solid-sonner": "catalog:" }, diff --git a/packages/ui/src/context/highlight-theme.tsx b/packages/ui/src/context/highlight-theme.tsx new file mode 100644 index 000000000000..a432d95244e1 --- /dev/null +++ b/packages/ui/src/context/highlight-theme.tsx @@ -0,0 +1,66 @@ +import type { Theme } from "@pierre/highlights" + +// A Highlights palette: each value names an app CSS variable without its `--`, so code follows the active app theme. +// Token types are grouped as in Pierre's own `pierre-dark` theme (for example, properties share the variable colour +// and all punctuation is one muted colour); each group uses the app syntax role with the same meaning. +// Token names fall back to their parent scope, so `keyword.control` uses `keyword`. +export const OpenCodeTheme = { + name: "OpenCode", + appearance: "dark", + cssVariables: { prefix: "--" }, + style: { + "editor.background": "color-background-stronger", + "editor.foreground": "text-base", + created: "syntax-diff-add", + deleted: "syntax-diff-delete", + modified: "syntax-diff-unknown", + syntax: { + attribute: "syntax-property", + background: "color-background-stronger", + boolean: "syntax-constant", + comment: "syntax-comment", + constant: "syntax-constant", + constructor: "syntax-type", + "diff.minus": "syntax-diff-delete", + "diff.plus": "syntax-diff-add", + embedded: "text-base", + emphasis: "text-base", + enum: "syntax-type", + foreground: "text-base", + function: "syntax-primitive", + hint: "syntax-comment", + keyword: "syntax-keyword", + "keyword.operator": "syntax-operator", + label: "syntax-constant", + link_text: "syntax-info", + link_uri: "syntax-keyword", + namespace: "syntax-constant", + number: "syntax-constant", + operator: "syntax-operator", + predictive: { color: "syntax-comment", font_style: "italic" }, + preproc: "syntax-keyword", + primary: "syntax-info", + property: "syntax-variable", + punctuation: "syntax-punctuation", + "punctuation.special": "syntax-keyword", + selector: "syntax-keyword", + "selector.class": "syntax-property", + "selector.id": "syntax-primitive", + "selector.pseudo": "syntax-operator", + string: "syntax-string", + "string.escape": "syntax-constant", + "string.regex": "syntax-regexp", + "string.special": "syntax-constant", + tag: "syntax-keyword", + "tag.component.jsx": "syntax-type", + text: "text-base", + "text.literal": "syntax-string", + title: { color: "syntax-keyword", font_weight: 700 }, + type: "syntax-type", + variable: "syntax-variable", + "variable.parameter": "text-base", + "variable.special": "syntax-constant", + variant: "syntax-constant", + }, + }, +} satisfies Theme diff --git a/packages/ui/src/context/marked-parser.test.ts b/packages/ui/src/context/marked-parser.test.ts index 30093e45726f..ec0dcbc60b9b 100644 --- a/packages/ui/src/context/marked-parser.test.ts +++ b/packages/ui/src/context/marked-parser.test.ts @@ -2,7 +2,7 @@ import { expect, test } from "bun:test" import { createMarkdownParser } from "./marked-parser" import { parseSmallMarkdown } from "./marked-base" -const parser = createMarkdownParser((code, language) => `
${code}
`) +const parser = createMarkdownParser() test.each([ ["OpenCode", "OpenCode"], @@ -84,12 +84,23 @@ test("lexes dollar-heavy paragraphs in near-linear time", () => { expect(performance.now() - startedAt).toBeLessThan(500) }) -test("uses the configured code highlighter", async () => { - expect(await parser.parse("```ts\nconst value = 1\n```\n")).toBe('
const value = 1
\n') +test("highlights fenced code with the app theme variables", async () => { + const html = await parser.parse("```ts\nconst value = 1\n```\n") + + expect(html).toStartWith('
')
+  expect(html).toContain('const ')
+})
+
+test("keeps the fence language for code without a lexer", async () => {
+  const html = await parser.parse("```mermaid\ngraph TD; A-->B\n```\n")
+
+  expect(html).toContain('graph TD; A-->B')
 })
 
 test.each(["```", "~~~"])("recognizes an empty %s fence at EOF", async (fence) => {
-  expect(await parser.parse(`foo\n${fence}`)).toBe('
foo
\n
\n')
+  expect(await parser.parse(`foo\n${fence}`)).toContain('
foo
\n
')
 })
 
 test("preserves emphasis when rejecting an outer reference link", async () => {
diff --git a/packages/ui/src/context/marked-parser.tsx b/packages/ui/src/context/marked-parser.tsx
index 70869eb544c8..06d9b47909f5 100644
--- a/packages/ui/src/context/marked-parser.tsx
+++ b/packages/ui/src/context/marked-parser.tsx
@@ -1,10 +1,28 @@
+import { codeToHtml, isSupportedLanguage } from "@pierre/highlights"
 import katex from "katex"
 import type { MarkedExtension, Tokens } from "marked"
-import markedShiki from "marked-shiki"
 import { createMarkdownBase } from "./marked-base"
+import { OpenCodeTheme } from "./highlight-theme"
 
-export function createMarkdownParser(highlight: (code: string, language: string) => string | Promise) {
-  return createMarkdownBase().use(markdownMath, markedShiki({ highlight }))
+export function createMarkdownParser() {
+  return createMarkdownBase().use(markdownMath, markdownCode)
+}
+
+const decoder = new TextDecoder()
+
+// The class keeps the fence's own language, even one without a lexer, so the viewer can label it and render Mermaid.
+const markdownCode: MarkedExtension = {
+  renderer: {
+    code(token) {
+      const language = (token.lang?.match(/^\S*/)?.[0] ?? "").toLowerCase().replace(/[^\w#+.-]/g, "") || "text"
+
+      const html = decoder.decode(
+        codeToHtml(token.text, { lang: isSupportedLanguage(language) ? language : "text", theme: OpenCodeTheme }),
+      )
+
+      return html.replace("", ``) + "\n"
+    },
+  },
 }
 
 // `$$...$$` renders as display math anywhere in a line.
diff --git a/packages/ui/src/context/marked-theme-register.tsx b/packages/ui/src/context/marked-theme-register.tsx
deleted file mode 100644
index 135381eab8d2..000000000000
--- a/packages/ui/src/context/marked-theme-register.tsx
+++ /dev/null
@@ -1,10 +0,0 @@
-import { registerCustomTheme } from "@pierre/diffs"
-import { OpenCodeTheme } from "./marked-theme"
-
-let registered = false
-
-export function registerOpenCodeTheme() {
-  if (registered) return
-  registered = true
-  registerCustomTheme("OpenCode", () => Promise.resolve(OpenCodeTheme))
-}
diff --git a/packages/ui/src/context/marked-theme.tsx b/packages/ui/src/context/marked-theme.tsx
deleted file mode 100644
index 527a35e678f0..000000000000
--- a/packages/ui/src/context/marked-theme.tsx
+++ /dev/null
@@ -1,372 +0,0 @@
-import type { ThemeRegistrationResolved } from "@pierre/diffs"
-
-export const OpenCodeTheme = {
-  name: "OpenCode",
-  bg: "var(--color-background-stronger)",
-  fg: "var(--text-base)",
-  colors: {
-    "editor.background": "var(--color-background-stronger)",
-    "editor.foreground": "var(--text-base)",
-    "gitDecoration.addedResourceForeground": "var(--syntax-diff-add)",
-    "gitDecoration.deletedResourceForeground": "var(--syntax-diff-delete)",
-    "gitDecoration.modifiedResourceForeground": "var(--syntax-diff-unknown)",
-    // "gitDecoration.conflictingResourceForeground": "#ffca00",
-    // "gitDecoration.modifiedResourceForeground": "#1a76d4",
-    // "gitDecoration.untrackedResourceForeground": "#00cab1",
-    // "gitDecoration.ignoredResourceForeground": "#84848A",
-    // "terminal.titleForeground": "#adadb1",
-    // "terminal.titleInactiveForeground": "#84848A",
-    // "terminal.background": "#141415",
-    // "terminal.foreground": "#adadb1",
-    // "terminal.ansiBlack": "#141415",
-    // "terminal.ansiRed": "#ff2e3f",
-    // "terminal.ansiGreen": "#0dbe4e",
-    // "terminal.ansiYellow": "#ffca00",
-    // "terminal.ansiBlue": "#008cff",
-    // "terminal.ansiMagenta": "#c635e4",
-    // "terminal.ansiCyan": "#08c0ef",
-    // "terminal.ansiWhite": "#c6c6c8",
-    // "terminal.ansiBrightBlack": "#141415",
-    // "terminal.ansiBrightRed": "#ff2e3f",
-    // "terminal.ansiBrightGreen": "#0dbe4e",
-    // "terminal.ansiBrightYellow": "#ffca00",
-    // "terminal.ansiBrightBlue": "#008cff",
-    // "terminal.ansiBrightMagenta": "#c635e4",
-    // "terminal.ansiBrightCyan": "#08c0ef",
-    // "terminal.ansiBrightWhite": "#c6c6c8",
-  },
-  tokenColors: [
-    {
-      scope: ["comment", "punctuation.definition.comment", "string.comment"],
-      settings: {
-        foreground: "var(--syntax-comment)",
-      },
-    },
-    {
-      scope: ["entity.other.attribute-name"],
-      settings: {
-        foreground: "var(--syntax-property)", // maybe attribute
-      },
-    },
-    {
-      scope: ["constant", "entity.name.constant", "variable.other.constant", "variable.language", "entity"],
-      settings: {
-        foreground: "var(--syntax-constant)",
-      },
-    },
-    {
-      scope: ["entity.name", "meta.export.default", "meta.definition.variable"],
-      settings: {
-        foreground: "var(--syntax-type)",
-      },
-    },
-    {
-      scope: ["meta.object.member"],
-      settings: {
-        foreground: "var(--syntax-primitive)",
-      },
-    },
-    {
-      scope: [
-        "variable.parameter.function",
-        "meta.jsx.children",
-        "meta.block",
-        "meta.tag.attributes",
-        "entity.name.constant",
-        "meta.embedded.expression",
-        "meta.template.expression",
-        "string.other.begin.yaml",
-        "string.other.end.yaml",
-      ],
-      settings: {
-        foreground: "var(--syntax-punctuation)",
-      },
-    },
-    {
-      scope: ["entity.name.function", "support.type.primitive"],
-      settings: {
-        foreground: "var(--syntax-primitive)",
-      },
-    },
-    {
-      scope: ["support.class.component"],
-      settings: {
-        foreground: "var(--syntax-type)",
-      },
-    },
-    {
-      scope: "keyword",
-      settings: {
-        foreground: "var(--syntax-keyword)",
-      },
-    },
-    {
-      scope: [
-        "keyword.operator",
-        "storage.type.function.arrow",
-        "punctuation.separator.key-value.css",
-        "entity.name.tag.yaml",
-        "punctuation.separator.key-value.mapping.yaml",
-      ],
-      settings: {
-        foreground: "var(--syntax-operator)",
-      },
-    },
-    {
-      scope: ["storage", "storage.type"],
-      settings: {
-        foreground: "var(--syntax-keyword)",
-      },
-    },
-    {
-      scope: ["storage.modifier.package", "storage.modifier.import", "storage.type.java"],
-      settings: {
-        foreground: "var(--syntax-primitive)",
-      },
-    },
-    {
-      scope: [
-        "string",
-        "punctuation.definition.string",
-        "string punctuation.section.embedded source",
-        "entity.name.tag",
-      ],
-      settings: {
-        foreground: "var(--syntax-string)",
-      },
-    },
-    {
-      scope: "support",
-      settings: {
-        foreground: "var(--syntax-primitive)",
-      },
-    },
-    {
-      scope: ["support.type.object.module", "variable.other.object", "support.type.property-name.css"],
-      settings: {
-        foreground: "var(--syntax-object)",
-      },
-    },
-    {
-      scope: "meta.property-name",
-      settings: {
-        foreground: "var(--syntax-property)",
-      },
-    },
-    {
-      scope: "variable",
-      settings: {
-        foreground: "var(--syntax-variable)",
-      },
-    },
-    {
-      scope: "variable.other",
-      settings: {
-        foreground: "var(--syntax-variable)",
-      },
-    },
-    {
-      scope: [
-        "invalid.broken",
-        "invalid.illegal",
-        "invalid.unimplemented",
-        "invalid.deprecated",
-        "message.error",
-        "markup.deleted",
-        "meta.diff.header.from-file",
-        "punctuation.definition.deleted",
-        "brackethighlighter.unmatched",
-        "token.error-token",
-      ],
-      settings: {
-        foreground: "var(--syntax-critical)",
-      },
-    },
-    {
-      scope: "carriage-return",
-      settings: {
-        foreground: "var(--syntax-keyword)",
-      },
-    },
-    {
-      scope: "string source",
-      settings: {
-        foreground: "var(--syntax-variable)",
-      },
-    },
-    {
-      scope: "string variable",
-      settings: {
-        foreground: "var(--syntax-constant)",
-      },
-    },
-    {
-      scope: [
-        "source.regexp",
-        "string.regexp",
-        "string.regexp.character-class",
-        "string.regexp constant.character.escape",
-        "string.regexp source.ruby.embedded",
-        "string.regexp string.regexp.arbitrary-repitition",
-        "string.regexp constant.character.escape",
-      ],
-      settings: {
-        foreground: "var(--syntax-regexp)",
-      },
-    },
-    {
-      scope: "support.constant",
-      settings: {
-        foreground: "var(--syntax-primitive)",
-      },
-    },
-    {
-      scope: "support.variable",
-      settings: {
-        foreground: "var(--syntax-variable)",
-      },
-    },
-    {
-      scope: "meta.module-reference",
-      settings: {
-        foreground: "var(--syntax-info)",
-      },
-    },
-    {
-      scope: "punctuation.definition.list.begin.markdown",
-      settings: {
-        foreground: "var(--syntax-punctuation)",
-      },
-    },
-    {
-      scope: ["markup.heading", "markup.heading entity.name"],
-      settings: {
-        fontStyle: "bold",
-        foreground: "var(--syntax-info)",
-      },
-    },
-    {
-      scope: "markup.quote",
-      settings: {
-        foreground: "var(--syntax-info)",
-      },
-    },
-    {
-      scope: "markup.italic",
-      settings: {
-        fontStyle: "italic",
-        // foreground: "",
-      },
-    },
-    {
-      scope: "markup.bold",
-      settings: {
-        fontStyle: "bold",
-        foreground: "var(--text-strong)",
-      },
-    },
-    {
-      scope: [
-        "markup.raw",
-        "markup.inserted",
-        "meta.diff.header.to-file",
-        "punctuation.definition.inserted",
-        "markup.changed",
-        "punctuation.definition.changed",
-        "markup.ignored",
-        "markup.untracked",
-      ],
-      settings: {
-        foreground: "var(--text-base)",
-      },
-    },
-    {
-      scope: "meta.diff.range",
-      settings: {
-        fontStyle: "bold",
-        foreground: "var(--syntax-unknown)",
-      },
-    },
-    {
-      scope: "meta.diff.header",
-      settings: {
-        foreground: "var(--syntax-unknown)",
-      },
-    },
-    {
-      scope: "meta.separator",
-      settings: {
-        fontStyle: "bold",
-        foreground: "var(--syntax-unknown)",
-      },
-    },
-    {
-      scope: "meta.output",
-      settings: {
-        foreground: "var(--syntax-unknown)",
-      },
-    },
-    {
-      scope: "meta.export.default",
-      settings: {
-        foreground: "var(--syntax-unknown)",
-      },
-    },
-    {
-      scope: [
-        "brackethighlighter.tag",
-        "brackethighlighter.curly",
-        "brackethighlighter.round",
-        "brackethighlighter.square",
-        "brackethighlighter.angle",
-        "brackethighlighter.quote",
-      ],
-      settings: {
-        foreground: "var(--syntax-unknown)",
-      },
-    },
-    {
-      scope: ["constant.other.reference.link", "string.other.link"],
-      settings: {
-        fontStyle: "underline",
-        foreground: "var(--syntax-unknown)",
-      },
-    },
-    {
-      scope: "token.info-token",
-      settings: {
-        foreground: "var(--syntax-info)",
-      },
-    },
-    {
-      scope: "token.warn-token",
-      settings: {
-        foreground: "var(--syntax-warning)",
-      },
-    },
-    {
-      scope: "token.debug-token",
-      settings: {
-        foreground: "var(--syntax-info)",
-      },
-    },
-  ],
-  semanticTokenColors: {
-    comment: "var(--syntax-comment)",
-    string: "var(--syntax-string)",
-    number: "var(--syntax-constant)",
-    regexp: "var(--syntax-regexp)",
-    keyword: "var(--syntax-keyword)",
-    variable: "var(--syntax-variable)",
-    parameter: "var(--syntax-variable)",
-    property: "var(--syntax-property)",
-    function: "var(--syntax-primitive)",
-    method: "var(--syntax-primitive)",
-    type: "var(--syntax-type)",
-    class: "var(--syntax-type)",
-    namespace: "var(--syntax-type)",
-    enumMember: "var(--syntax-primitive)",
-    "variable.constant": "var(--syntax-constant)",
-    "variable.defaultLibrary": "var(--syntax-unknown)",
-  },
-} as unknown as ThemeRegistrationResolved
diff --git a/packages/ui/src/context/marked.tsx b/packages/ui/src/context/marked.tsx
deleted file mode 100644
index e06aabc8ec3c..000000000000
--- a/packages/ui/src/context/marked.tsx
+++ /dev/null
@@ -1,31 +0,0 @@
-import { getSharedHighlighter } from "@pierre/diffs"
-import { bundledLanguages, type BundledLanguage } from "shiki"
-import { createSimpleContext } from "./helper"
-import { createMarkdownParser } from "./marked-parser"
-import { registerOpenCodeTheme } from "./marked-theme-register"
-
-export { OpenCodeTheme } from "./marked-theme"
-
-registerOpenCodeTheme()
-
-export const { use: useMarked, provider: MarkedProvider } = createSimpleContext({
-  name: "Marked",
-  init: () =>
-    createMarkdownParser(async (code, language) => {
-      const highlighter = await getSharedHighlighter({
-        themes: ["OpenCode"],
-        langs: [],
-        preferredHighlighter: "shiki-wasm",
-      })
-
-      const name = language in bundledLanguages ? language : "text"
-
-      if (!highlighter.getLoadedLanguages().includes(name)) await highlighter.loadLanguage(name as BundledLanguage)
-
-      return highlighter.codeToHtml(code, {
-        lang: name,
-        theme: "OpenCode",
-        tabindex: false,
-      })
-    }),
-})
diff --git a/packages/ui/src/i18n/en.ts b/packages/ui/src/i18n/en.ts
index 783b2e06b784..7186c1f0f472 100644
--- a/packages/ui/src/i18n/en.ts
+++ b/packages/ui/src/i18n/en.ts
@@ -75,12 +75,15 @@ const source = {
   "ui.sessionTurn.retry.attemptWaiting.other": "Attempt {{attempt}} - retrying in {{count}}s",
   "ui.sessionTurn.retry.attemptRetryingNow": "Attempt {{attempt}} - retrying",
   "ui.sessionTurn.retry.geminiHot": "gemini is way too hot right now",
+  "ui.sessionTurn.freeLimit.title": "You've used today's free messages.",
+  "ui.sessionTurn.freeLimit.lineReset": "You've used today's free messages. They reset at {{time}}.",
   "ui.sessionTurn.error.freeUsageExceeded": "Free usage exceeded",
   "ui.sessionTurn.error.addCredits": "Add credits",
 
   "dialog.usageExceeded.freeTier.title": "Free limit reached",
   "dialog.usageExceeded.freeTier.description":
     "Subscribe to OpenCode Go for reliable access to the best open-source models for $10/month.",
+  "dialog.usageExceeded.freeTier.reset": "Your free messages reset at {{time}}.",
   "dialog.usageExceeded.freeTier.actionLabel": "Subscribe",
   "dialog.usageExceeded.accountRateLimit.title": "Go limit reached",
   "dialog.usageExceeded.accountRateLimit.description":
diff --git a/packages/ui/test/solid-jsx.ts b/packages/ui/test/solid-jsx.ts
new file mode 100644
index 000000000000..4dd5861f43e3
--- /dev/null
+++ b/packages/ui/test/solid-jsx.ts
@@ -0,0 +1,20 @@
+import { plugin } from "bun"
+
+// Bun transpiles TSX against the tsconfig `jsxImportSource`, but `solid-js/jsx-runtime` has no runtime
+// exports because Solid's compiler normally replaces JSX. Tests run without that compiler, so compile
+// JSX against Solid's hyperscript runtime instead.
+const transpiler = new Bun.Transpiler({
+  loader: "tsx",
+  trimUnusedImports: true,
+  tsconfig: { compilerOptions: { jsx: "react-jsx", jsxImportSource: "solid-js/h" } },
+})
+
+plugin({
+  name: "solid-jsx-runtime",
+  setup(build) {
+    build.onLoad({ filter: /\.[jt]sx$/ }, async (args) => ({
+      contents: transpiler.transformSync(await Bun.file(args.path).text()),
+      loader: "js",
+    }))
+  },
+})
diff --git a/packages/util/package.json b/packages/util/package.json
index 67f9234f7b86..1f2022d22a57 100644
--- a/packages/util/package.json
+++ b/packages/util/package.json
@@ -1,7 +1,7 @@
 {
   "$schema": "https://json.schemastore.org/package.json",
   "name": "@opencode/util",
-  "version": "2.1.0",
+  "version": "2.1.2",
   "type": "module",
   "license": "MIT",
   "repository": {
diff --git a/patches/@silvia-odwyer%2Fphoton-node@0.3.4.patch b/patches/@silvia-odwyer%2Fphoton-node@0.3.4.patch
index f7267890aede..da41cb715e0a 100644
--- a/patches/@silvia-odwyer%2Fphoton-node@0.3.4.patch
+++ b/patches/@silvia-odwyer%2Fphoton-node@0.3.4.patch
@@ -274,9 +274,9 @@ index 8f4144d..29a964a 100644
      const offset = table.grow(4);
      table.set(0, undefined);
 @@ -4509,7 +4510,8 @@ module.exports.__wbindgen_init_externref_table = function() {
-      ;
-  };
-  
+     ;
+ };
+ 
 -const path = require('path').join(__dirname, 'photon_rs_bg.wasm');
 +// Allow opencode's Bun compiled binary to point photon-node at its embedded wasm asset.
 +const path = globalThis.__OPENCODE_PHOTON_WASM_PATH || require('path').join(__dirname, 'photon_rs_bg.wasm');
diff --git a/script/duplicate-pr.ts b/script/duplicate-pr.ts
deleted file mode 100755
index 901756f8818a..000000000000
--- a/script/duplicate-pr.ts
+++ /dev/null
@@ -1,81 +0,0 @@
-#!/usr/bin/env bun
-
-/* TODO: Migrate this script to the new in-process @opencode/sdk.
-import path from "path"
-import { pathToFileURL } from "bun"
-import { createOpencode } from "@opencode/sdk"
-import { parseArgs } from "util"
-
-async function main() {
-  const { values, positionals } = parseArgs({
-    args: Bun.argv.slice(2),
-    options: {
-      file: { type: "string", short: "f" },
-      help: { type: "boolean", short: "h", default: false },
-    },
-    allowPositionals: true,
-  })
-
-  if (values.help) {
-    console.log(`
-Usage: bun script/duplicate-pr.ts [options] 
-
-Options:
-  -f, --file    File to attach to the prompt
-  -h, --help          Show this help message
-
-Examples:
-  bun script/duplicate-pr.ts -f pr_info.txt "Check the attached file for PR details"
-`)
-    process.exit(0)
-  }
-
-  const message = positionals.join(" ")
-  if (!message) {
-    console.error("Error: message is required")
-    process.exit(1)
-  }
-
-  const opencode = await createOpencode({ port: 0 })
-
-  try {
-    const parts: Array<{ type: "text"; text: string } | { type: "file"; url: string; filename: string; mime: string }> =
-      []
-
-    if (values.file) {
-      const resolved = path.resolve(process.cwd(), values.file)
-      const file = Bun.file(resolved)
-      if (!(await file.exists())) {
-        console.error(`Error: file not found: ${values.file}`)
-        process.exit(1)
-      }
-      parts.push({
-        type: "file",
-        url: pathToFileURL(resolved).href,
-        filename: path.basename(resolved),
-        mime: "text/plain",
-      })
-    }
-
-    parts.push({ type: "text", text: message })
-
-    const session = await opencode.client.session.create()
-    const result = await opencode.client.session
-      .prompt({
-        path: { id: session.data!.id },
-        body: {
-          agent: "duplicate-pr",
-          parts,
-        },
-        signal: AbortSignal.timeout(120_000),
-      })
-      .then((x) => x.data?.parts?.find((y) => y.type === "text")?.text ?? "")
-
-    console.log(result.trim())
-  } finally {
-    opencode.server.close()
-  }
-}
-
-void main()
-*/
diff --git a/services/update/package.json b/services/update/package.json
index 39137bf16a08..e7b3a28fed99 100644
--- a/services/update/package.json
+++ b/services/update/package.json
@@ -1,7 +1,7 @@
 {
   "$schema": "https://json.schemastore.org/package.json",
   "name": "@opencode/update",
-  "version": "2.1.0",
+  "version": "2.1.2",
   "private": true,
   "type": "module",
   "scripts": {
diff --git a/services/www/openapi.json b/services/www/openapi.json
index 829c7e20c1f5..1c68daeaadd1 100644
--- a/services/www/openapi.json
+++ b/services/www/openapi.json
@@ -5382,6 +5382,68 @@
         }
       }
     },
+    "/api/experimental/voice/models": {
+      "get": {
+        "tags": ["voice"],
+        "operationId": "experimental.voice.models",
+        "parameters": [],
+        "security": [],
+        "responses": {
+          "200": {
+            "description": "VoiceModelsResponse",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/VoiceModelsResponse"
+                }
+              }
+            }
+          },
+          "400": {
+            "description": "InvalidRequestError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/InvalidRequestErrorEncoded"
+                }
+              }
+            }
+          },
+          "401": {
+            "description": "UnauthorizedError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/UnauthorizedErrorEncoded"
+                }
+              }
+            }
+          },
+          "404": {
+            "description": "LocationNotFoundError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/LocationNotFoundErrorEncoded"
+                }
+              }
+            }
+          },
+          "503": {
+            "description": "ServiceUnavailableError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/ServiceUnavailableErrorEncoded"
+                }
+              }
+            }
+          }
+        },
+        "description": "List supported voice models, their transports, whether the server has a usable API key for each provider or the connected OpenCode Console hosts the model, and the Console workspace a hosted model bills to.",
+        "summary": "List voice models"
+      }
+    },
     "/api/experimental/voice/session": {
       "post": {
         "tags": ["voice"],
@@ -5440,7 +5502,7 @@
             }
           }
         },
-        "description": "Create a server-brokered voice session using the server's configured provider credentials and return the WebRTC SDP answer or WebSocket ephemeral token.",
+        "description": "Create a server-brokered voice session using the server's configured provider credentials, or the connected OpenCode Console when it hosts the model, and return the WebRTC SDP answer or WebSocket ephemeral token.",
         "summary": "Create voice session",
         "requestBody": {
           "content": {
@@ -5474,6 +5536,87 @@
         }
       }
     },
+    "/api/experimental/voice/session/{sessionID}/instructions": {
+      "post": {
+        "tags": ["voice"],
+        "operationId": "experimental.voice.instructions",
+        "parameters": [
+          {
+            "name": "sessionID",
+            "in": "path",
+            "schema": {
+              "type": "string"
+            },
+            "required": true
+          }
+        ],
+        "security": [],
+        "responses": {
+          "204": {
+            "description": ""
+          },
+          "400": {
+            "description": "InvalidRequestError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/InvalidRequestErrorEncoded"
+                }
+              }
+            }
+          },
+          "401": {
+            "description": "UnauthorizedError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/UnauthorizedErrorEncoded"
+                }
+              }
+            }
+          },
+          "404": {
+            "description": "LocationNotFoundError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/LocationNotFoundErrorEncoded"
+                }
+              }
+            }
+          },
+          "503": {
+            "description": "ServiceUnavailableError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/ServiceUnavailableErrorEncoded"
+                }
+              }
+            }
+          }
+        },
+        "description": "Replace the backend instructions of a running session the connected OpenCode Console hosts, whose data channel can't change them.",
+        "summary": "Replace voice session instructions",
+        "requestBody": {
+          "content": {
+            "application/json": {
+              "schema": {
+                "type": "object",
+                "properties": {
+                  "instructions": {
+                    "type": "string"
+                  }
+                },
+                "required": ["instructions"],
+                "additionalProperties": false
+              }
+            }
+          },
+          "required": true
+        }
+      }
+    },
     "/api/provider": {
       "get": {
         "tags": ["provider"],
@@ -14324,6 +14467,17 @@
                 "type": "integer",
                 "minimum": 0
               },
+              "location_idle_timeout": {
+                "anyOf": [
+                  {
+                    "type": "string"
+                  },
+                  {
+                    "type": "boolean",
+                    "enum": [false]
+                  }
+                ]
+              },
               "policies": {
                 "type": "array",
                 "items": {
@@ -19794,6 +19948,10 @@
             },
             "required": ["body"],
             "additionalProperties": false
+          },
+          "retryAfterMs": {
+            "type": "integer",
+            "minimum": 0
           }
         },
         "required": ["type", "message"],
@@ -20525,6 +20683,78 @@
         "required": ["_tag", "providerID", "message"],
         "additionalProperties": false
       },
+      "VoiceModel": {
+        "type": "object",
+        "properties": {
+          "model": {
+            "type": "string"
+          },
+          "provider": {
+            "type": "string"
+          },
+          "transports": {
+            "type": "array",
+            "items": {
+              "$ref": "#/components/schemas/VoiceTransportType"
+            }
+          },
+          "available": {
+            "type": "boolean"
+          },
+          "integration": {
+            "anyOf": [
+              {
+                "type": "string"
+              },
+              {
+                "type": "null"
+              }
+            ]
+          },
+          "workspace": {
+            "anyOf": [
+              {
+                "type": "string"
+              },
+              {
+                "type": "null"
+              }
+            ]
+          }
+        },
+        "required": ["model", "provider", "transports", "available"],
+        "additionalProperties": false
+      },
+      "VoiceModelsResponse": {
+        "type": "object",
+        "properties": {
+          "data": {
+            "type": "object",
+            "properties": {
+              "models": {
+                "type": "array",
+                "items": {
+                  "$ref": "#/components/schemas/VoiceModel"
+                }
+              },
+              "default": {
+                "anyOf": [
+                  {
+                    "type": "string"
+                  },
+                  {
+                    "type": "null"
+                  }
+                ]
+              }
+            },
+            "required": ["models"],
+            "additionalProperties": false
+          }
+        },
+        "required": ["data"],
+        "additionalProperties": false
+      },
       "VoiceSessionResponse": {
         "type": "object",
         "properties": {
@@ -20533,6 +20763,23 @@
             "properties": {
               "transport": {
                 "$ref": "#/components/schemas/VoiceTransport"
+              },
+              "session": {
+                "anyOf": [
+                  {
+                    "type": "object",
+                    "properties": {
+                      "id": {
+                        "type": "string"
+                      }
+                    },
+                    "required": ["id"],
+                    "additionalProperties": false
+                  },
+                  {
+                    "type": "null"
+                  }
+                ]
               }
             },
             "required": ["transport"],
@@ -20562,6 +20809,10 @@
           }
         ]
       },
+      "VoiceTransportType": {
+        "type": "string",
+        "enum": ["webrtc", "websocket"]
+      },
       "VoiceWebRTCTransport": {
         "type": "object",
         "properties": {
diff --git a/services/www/public/openapi.json b/services/www/public/openapi.json
index 829c7e20c1f5..1c68daeaadd1 100644
--- a/services/www/public/openapi.json
+++ b/services/www/public/openapi.json
@@ -5382,6 +5382,68 @@
         }
       }
     },
+    "/api/experimental/voice/models": {
+      "get": {
+        "tags": ["voice"],
+        "operationId": "experimental.voice.models",
+        "parameters": [],
+        "security": [],
+        "responses": {
+          "200": {
+            "description": "VoiceModelsResponse",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/VoiceModelsResponse"
+                }
+              }
+            }
+          },
+          "400": {
+            "description": "InvalidRequestError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/InvalidRequestErrorEncoded"
+                }
+              }
+            }
+          },
+          "401": {
+            "description": "UnauthorizedError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/UnauthorizedErrorEncoded"
+                }
+              }
+            }
+          },
+          "404": {
+            "description": "LocationNotFoundError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/LocationNotFoundErrorEncoded"
+                }
+              }
+            }
+          },
+          "503": {
+            "description": "ServiceUnavailableError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/ServiceUnavailableErrorEncoded"
+                }
+              }
+            }
+          }
+        },
+        "description": "List supported voice models, their transports, whether the server has a usable API key for each provider or the connected OpenCode Console hosts the model, and the Console workspace a hosted model bills to.",
+        "summary": "List voice models"
+      }
+    },
     "/api/experimental/voice/session": {
       "post": {
         "tags": ["voice"],
@@ -5440,7 +5502,7 @@
             }
           }
         },
-        "description": "Create a server-brokered voice session using the server's configured provider credentials and return the WebRTC SDP answer or WebSocket ephemeral token.",
+        "description": "Create a server-brokered voice session using the server's configured provider credentials, or the connected OpenCode Console when it hosts the model, and return the WebRTC SDP answer or WebSocket ephemeral token.",
         "summary": "Create voice session",
         "requestBody": {
           "content": {
@@ -5474,6 +5536,87 @@
         }
       }
     },
+    "/api/experimental/voice/session/{sessionID}/instructions": {
+      "post": {
+        "tags": ["voice"],
+        "operationId": "experimental.voice.instructions",
+        "parameters": [
+          {
+            "name": "sessionID",
+            "in": "path",
+            "schema": {
+              "type": "string"
+            },
+            "required": true
+          }
+        ],
+        "security": [],
+        "responses": {
+          "204": {
+            "description": ""
+          },
+          "400": {
+            "description": "InvalidRequestError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/InvalidRequestErrorEncoded"
+                }
+              }
+            }
+          },
+          "401": {
+            "description": "UnauthorizedError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/UnauthorizedErrorEncoded"
+                }
+              }
+            }
+          },
+          "404": {
+            "description": "LocationNotFoundError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/LocationNotFoundErrorEncoded"
+                }
+              }
+            }
+          },
+          "503": {
+            "description": "ServiceUnavailableError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/ServiceUnavailableErrorEncoded"
+                }
+              }
+            }
+          }
+        },
+        "description": "Replace the backend instructions of a running session the connected OpenCode Console hosts, whose data channel can't change them.",
+        "summary": "Replace voice session instructions",
+        "requestBody": {
+          "content": {
+            "application/json": {
+              "schema": {
+                "type": "object",
+                "properties": {
+                  "instructions": {
+                    "type": "string"
+                  }
+                },
+                "required": ["instructions"],
+                "additionalProperties": false
+              }
+            }
+          },
+          "required": true
+        }
+      }
+    },
     "/api/provider": {
       "get": {
         "tags": ["provider"],
@@ -14324,6 +14467,17 @@
                 "type": "integer",
                 "minimum": 0
               },
+              "location_idle_timeout": {
+                "anyOf": [
+                  {
+                    "type": "string"
+                  },
+                  {
+                    "type": "boolean",
+                    "enum": [false]
+                  }
+                ]
+              },
               "policies": {
                 "type": "array",
                 "items": {
@@ -19794,6 +19948,10 @@
             },
             "required": ["body"],
             "additionalProperties": false
+          },
+          "retryAfterMs": {
+            "type": "integer",
+            "minimum": 0
           }
         },
         "required": ["type", "message"],
@@ -20525,6 +20683,78 @@
         "required": ["_tag", "providerID", "message"],
         "additionalProperties": false
       },
+      "VoiceModel": {
+        "type": "object",
+        "properties": {
+          "model": {
+            "type": "string"
+          },
+          "provider": {
+            "type": "string"
+          },
+          "transports": {
+            "type": "array",
+            "items": {
+              "$ref": "#/components/schemas/VoiceTransportType"
+            }
+          },
+          "available": {
+            "type": "boolean"
+          },
+          "integration": {
+            "anyOf": [
+              {
+                "type": "string"
+              },
+              {
+                "type": "null"
+              }
+            ]
+          },
+          "workspace": {
+            "anyOf": [
+              {
+                "type": "string"
+              },
+              {
+                "type": "null"
+              }
+            ]
+          }
+        },
+        "required": ["model", "provider", "transports", "available"],
+        "additionalProperties": false
+      },
+      "VoiceModelsResponse": {
+        "type": "object",
+        "properties": {
+          "data": {
+            "type": "object",
+            "properties": {
+              "models": {
+                "type": "array",
+                "items": {
+                  "$ref": "#/components/schemas/VoiceModel"
+                }
+              },
+              "default": {
+                "anyOf": [
+                  {
+                    "type": "string"
+                  },
+                  {
+                    "type": "null"
+                  }
+                ]
+              }
+            },
+            "required": ["models"],
+            "additionalProperties": false
+          }
+        },
+        "required": ["data"],
+        "additionalProperties": false
+      },
       "VoiceSessionResponse": {
         "type": "object",
         "properties": {
@@ -20533,6 +20763,23 @@
             "properties": {
               "transport": {
                 "$ref": "#/components/schemas/VoiceTransport"
+              },
+              "session": {
+                "anyOf": [
+                  {
+                    "type": "object",
+                    "properties": {
+                      "id": {
+                        "type": "string"
+                      }
+                    },
+                    "required": ["id"],
+                    "additionalProperties": false
+                  },
+                  {
+                    "type": "null"
+                  }
+                ]
               }
             },
             "required": ["transport"],
@@ -20562,6 +20809,10 @@
           }
         ]
       },
+      "VoiceTransportType": {
+        "type": "string",
+        "enum": ["webrtc", "websocket"]
+      },
       "VoiceWebRTCTransport": {
         "type": "object",
         "properties": {
diff --git a/services/www/src/docs/content/cli/web.mdx b/services/www/src/docs/content/cli/web.mdx
index da553409204c..de349965d256 100644
--- a/services/www/src/docs/content/cli/web.mdx
+++ b/services/www/src/docs/content/cli/web.mdx
@@ -38,9 +38,10 @@ platform:
 | Platform              | Default port |
 | --------------------- | ------------ |
 | macOS, Linux, and WSL | 49374        |
-| Windows               | 45278        |
+| Windows               | 28894        |
 
-Windows reserves ports from 49152 up. Print the current address with
+Windows reserves ports from 49152 up, and WSL mirrored networking reserves
+another block below that. Print the current address with
 `opencode service status`:
 
 ```bash
diff --git a/services/www/src/docs/content/config.mdx b/services/www/src/docs/content/config.mdx
index 314088d13b8b..b89d3ea1626c 100644
--- a/services/www/src/docs/content/config.mdx
+++ b/services/www/src/docs/content/config.mdx
@@ -410,20 +410,6 @@ interval and 30-minute active window.
 See the [warming guide](/warming) for request behavior, customization,
 and cost considerations.
 
-### Experimental
-
-Experimental features may change or be removed in a future release.
-
-Set `todo` to `true` to give agents a `todowrite` tool for tracking multi-step work.
-
-```jsonc
-{
-  "experimental": {
-    "todo": true,
-  },
-}
-```
-
 ### Skills
 
 Add directories or URLs that OpenCode should search for agent skills.
@@ -623,3 +609,58 @@ These are useful for:
 
 - Keeping sensitive data like API keys in separate files.
 - Sharing common configuration snippets across multiple config files.
+
+## Experimental
+
+
+  Experimental settings may change, break, or be removed in any release without notice. Don't rely on them for stable
+  workflows.
+
+
+All experimental settings live under the `experimental` key:
+
+```jsonc title="opencode.jsonc"
+{
+  "$schema": "https://opencode.ai/config.json",
+  "experimental": {
+    "todo": true,
+    "subagent_depth": 2,
+    "portable_shell_scanner": true,
+    "location_idle_timeout": "4 hours",
+  },
+}
+```
+
+### Todos
+
+Set `todo` to `true` to give agents a `todowrite` tool for tracking multi-step work. Defaults to `false`.
+
+### Subagent depth
+
+`subagent_depth` sets how deeply subagents can start their own subagents. Defaults to `1`, so a subagent cannot
+start another subagent.
+
+### Portable shell scanner
+
+Set `portable_shell_scanner` to `true` to replace the default tree-sitter shell scanner used by permission checks.
+Defaults to `false`. See [Permissions](/permissions#scanner) for how it behaves.
+
+### Location idle timeout
+
+`location_idle_timeout` controls how long a project location can go without session activity before OpenCode unloads
+its services. It takes a duration such as `"2 hours"` and defaults to `"60 minutes"`. Running sessions and background
+jobs keep a location loaded. Set it to `false` to never unload idle locations. Project config can override the global
+value.
+
+```jsonc
+{
+  "experimental": {
+    "location_idle_timeout": false,
+  },
+}
+```
+
+### Deprecated policies
+
+`experimental.policies` is deprecated. Use top-level [`policies`](/policies) instead. The old key still works and
+merges into `policies`.
diff --git a/services/www/src/docs/content/console/go.mdx b/services/www/src/docs/content/console/go.mdx
index 6079372ca738..6c8ebb7e0a58 100644
--- a/services/www/src/docs/content/console/go.mdx
+++ b/services/www/src/docs/content/console/go.mdx
@@ -136,7 +136,7 @@ Token prices are per 1M tokens.
     | Kimi K2.7 Code                          | $0.95  | $4.00  | $0.19       | -            | **$60**                                               |
     | LongCat-2.0                             | $0.30  | $1.20  | $0.006      | -            | **$60**                                               |
     | LongCat 2.5 Preview Free                | Free   | Free   | Free        | -            | **Unlimited**
limited time        |
-    | Step 5 Preview                          | $1.00  | $2.70  | $0.10       | -            | ~~$15~~ **$60**
4x · Limited time |
+    | Step 5 Preview                          | $1.00  | $2.70  | $0.10       | -            | **$15**                                               |
     | Beam Free                               | Free   | Free   | Free        | -            | **Unlimited**
limited time        |
     | MiMo-V2.6-Flash                         | $0.14  | $0.28  | $0.0028     | -            | **$60**                                               |
     | MiMo-V2.6-Pro                           | $0.435 | $0.87  | $0.003625   | -            | **$15**                                               |
@@ -178,7 +178,7 @@ Token prices are per 1M tokens.
     | Kimi K2.7 Code                          | $0.95  | $4.00  | $0.19       | -            | **$180**                                                |
     | LongCat-2.0                             | $0.30  | $1.20  | $0.006      | -            | **$240**                                                |
     | LongCat 2.5 Preview Free                | Free   | Free   | Free        | -            | **Unlimited**
limited time          |
-    | Step 5 Preview                          | $1.00  | $2.70  | $0.10       | -            | ~~$60~~ **$240**
4x · Limited time  |
+    | Step 5 Preview                          | $1.00  | $2.70  | $0.10       | -            | **$60**                                                 |
     | Beam Free                               | Free   | Free   | Free        | -            | **Unlimited**
limited time          |
     | MiMo-V2.6-Flash                         | $0.14  | $0.28  | $0.0028     | -            | **$120**                                                |
     | MiMo-V2.6-Pro                           | $0.435 | $0.87  | $0.003625   | -            | **$60**                                                 |
@@ -236,7 +236,7 @@ Plus estimates scale with each model's higher usage limit.
     | Kimi K2.7 Code                                                   | 1,350                | 3,380                  | 6,750                    |
     | LongCat-2.0                                                      | 11,400               | 28,600                 | 57,200                   |
     | LongCat 2.5 Preview Free                                         | Unlimited            | Unlimited              | Unlimited                |
-    | Step 5 Preview
4x · Limited time             | ~~430~~
**1,720** | ~~1,070~~
**4,280** | ~~2,130~~
**8,520** |
+    | Step 5 Preview                                                   | 430                  | 1,070                  | 2,130                    |
     | Beam Free                                                        | Unlimited            | Unlimited              | Unlimited                |
     | MiMo-V2.6-Flash                                                  | 30,100               | 75,200                 | 150,400                  |
     | MiMo-V2.6-Pro                                                    | 3,250                | 8,150                  | 16,300                   |
@@ -271,7 +271,7 @@ Plus estimates scale with each model's higher usage limit.
     | Kimi K2.7 Code                                                   | 4,050                  | 10,140                   | 20,250                   |
     | LongCat-2.0                                                      | 45,600                 | 114,400                  | 228,800                  |
     | LongCat 2.5 Preview Free                                         | Unlimited              | Unlimited                | Unlimited                |
-    | Step 5 Preview
4x · Limited time             | ~~1,720~~
**6,880** | ~~4,280~~
**17,120** | ~~8,520~~
**34,080** |
+    | Step 5 Preview                                                   | 1,720                  | 4,280                    | 8,520                    |
     | Beam Free                                                        | Unlimited              | Unlimited                | Unlimited                |
     | MiMo-V2.6-Flash                                                  | 60,200                 | 150,400                  | 300,800                  |
     | MiMo-V2.6-Pro                                                    | 13,000                 | 32,600                   | 65,200                   |
diff --git a/services/www/src/docs/content/console/models.mdx b/services/www/src/docs/content/console/models.mdx
index b8ebcf31bbab..41210ea06019 100644
--- a/services/www/src/docs/content/console/models.mdx
+++ b/services/www/src/docs/content/console/models.mdx
@@ -84,6 +84,7 @@ You can also access the models directly through the following API endpoints.
 | Grok 4.6                        | grok-4.6                        | `https://opencode.ai/inference/openai/v1/responses`                        | `@ai-sdk/openai`            |
 | Grok 4.5                        | grok-4.5                        | `https://opencode.ai/inference/openai/v1/responses`                        | `@ai-sdk/openai`            |
 | Grok Build 0.1                  | grok-build-0.1                  | `https://opencode.ai/inference/openai/v1/responses`                        | `@ai-sdk/openai`            |
+| Muse Wonder 1.0                 | muse-wonder-1.0                 | `https://opencode.ai/inference/openai/v1/responses`                        | `@ai-sdk/openai`            |
 | Muse Spark 1.3                  | muse-spark-1.3                  | `https://opencode.ai/inference/openai/v1/responses`                        | `@ai-sdk/openai`            |
 | Muse Spark 1.2                  | muse-spark-1.2                  | `https://opencode.ai/inference/openai/v1/responses`                        | `@ai-sdk/openai`            |
 | Qwen3.8 Max                     | qwen3.8-max                     | `https://opencode.ai/inference/openai/v1/chat/completions`                 | `@ai-sdk/openai-compatible` |
@@ -114,7 +115,7 @@ You can also access the models directly through the following API endpoints.
 | DeepSeek V4.1 Flash Free        | deepseek-v4.1-flash-free        | `https://opencode.ai/inference/openai/v1/chat/completions`                 | `@ai-sdk/openai-compatible` |
 | MiniMax M3.1 Flash Free         | minimax-m3.1-flash-free         | `https://opencode.ai/inference/openai/v1/chat/completions`                 | `@ai-sdk/openai-compatible` |
 | Beam Free                       | beam-free                       | `https://opencode.ai/inference/openai/v1/chat/completions`                 | `@ai-sdk/openai-compatible` |
-| LL Alpha Free                   | ll-alpha-free                   | `https://opencode.ai/inference/openai/v1/chat/completions`                 | `@ai-sdk/openai-compatible` |
+| Flint 40B Free                  | flint-40b-free                  | `https://opencode.ai/inference/openai/v1/chat/completions`                 | `@ai-sdk/openai-compatible` |
 | LongCat 2.5 Preview Free        | longcat-2.5-preview-free        | `https://opencode.ai/inference/openai/v1/chat/completions`                 | `@ai-sdk/openai-compatible` |
 | Exo Free                        | exo-free                        | `https://opencode.ai/inference/openai/v1/chat/completions`                 | `@ai-sdk/openai-compatible` |
 | MiMo-V2.6-Flash Free            | mimo-v2.6-flash-free            | `https://opencode.ai/inference/openai/v1/chat/completions`                 | `@ai-sdk/openai-compatible` |
@@ -208,7 +209,7 @@ Console uses pay-as-you-go pricing. Below are the prices **per 1M tokens**.
 | DeepSeek V4.1 Flash Free          | Free            | Free            | Free            | -            |
 | MiniMax M3.1 Flash Free           | Free            | Free            | Free            | -            |
 | Beam Free                         | Free            | Free            | Free            | -            |
-| LL Alpha Free                     | Free            | Free            | Free            | -            |
+| Flint 40B Free                    | Free            | Free            | Free            | -            |
 | LongCat 2.5 Preview Free          | Free            | Free            | Free            | -            |
 | Exo Free                          | Free            | Free            | Free            | -            |
 | MiMo-V2.6-Flash Free              | Free            | Free            | Free            | -            |
@@ -271,14 +272,15 @@ Console uses pay-as-you-go pricing. Below are the prices **per 1M tokens**.
 | Grok 4.5 (≤ 200K tokens)          | $2.00           | $6.00           | $0.30           | -            |
 | Grok 4.5 (> 200K tokens)          | $4.00           | $12.00          | $0.60           | -            |
 | Grok Build 0.1                    | $1.00           | $2.00           | $0.20           | -            |
+| Muse Wonder 1.0                   | $4.00           | $20.00          | $0.20           | -            |
 | Muse Spark 1.3                    | $1.25           | $4.25           | $0.15           | -            |
 | Muse Spark 1.2                    | $1.25           | $4.25           | $0.15           | -            |
 | GPT 6 Astra (≤ 272K tokens)       | $10.00          | $50.00          | $1.00           | $12.50       |
 | GPT 6 Astra (> 272K tokens)       | $20.00          | $75.00          | $2.00           | $25.00       |
 | GPT 6 Sol (≤ 272K tokens)         | $2.00           | $10.00          | $0.20           | $2.50        |
 | GPT 6 Sol (> 272K tokens)         | $4.00           | $15.00          | $0.40           | $5.00        |
-| GPT 6.1 Sol (≤ 272K tokens)       | $2.00           | $10.00          | $0.20           | $2.50        |
-| GPT 6.1 Sol (> 272K tokens)       | $4.00           | $15.00          | $0.40           | $5.00        |
+| GPT 6.1 Sol (≤ 272K tokens)       | $2.00           | $10.00          | $0.10           | $2.50        |
+| GPT 6.1 Sol (> 272K tokens)       | $4.00           | $15.00          | $0.20           | $5.00        |
 | GPT 6 Luna (≤ 272K tokens)        | $0.10           | $0.50           | $0.01           | $0.125       |
 | GPT 6 Luna (> 272K tokens)        | $0.20           | $0.75           | $0.02           | $0.25        |
 | GPT 5.6 Sol (≤ 272K tokens)       | $4.00           | $20.00          | $0.40           | $5.00        |
@@ -331,10 +333,12 @@ models to generate session titles.
 - Big Pickle is a stealth model that's free on OpenCode for a limited time. The team is using this time to collect feedback and improve the model.
 - MiniMax M3.1 Flash Free is free on OpenCode for a limited time. Its provider follows a zero-retention policy and does not use your data for model training.
 - Beam Free is free on OpenCode for a limited time. Its provider does not use your data for model training and retains it for 30 days for safety reasons. It's not available in Angola, Belarus, Cambodia, China (including Hong Kong and Macau), Cuba, the Democratic Republic of the Congo, Equatorial Guinea, Iran, Iraq, Laos, Libya, Nigeria, North Korea, Russia, Syria, Turkmenistan, Uzbekistan, Venezuela, or the Crimea, Donetsk, and Luhansk regions of Ukraine.
-- LL Alpha Free is a stealth model that's free on OpenCode for a limited time. The team is using this time to collect feedback and improve the model.
+- Flint 40B Free is free on OpenCode for a limited time. The team is using this time to collect feedback and improve the model.
 - LongCat 2.5 Preview Free is free on OpenCode for a limited time. Its provider follows a zero-retention policy and does not use your data for model training.
 - Jev 1.13 Free is available on OpenCode for a limited time.
 
+To hide free models in OpenCode, set `OPENCODE_DISABLE_FREE_MODELS=1`. See [free models](/models#free-models).
+
 [Contact us](mailto:help@anoma.ly) if you have any questions.
 
 ### Auto-recharge and budgets
@@ -370,7 +374,7 @@ models to generate session titles.
 All our models are hosted in the US and EU. Our providers follow a zero-retention policy and do not use your data for model training, with the following exceptions:
 
 - Big Pickle: During its free period, collected data may be used to improve the model.
-- LL Alpha Free: During its free period, collected data may be used to improve the model. Data is retained for 30 days.
+- Flint 40B Free: During its free period, collected data may be used to improve the model. Data is retained for 30 days.
 - Exo Free: During its free period, collected data may be used to improve the model.
 - MiMo-V2.6-Flash Free: During its free period, collected data may be used to improve the model.
 - MiMo-V2.5 Free: During its free period, collected data may be used to improve the model.
diff --git a/services/www/src/docs/content/models.mdx b/services/www/src/docs/content/models.mdx
index 6549acc2d29d..b23204cbc42f 100644
--- a/services/www/src/docs/content/models.mdx
+++ b/services/www/src/docs/content/models.mdx
@@ -57,6 +57,24 @@ falls back to the newest available supported model.
 
 See [Config](/config) for configuration locations and precedence.
 
+## Free models
+
+Set `OPENCODE_DISABLE_FREE_MODELS=1` (or `true`) on the server process to hide [free OpenCode models](/console/models#free-models):
+
+```bash
+OPENCODE_DISABLE_FREE_MODELS=1 opencode serve
+```
+
+For the shared background server:
+
+```bash
+opencode service set env OPENCODE_DISABLE_FREE_MODELS 1
+```
+
+Free means an `opencode` or `opencode-go` model with no input price. These models disappear from the catalog and defaults;
+explicitly selecting one fails with `Model unavailable`. Signed out, the `opencode` provider is absent until you connect
+an OpenCode Console account or set `OPENCODE_API_KEY`, which restores its paid models.
+
 ## Variants
 
 Variants are named options for one model, often used for reasoning effort or token budgets. Add `#variant` when selecting
diff --git a/services/www/src/docs/content/permissions.mdx b/services/www/src/docs/content/permissions.mdx
index e9c997253a15..31d26e6dcf03 100644
--- a/services/www/src/docs/content/permissions.mdx
+++ b/services/www/src/docs/content/permissions.mdx
@@ -190,7 +190,8 @@ so authorize only trusted directory boundaries.
 
 ## Scanner
 
-Enable the experimental portable shell scanner with:
+Enable the experimental portable shell scanner with the following setting. Like all
+[experimental settings](/config#experimental), it may change or be removed in a future release.
 
 ```jsonc
 {
diff --git a/sst-env.d.ts b/sst-env.d.ts
index bcfa83ba237d..781d74e87f88 100644
--- a/sst-env.d.ts
+++ b/sst-env.d.ts
@@ -19,32 +19,14 @@ declare module "sst" {
       "value": string
     }
     "Api": import("@cloudflare/workers-types").Service
-    "Bucket": import("@cloudflare/workers-types").R2Bucket
     "Console": {
       "type": "sst.cloudflare.SolidStart"
       "url": string
     }
-    "DISCORD_SUPPORT_BOT_TOKEN": {
-      "type": "sst.sst.Secret"
-      "value": string
-    }
-    "DISCORD_SUPPORT_CHANNEL_ID": {
-      "type": "sst.sst.Secret"
-      "value": string
-    }
     "EMAILOCTOPUS_API_KEY": {
       "type": "sst.sst.Secret"
       "value": string
     }
-    "EnterpriseStorage": import("@cloudflare/workers-types").R2Bucket
-    "FEISHU_APP_ID": {
-      "type": "sst.sst.Secret"
-      "value": string
-    }
-    "FEISHU_APP_SECRET": {
-      "type": "sst.sst.Secret"
-      "value": string
-    }
     "GITHUB_APP_ID": {
       "type": "sst.sst.Secret"
       "value": string